Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The effect of cationically-modified phosphorylcholine polymers on human osteoblasts in vitro and their effect on bone formation in vivo

The effect of cationically-modified phosphorylcholine polymers on human osteoblasts in vitro and... J Mater Sci: Mater Med (2017) 28:144 DOI 10.1007/s10856-017-5958-8 BIOCOMPATIBILITY STUDIES Original Research The effect of cationically-modified phosphorylcholine polymers on human osteoblasts in vitro and their effect on bone formation in vivo 1 1 2 2 ● ● ● ● Jonathan M. Lawton Mariam Habib Bingkui Ma Roger A. Brooks 1 3 2 1 ● ● ● Serena M. Best Andrew L. Lewis Neil Rushton William Bonfield Received: 30 January 2017 / Accepted: 3 August 2017 / Published online: 17 August 2017 © Springer Science+Business Media, LLC 2017 Abstract The effect of introducing cationic charge into biotolerant, stimulating the production of fibrous tissue and phosphorylcholine (PC)-based polymers has been investi- areas of loosely associated matrix (LAM) around the gated in this study with a view to using these materials as implant. Their development, as formulated in this study, as coatings to improve bone formation and osseointegration at bone interfacing implant coatings is therefore not warranted. the bone-implant interface. PC-based polymers, which have been used in a variety of medical devices to improve bio- Graphical abstract compatibility, are associated with low protein adsorption resulting in reduced complement activation, inflammatory response and cell adhesion. However, in some applications, such as orthopaedics, good integration between the implant and bone is needed to allow the distribution of loading stresses and a bioactive response is required. It has pre- viously been shown that the incorporation of cationic charge into PC-based polymers may increase protein adsorption that stimulates subsequent cell adhesion. In this paper, the effect of cationic charge in PC-based polymers on human osteoblasts (HObs) in vitro and the effect of these polymers on bone formation in the rat tibia was assessed. Increasing PC positive surface charge increased HOb cell adhesion and stimulated increased cell differentiation and the production of calcium phosphate deposits. However, when implanted in bone these materials were at best Jonathan M. Lawton and Mariam Habib contributed equally to this work * Andrew L. Lewis 1 Introduction andrew.lewis@btgplc.com Advances in modern medicine, such as the introduction of Department of Materials Science and Metallurgy, Cambridge penicillin, antiseptics, and vaccinations, have significantly Centre for Medical Materials, University of Cambridge, New Museum Site, Cambridge CB2 3QZ, UK increased human life expectancy [1]. The ageing population in the West (in the UK and US alone there are over 100 Orthopaedic Research Unit, University of Cambridge, Addenbrookes Hospital, Hills Road, Cambridge CB2 2QQ, UK million individuals over the age of 50) [2], coupled with the deterioration of bone stock with age (particularly for woman Biocompatibles UK Ltd, Chapman House, Farnham Business Park, Weydon Lane, Farnham, Surrey GU9 8QL, UK of post-menopausal age) [3], mean that the clinical need for 144 Page 2 of 13 J Mater Sci: Mater Med (2017) 28:144 tissue repair and replacement has never been greater. Bone biomineralization [24–26]. This suggested the potential of tissue loss, predominantly due to degenerative diseases PC polymer to also interact with calcium and form a stable (such as osteoporosis), osteosarcoma and trauma [4], has led interface with bone mineral. However, osteoblast activity at to the design of a variety of surgical techniques, medical the interface would be required for osteointegration and devices and specialised materials. These range from total unmodified PC polymer has reduced cellular interaction. hip and knee replacements to fracture fixation devices and Attempts have been made to modify PC polymers to pro- bone stock replacements. duce materials with inherent biocompatibility but also with Phosphorylcholine (PC) materials are bio-inspired poly- additional components to invoke specific biological inter- mers that mimic the extracellular surface of red blood cells, actions. In one study, a PC-containing copolymer of N- containing an exact chemical copy of the predominant isopropylacrylamide and N-(n-octadecyl) acrylamide was zwitterionic phospholipid headgroup found in the cell lipid shown to preferentially adhere U937 macrophages, which membrane. Unlike most biomaterials, the well-hydrated and subsequently increased expression of TNF-α in response to 2+ 3+ neutrally-charged PC surface allows for the interaction of Co and Co ions [27]. In another approach, researchers proteins without inducing shape changes in the protein’s have also shown that the introduction of surface charge into three-dimensional structure and thus reduce irreversible biomaterials increases cell spreading, adhesion, growth and protein adsorption [5]. Furthermore, this decrease in protein proliferation [28–31]. Rose et al [32] demonstrated that the adsorption results in decreased blood clotting [6], cellular introduction of a cationic moiety (choline methacrylate adhesion, and in a reduction in the inflammatory response chloride salt – CMA) into PC polymers also significantly and fibrous capsule formation [7]. Such properties have increased the amount of protein adsorption and subsequent resulted in PC materials being used for a variety of bio- cell attachment, where the amount of protein adsorption and medical applications where a passive interaction between cell attachment was found to be protein, cell and CMA the material and the body is required. Examples include monomer content specific. The authors [32] concluded that coating blood contact devices, such as coronary guide wires the increase in PC bioactivity was caused by the increase in [8] and stents [9], extracorporeal circuits [10], and for the polymer surface charge, whereby the predominantly contact lenses [11]. PC materials have also found utility in negatively charged biomolecules interact electrostatically the orthopaedic field, as superlubricious, low-wear surfaces with the positively charged surfaces. Moreover, Palmer et al grafted onto polyethylene acetabular liners [12], for which [33] have shown that cationically- modified PC materials there are now 3 year data from an 80 patient study [13], can also act as drug delivery vehicles, where negatively colbalt-chromium-molybdenum metal alloy bearings [14, charged therapeutic biomolecules interact reversibly with 15] and poly(ether-ether ketone) orthopaedic bearing sur- PC’s positive charge and are released in a controlled man- faces [16]. ner. In a comparative study of PC and cationically-modified However, in some clinical applications a “bioinert” PC coatings on titanium porous oxide surface implants response is not the most appropriate. For example, ortho- placed in trabecular and cortical bone of rabbits [34], paedic implants, such as artificial hips, require a degree of although there was no significant differences in bone den- interaction between the femoral stem and the surrounding sity between the coatings or uncoated control, there was a bone tissue in order to enhance load transfer and to decrease higher bone-implant contact for the cationically-charged PC micromotion and stress shielding of the implant, which can coating and uncoated control compared to the PC coating lead to implant loosening and premature failure [17, 18]. (p < 0.05) at the 6 week explant point. This somewhat Furthermore, increasing the speed and quantity of bone- indicates that the surface charge is able to orchestrate cel- implant integration allows early loading, resulting in lular interaction, overcoming the inherent anti-adhesiveness quicker patient mobility, improved patient well-being, of the PC material, but only histometric and biomechanical shorter hospital stays, and a reduction in healthcare costs analysis was performed and hence no mechanism proposed. [1]. An important development in this regard was plasma- The aim of current work presented herein was to investigate sprayed hydroxyapatite coatings which produce improved the effects of high and low charge levels of cationically- bone bonding of implants to bone and increased longevity charged PC materials on human osteoblast cells (HObs). of some joint replacements [19, 20]. However, the risks Osteoblasts are skeletal cells responsible for bone formation associated with these relatively thick coatings is delamina- [35]. They are derived from messenchymal stem cells found tion and there is continued concern about the resorption of in the bone marrow [36] via osteoprogenitor cells and they HA by interfacial remodelling in some applications [21– secrete extracellular matrix (type I collagen) and non- 23]. The phosphorylcholine side chain of the PC-polymer collagenous material that is involved in the nucleation and contains the phosphate moiety present in phosphatidylcho- formation of bone mineral [37]. Hence the cells are likely line and phosphatidylserine, the latter being membrane lipid being influenced either directly or indirectly by the implant component that is able to bind calcium and is implicated in surface characteristics. We describe in vitro experiments in J Mater Sci: Mater Med (2017) 28:144 Page 3 of 13 144 order to shed light on potential mechanisms of interaction 2 h at room temperature, the polymers were cured onto the between biomaterial and tissue and investigate the in vivo plates and flasks at 70 °C for 72 h. The coated culture dishes longevity of any beneficial effect in a 14 week rat tibia were then sterilised under UV light for 2 h. implantation model. 2.1.2 Cell culture 2 Materials and methods Human osteoblasts (HObs) (PromoCell, Heidelberg, Ger- many) isolated from hip bone were grown in 75 cm tissue 2.1 In vitro experiments culture flasks in McCoy’s 5a growth medium (Invitrogen, Paisley, UK) containing 10% foetal bovine serum (FBS), 2.1.1 Materials 1% glutamine and 30 μg/ml vitamin C (standard culture conditions) (Sigma, Poole, Dorset, UK). After cryogenic The PC polymers used in this experiment have previously recovery and 24 h incubation, cell adherence was checked been characterised [38]. They are methacrylate-based and the medium replaced; thereafter the medium was polymers synthesised using free radical polymerisation of replaced every 2 days. At 60–80% confluence, the cells 2-methacryloxyethylphosphorylcholine (MPC), lauryl- were washed three times with Hank’s Balanced Salt Solu- methacrylate (LMA), hydroxypropylmethacrylate (HPMA) tion (HBSS) (Invitrogen) and incubated (RT for 5 min) and trimethoxysilylmethacrylate (TMSMA). Positive before trypsin/ethylenediaminetetraacetic acid (EDTA) charge was introduced by including choline methacrylate (Sigma) was used to detach the cells from the flask. Com- (CMA). A schematic representation of this polymer is plete cellular detachment was confirmed by phase-contrast shown in Fig. 1 and the constituents (weight percentages) microscopy. After 5 min, McCoy’s growth medium was and their function are detailed in Table 1. In addition to the added to neutralise the trypsin/EDTA. The cell suspension other polymers, PC5 was used in the in vitro experiments was then centrifuged at 220 × g for 4 min at room tem- and PC6 was used to coat the in vivo implants due to the perature before the supernatant was aspirated and the cell availability of these two substrates when these two experi- pellet re-suspended in 1 mL of McCoy’s medium and ments were carried out. The difference in response to the counted using a haemocytometer. small difference in cationic charge between these two PC polymers is unlikely to be significant. The polymers were 2.1.3 Cell number supplied by Biocompatibles UK Ltd (Farnham, Surrey, UK). HOb cells were cultured on non-coated, PC5 and PC20 Each PC polymer was dissolved in ethanol at a con- coated T25 flasks for 28 days in medium (as previously centration of 10 mg/ml and the solution filtered through a described), incubated at 37 °C in 5% CO humidity. Cell 0.2 μm filter. Tissue culture plates and flasks were then number was determined at 6 h, 1, 2, 4, 7, 14 and 28 days TM filled with the polymer-ethanol solution and left for 5 min at using Vialight HS proliferation/cytotoxicity kit (Lonza, room temperature before being emptied. After air-drying for UK). In this assay, luciferase enzyme reacts with adenosine trisphosphate (ATP) (found in all metabolically active cells) to emit light and its intensity is indicative of cell number. At each time point the culture media was aspirated and the cells washed 3 times with phosphate buffer saline solution (PBS) (Sigma). For each flask 1 mL 0.5% v/v Triton X-100 (VWR, Lutterworth, UK) in PBS was added and two freeze thaw cycles (15 min at −70 °C and 37 °C) were used to lyse the cell suspension. 180 μL of the lysate was then trans- ferred in duplicate to a white-walled 96-well plate and 20 μL of ATP monitoring reagent added to each well. The ATP concentration was then immediately read (TopCount. TM NXT , Packard BioScience Co, Meriden, USA). Using a standard curve (light intensities of pre-determined HOb cell numbers of 0, 2500, 5000, 10000, 20000 and 40000 on uncoated well plates) the cell numbers on the substrates was determined. Cell number was also determined for cells on PC0 at 6 h, 1 and 2 days. Three separate flasks were used Fig. 1 A schematic chemical representation of a cationically charged phosphorylcholine polymer for each substrate and time point. 144 Page 4 of 13 J Mater Sci: Mater Med (2017) 28:144 Table 1 Constituents of the Monomer content (wt%) phosphorylcholine materials (weight percent) Monomer PC20 PC6 PC5 PC0 Component function MPC v 22 29 28 29 Phosphorylcholine head group—charge neutrality—non- thrombogenic LMA w 41 50 50 51 Lauryl group—alkyl chain component—hydrophobic—substrate adsorption HPMA x 12 12 12 15 Aids crosslinking and film formation TSMA y 5 4 5 5 Methoxysilyl crosslinker—mechanical stability. Aids substrate adhesion TMA z 20 6 5 0 Cationic charge 2.1.4 Cell differentiation characterise the elemental constituents of mineral exudates on the cells and ECM (extracellular matrix). HOb cells were cultured on non-coated, PC5 and PC20 coated T25 flasks for 28 days in McCoy’s medium. Cell 2.2 In vivo experiment differentiation was determined by identifying the enzyme alkaline phosphatase (ALP), a widely recognised [39] 2.2.1 Materials enzyme marker of osteoblast differentiation associated with skeletal mineralisation [40]. 0.5 M 2-amino-2-methyl-1- Thirty surgical grade stainless steel (316 L) pins (R J Lay- propanol (AMP) substrate buffer (Sigma) was prepared in land, Rayleigh, UK) were cleaned using dry Emory cloth distilled water (pH 10) and supplemented with 2 mM paper (P1200, T444 Norton, Tufbak Durite) and sonicated magnesium chloride (MgCl ) and 9 mM p-nitrophenol for 5 min in both acetone and ethanol. The pins were phosphate (p-NPP) (Sigma). In an alkaline solution, ALP equally assigned to one of five groups and coated accord- catalyses p-nitrophenyl phosphate to p-nitrophenol, ingly. Six pins were uncoated (negative control). Six pins appearing yellow in colour. Cells were lysed at 1, 2, 4, 7, 14 were plasma-spray coated with 60–100 μm of hydro- and 28 days in 3 flasks per substrate per timepoint (as xyapatite (Plasma Biotal, Tideswell, UK) (positive control). described for ATP measurement) and 50 μL of each lysate Three groups of six were then double dip coated at 4 mm/ in duplicate and 50 μL of a range of p-nitrophenol con- sec in 5 mg/ml polymer concentrations in ethanol of PC0, centrations, diluted to produce a standard curve were PC6 or PC20 respectively (as previously described in Table transferred to a 96-well plate. Subsequently, 50 μLofAMP 1), resulting in polymer coatings approximately 30–50 nm buffer was added to each well and the plate incubated at thick. The polymer coated pins were then cured at 70 °C for 37 °C for 15 min. Absorbance was immediately read at 405 4 h. All pins were designed to have equal dimensions prior nm using an EL800 Universal Microplate reader (Bio-tek to implantation (Fig. 2a) and were sterilised using gamma Instruments, Inc., Winooski, USA). A standard curve of irradiation (25 kGy, Isotron, UK). absorbance as function of p-nitrophenol concentration was generated and used to determine the total ALP content of 2.2.2 Surgery each test sample. Optical Imaging and Energy Dispersive X-ray (EDX) The surgical method for implantation has previously been Characterisation of Cells HOb cells were cultured (seeding described by Allen et al [41] and others [42, 43]. All pro- 3 2 density of 12 × 10 /cm ) in T25 flasks coated with PC0, cedures underwent ethical review and were carried out in PC5 or PC20 and on a non-coated tissue culture plastic accordance with the regulations as set out in the Animal control for 48 h and 28 days. Optical imaging using a Leitz (Scientific Procedures) Act 1986. Thirty mature Sprague Labovert phase contrast microscope was carried out on all Dawley rats, weighing between 300–350 g (Harlam UK substrates at 48 h. For SEM, the cells on the substrates at 48 Ltd, Bicester, UK) were randomly allocated to one of the h and 28 days, were washed with PBS, fixed in 4% paraf- five groups. The number of animals used per group was ormaldehyde (Sigma) and washed three times with distilled based upon previous studies which showed significant water. A square piece, approximately 20 × 20 mm, of each results with groups of 6. One pin was implanted into the flask was cut and mounted onto an SEM stub using double- right tibia of each animal (Fig. 2b). sided carbon tape and sputter coated with carbon. A JEOL All drugs were obtained from National Veterinary Ser- XL30 scanning electron microscope (SEM) at 5 kV vices, Stoke-on-Trent, UK. Briefly, anaesthesisa was was used to image the samples and EDX was used to induced using a gaseous mixture of oxygen and halothane J Mater Sci: Mater Med (2017) 28:144 Page 5 of 13 144 Fig. 2 a Schematic representation of the stainless steel pins used for implantation; b a radiograph indicating the placement of the surgical pin in the rat tibia; c diagrammatic map of a laterel sectioned pin. The figure indicates the locations where images for histological analysis were captured (1–5) and the local distinction between the interface and interfacial area (4%), at a rate of 6 litres/min in an anaesthetic chamber. de-fatted using continuously topped-up acetone under Midazolam (3 mg/kg) was given by intra-peritoneal injec- vacuum for 7 days. The tibias were then embedded in poly tion and the rats maintained with halothane 4% delivered (methyl methacrylate) (PMMA). Methyl methacrylate through an anaesthetic mask. The right hind limb was (Sigma) containing 2.5% w/v benzoyl peroxide (catalyst) shaved, peri-operative analgesia administered sub- and 2.5% v/v dibutyl phthalate (plasticiser) was infiltrated cutaniously (Rimadyl—carprofen, 5 mg/ml) and the rat into the specimens at 4 °C under vaccuum for 7 days. The placed supine. A 1.5 cm incision was made lateral to the methacrylate solution was then replaced with fresh solution patella and a lateral capsulotomy performed allowing and polymerised. Once polymerised, 300–500 μm thick medial dislocation of the patella exposing the tibial plateau. longtitudinal sections (sagital plane of the pin and tibia) A 1.5 mm diameter hand drill was used to drill through the were cut using a diamond saw (100 CA blade 0.1 mm/sec, centre of the tibial plateau to a depth of 10 mm and counter Accuton 5, Struers, Glasgow, UK) and glued (Cyno 40, sunk (using a 4 mm counter-bore) to allow the head of the Delta, Leeds, UK) onto frosted glass microscpe slides implant to be flush with articulating surfaces. The pin was before being ground and polished to 100 μm thick using a then press-fit into the tibia, the patella reduced and the graded series of silicon carbide papers (Struers, Denmark). incision closed. Immediately after surgery the rats received 150 mg/kg of antibiotic (Synulox—coamoxyclavulanic 2.2.4 Microscopy and histological analysis acid) intramuscularly and a sub-cutaneous injection of 0.15 mg/kg analgesia (temgesic: buprenorphine). The rats were Slides were washed in liquid detergent and water for 1 min, allowed to recover in IVC caging systems (Techniplast UK rinsed in water for 3 min and stained with toluidine blue Ltd, Northants, UK) and after 5 h a second injection of (pH 9, 56 °C, 30 min). Five fields (Fig. 2c) for each slide temgesic (0.1 mg/kg) was administered. were digitally photographed using a Leitz Dialux 20 light microscope (X100) with attached camera and stored on a 2.2.3 Tissue processing computer (Aquis Image Acquisition software—Synoptics, Cambridge, UK). The rats were sacrificed at 14 weeks according to Home The head of the pin was ignored due to relatively poor Office (UK) schedule 1. The right tibia was removed and bone apposition in this region. Such a response may be due fixed in ice-cold paraformaldehyde in 0.1 M phosphate to stress shielding (poor stress distribution through the pin) buffer (pH 7.4) with 0.1% w/v sucrose and 0.05% v/v resulting in bone resorption rather than due to the problems gluteraldehyde at 4 °C. After 5 h the tibias were washed associated with surgery and experimental technique and has three times in PBS and dehydrated in ascending con- been observed previously [33]. The image was then edited centrations of methylated spirit (50–100%) before being in Adobe Photoshop 6.0 and cropped to allow a distance of 144 Page 6 of 13 J Mater Sci: Mater Med (2017) 28:144 50 μm from the pin surface into the surrounding bone and tissue plastic control. Interestingly, the number of cells bone marrow leaving an image of the bone-implant inter- decreased on PC5 with increasing time, up to 28 days. After face (Figs. 2c and 8). Four tissue types, bone, marrow, day 1, significant differences (P < 0.001) existed between fibrous tissue and loosely associated matrix (LAM), were PC5 and PC20 and control surfaces for all time points. This then identified and colour coded. The coloured images were indicates that HOb cells proliferated well on PC20, how- then processed into binary images. Scion Image Analysis ever, proliferation on control surfaces was significantly (Scion Corporation, USA) was used to measure the areas of greater at all time points (P < 0.001). the four tissue types and these were expressed as a per- Figure 4b shows the effect of time and charged PC centage of the total area for each field position on each substrate on cell differentiation. The ALP activity per individual pin (interfacial area). The percentage of each thousand cells decreased on PC20 over the first 3 days of tissue was also calculated along the length of the interface culture and on the control substrate between day 1 and 2. between the pin and the bone (interface). The five positions Thereafter ALP activity did not change significantly and along the pin were then summed and averaged. there was no significant difference between ALP activity on PC20 and control substrates after day 2. Alkaline phos- phatase activity and the number of cells on PC5 was too low 2.2.5 Statistical analysis to give an accurate measure of normalised activity For both the in vivo and in vitro experiments the data was compared using a one-way ANOVA. Identified trends were further investigated using either specific t-testing or Bon- feronii analysis to highlight significant differences. The distribution of the data is represented using standard error. 3 Results 3.1 In vitro experiments Figure 3 indicates the number of HOb cells on PC coated and control surfaces over 48 h. At each time point, tissue plastic control surfaces exhibited the largest number of cells. Significant differences (P < 0.001) were observed between the control and PC0 and between PC20 and PC0 surfaces at all three time points. However, significant differences (P < 0.001) between PC20 and PC5 were only seen at 24 and 48 h. Figure 4a indicates the number of HOb cells on charge- modified PC coated and control surfaces over 28 days. The number of HOb cells increased with time on PC20 and Fig. 4 a The effect of cationically-modified PC on the number of HOb cells in vitro over a 28-day period in culture as compared to control Fig. 3 The effect of PC and cationically-modified PC on the number surfaces. The number of cells on control surfaces was significantly of HOb cells in vitro after 6, 24 and 48 h in culture as compared to greater (p < 0.001) than on PC20 and PC5 at all time points and on control surfaces. *p < 0.001 compared to other materials at the PC20 than PC5 after day1; b corresponding effect of cationically- equivalent time point. +p < 0.001 compared to PC20 and control at modified PC on HOb cell ALP production in vitro over a 28-day the equivalent time points period as compared to control surfaces J Mater Sci: Mater Med (2017) 28:144 Page 7 of 13 144 Fig. 5 Photomicrographs of HOb Cells on a PC0, b PC5, c PC20 and d control surfaces at 48 h in vitro. (Scale bars = 100 μm) Figure 4a shows the effect of PC charge content on HOb tissue than both bone and marrow and LAM. For SS there cell proliferation and spreading at 48 h and supports the data was also significantly more bone and marrow than LAM. presented in Fig. 3. PC20 and the control surfaces showed HA had the highest bone and marrow apposition compared the largest number of cells, and on these surfaces the HOb to the other four surfaces and also less fibrous tissue (p < cells have developed fully-spread morphologies (Fig. 5). 0.05). In a comparison of the PC materials, PC0 had sig- PC0 supported minimal cell adhesion and, although there nificantly (p < 0.05) more fibrous tissue than PC6 but there were more cells on PC5, the cells on both these substrates were no significant differences between PC0 and PC20 or remained rounded. The photomicrographs in Fig. 5 also between PC0 and SS. Furthermore, there was no significant indicate that the cells on tissue plastic controls and on difference between any of the PC groups in relation to the PC20 showed the presence of mineral deposits. formation of LAM. Figure 6 shows SEM with elemental analysis and sup- Figure 7b indicates the amount (percentage) of the dif- ports Fig. 5, confirming both the presence of mineral ferent tissue types in the bone-implant interfacial area deposits on PC20 and control surfaces at 48 h and that the defined in Fig. 2c. The data complements that described in number of these mineral deposits increased with time (up to Fig. 7a, however, in this instance bone and marrow, sig- 28 days). Figure 6 also shows that the mineral deposits on nificantly (p < 0.05) dominate over fibrous tissue and LAM both PC20 and control surfaces at 28 days contained both formation for all surface coatings. For PC6, PC0 and SS calcium and phosphorus. there was significantly (p < 0.05) more fibrous tissue than LAM. In a similar manner to the implant interface, Fig. 7b 3.2 In vivo results indicates that HA had, significantly (p < 0.05), the most bone and marrow apposition and the least fibrous tissue formation in comparison to the other four surfaces. PC0 had Figure 7a indicates the amount (percentage) of different types of tissue at the bone-implant interface. For HA there significantly (p < 0.05) more fibrous tissue than PC6. Analysis of these data shows that there is a trend for a linear was significantly (p < 0.05) more bone and marrow than both fibrous tissue and LAM. Both cationically charged PC increase in LAM with increasing CMA content (R = 0.9) materials produced a similar response, with no significant in the implant interfacial area; a weak linear dependency of differences found between any of the tissue types. For both LAM with CMA content was also found at the implant PC0 and SS there was significantly (p < 0.05) more fibrous interface (R = 0.7). Typical histological micrographs of the 144 Page 8 of 13 J Mater Sci: Mater Med (2017) 28:144 Fig. 6 SEM images showing osteoblast cells on a PC20 and b control at 48 h and c PC20 and d control at 28 days. An increase in the amount of mineral deposits (MD) were seen on both PC20 (c) and control (d) surfaces at 28 days. Corresponding elemental analysis is shown in the inset for each image and demonstrates that these mineral deposits contain calcium and phosphorus bone-implant interface are shown in Fig. 8. These show a thin layer of fibrous tissue between the implant and bone (Fig. 8a–c) with the exception of HA where the bone is directly apposed to the ceramic (Fig. 8d). The LAM is primarily seen adjacent to the PC coated implants (Fig. 8b, c). 4 Discussion 4.1 In vitro response It has been well documented that non-charge modified phosphorylcholine surfaces (PC0) significantly reduce cel- lular adhesion compared to tissue culture plastic controls [7, 9, 32]. Within 1 min of implantation most biomaterial sur- faces are coated with a protein film [44], which in an in vivo environment can comprise over 200 different proteins [45]. Proteins adsorb onto surfaces to minimise surface free energy. This can occur through protein conformational Fig. 7 The effect of PC and cationically-modified PC on the percen- change, partial dehydration of the protein and surface, and tage of bone and marrow, fibrous tissue and loosely associated matrix the redistribution of surface charges [46]. Subsequent cel- (LAM) compared to hydroxyapatite (HA) and stainless steel (SS) lular interactions are largely governed by the nature and positive and negative controls at a the bone-implant interface. * Sig- conformation of these adsorbed proteins whereby the nificantly more (p < 0.05) than the other two tissues on that material. +Signficantly different amounts of tissue on HA (p < 0.05) than the adsorbed proteins can act as ligands for cell-surface recep- same tissue on the other 4 materials. b In the bone-implant interfacial tors [47]. PC0 surfaces interact with proteins without area inducing conformational shape changes in their structure J Mater Sci: Mater Med (2017) 28:144 Page 9 of 13 144 Fig. 8 Typical experimental micrographs indicating different tissue types in the bone-implant interfacial area, where M is marrow, B is bone, FT is fibrous tissue and LAM is loosely associated matrix, for a stainless steel; b, c PC20; d hydroxyapatite (HA) resulting in a decrease in protein adsorption. It is thought Figure 4a shows that HOb cells attached to PC5 after 24 that their highly hydrophilic phosphorylcholine zwitterionic h are in a non-proliferate state. This is in stark contrast to head group (MPC monomer, Fig. 1) results in the formation the number of cells proliferating on both tissue culture of a hydration layer that allows proteins to interact with the plastic and PC20 up to 28 days. Generally, initial cell surface reversibly whereby the hydrated layer limits the adhesion occurs in the first 30–120 min of cell-surface dehydration step involved in adsorption [48]. This zwitter- contact followed by a stronger attachment involving the ionic group may also decrease protein adsorption through secretion and development of ECM, cellular migration and charge neutrality. The subsequent decrease in protein growth [50]. The larger number of proliferating osteoblasts, adsorption significantly reduces cellular adhesion, including at 48 h, on PC20 and the control in comparison to PC5 can monocytes, macrophages, fibroblasts and human granulo- also be seen visually in Fig. 5. This suggests that there is cytes [49]. Such a response was also observed in HOb cells only a weak interaction formed between the PC5 surface (Fig. 3), whereby there are significantly (p < 0.001) less and the osteoblast cells, reducing stimulation of the cells to HOb cells on PC0 at 6, 24 and 48 h compared to PC5, PC20 produce ECM. The lack of ECM production is likely to and tissue culture plastic. The lack of cell adhesion to PC0 cause the subsequent loss of cells by apoptosis, which is can also be seen visually in Fig. 5. seen at the later time points (Fig. 4a). The weaker interac- Figure 3 indicates that over early time periods (up to 48 h) tion between PC5 and the osteoblast cells is likely to relate increasing CMA content increases the number of adherent to the charge sensitive adsorption of proteins. For example, HOB cells. This is in agreement with the findings of Rose PC5 may not contain enough positive charge to adsorb et al [32] who showed that increasing CMA content in PC vitronectin or fibronectin, adhesive glycoproteins that are polymers increased cell adherence. The increase in cell known to play a key role in the anchorage of osteoblasts attachment is likely to be caused by either the non-specific [50]. electrostatic interaction of the predominantly negative cells The results for ALP production by the cells shown in and the positively charged surface or through an increase in Fig. 4b have been adjusted for the change in cell number in the number of cellular adhesive proteins adsorbed onto the the cultures shown in Fig. 4a. ALP is an early marker of surface of the PC polymer caused by the increase in CMA osteoblast differentiation [51] and is only up-regulated by content. It is possible that both these mechanisms are run- osteoblast cells after proliferation has finished and prior to ning concurrently. Rose et al [32] also showed that the mineralisation of the ECM. Once mineralisation starts, ALP inclusion of CMA content in PC polymers resulted in an production ceases [52]. On PC20 and control substrates the increase in protein adsorption compared to their non- osteoblasts continue proliferating throughout the culture charged counterparts and that specific proteins have a period, as seen in Fig. 4a, and do not reach confluence, minimal CMA or charge content requirement. Such which would enable the onset of differentiation throughout dependencies may help explain the results observed in Fig. the culture and increase ALP production. ALP production 4a that indicate the effect of PC cationic charge on HOb cell per thousand cells decreases over the first 2 days in these proliferation up to 28 days. cultures as the newly divided cells are not producing ALP. 144 Page 10 of 13 J Mater Sci: Mater Med (2017) 28:144 Thereafter it is likely that groups of cells are showing PC0 show significantly (p < 0.05) more fibrous tissue than increased ALP but not the proliferating cells in the culture; any other tissue type (Fig. 7). This is at odds with work by thus, the ALP activity per thousand cells does not show a Goreish et al. [7] who showed that intramuscular PC0 marked change. On PC5 the ALP production remains implants in the rabbit were surrounded by 40% less constant over the initial 2 days as cell number does not inflammatory cells than polyethylene controls, concluding increase after which the very low cell number and ALP that this was due to PC0’s well hydrated surface and passive production prevent an accurate measure of ALP production interaction with the surrounding tissue. Implant location per thousand cells. (bone is rich in pre-matrix producing cells that can differ- Evidence of mineral formation on PC20 is supported by entiate into fibroblasts), the nature of the controls (poly- Fig. 5, which shows that on both PC20 and tissue plastic ethylene is known to elicit a high fibrous response and control surfaces HOb cells have developed fully spread stainless steel (SS) 316 is reasonably “bioinert” in an oss- morphologies that contain deposits of calcium phosphate eous environment), as well as the different animal studies (Fig. 6), which increased with time, up to 28 days. How- used may explain the observed differences. ever, the EDX spectra (insets of Fig. 6a-d) showed that the The reasons for the effect of PC CMA content on bone calcium phosphate (Ca:P) ratio of both PC20 and control and marrow apposition is also debatable. Cationic charge was approximately 0.6, considerably lower than hydro- decreases fibrous tissue formation with a consequential xyapatite (the mineral phase of bone) which has a Ca:P ratio increase in bone and marrow apposition suggesting that of between 1.5 and 1.67 [53]. A lower Ca:P ratio is indi- increasing surface charge increases osseous-implant inte- cative of calcium deficient mineral. Such a calcium deficient gration (Fig. 7a). However, this increase is not linear with HA maybe the due to the presence of the cationic surface increasing CMA content and both charged polymers (PC6 2+ charge of the PC that may be preventing Ca deposition and PC20) show similar amounts of both bone and marrow into the mineral. and fibrous tissue (Fig. 7). Furthermore, no significant dif- ference in bone and marrow and fibrous tissue exist between 4.2 In vivo response the negative control (SS) and the two charged PC coatings. One might expect that increasing surface charge would Figure 7a, b shows the amount (percentage) of different increase fibrous tissue formation, due to surface charge tissue types at the bone-implant interface and interfacial aggravates the foreign body response. However, Rose et al. area respectively. The results indicate that the HA coating [32] showed that although the presence of surface charge in PC increases the number of attached fibroblast cells, unlike creates excellent bone apposition against the implant, showing significantly (p < 0.05) more bone and marrow other cells (eg monocytes and granulocytes), the number of implant apposition compared with the other four substrates. fibroblasts on all charged materials PC materials with up to The degree of this apposition can be seen in Fig. 8d and is 30% CMA was maintained at a constant level. in agreement with the literature [54, 55]. Hydroxyapatite An explanation for the decrease in both bone and marrow (Ca (PO )OH) based coatings create an interactive bond apposition and fibrous tissue formation with increasing PC 5 4 between the bone and implant that is characterised by the CMA content is the increase in LAM with increasing sur- presence of a thin calcium phosphate (apatite) layer, which face charge (Fig. 7a). LAM is shown visually in Fig. 8c, d forms early in the implantation process. Kokubo et al. [56] and can be characterised as a fluid-containing capsule argue that this intermediate apatite layer allows for the comprised of a few cells and loosely aggregated matrix preferential proliferation and differentiation of osteoblasts surrounded by a fibrous membrane. No reference in the on its surface, stimulating the formation of new bone. literature exists as to the formation of such tissue at the In contrast, the phosphorylcholine and the stainless steel implant-bone interface, though such tissue may be descri- implants are apposed with fibrous tissue and LAM (Figs. 7 bed as lesions; zones of tissue which have impaired function and 8a–c). Fibrous encapsulation of so-called “bio-tolerant” as a result of damage or disease [59]. The correlation (R )of polymeric and metallic implants at the bone-implant inter- increasing LAM formation with CMA content increases face has been well documented and is a consequence of the from 0.7 to 0.9 when examining the interfacial area (Fig. foreign-body reaction [57, 58]. After the immediate protein 7b) of the implant as apposed to the interface (Fig. 7b) and adsorption that follows implantation, local tissue damage is due to LAM extending from the surface of the implant causes the recruitment of white blood cells, including and occupying space (Fig. 8b, c). LAM formation maybe macrophages, which subsequently bind to the surface of the caused by the preferential attachment of inflammatory cells implant. However, these phagocytic cells cannot ingest the such as monocytes that may become over stimulated in the implant and fuse to form giant cells, recruiting fibroblasts, production of tumour necrosis factor (TNF) -α, a chemo- which subsequently secrete collagen matrix resulting in the tactic pro-inflammatory cytokine that can cause the forma- fibrous encapsulation of the implant [45]. Both SS and tion of necrotic tissue [60]. In addition, the increase in J Mater Sci: Mater Med (2017) 28:144 Page 11 of 13 144 charge may also be acting on cells in such a way that it post-implantation to allow for a period of fuller integration results in local cell apoptosis. with the implant, as previous work had seen some differ- Rose et al. [32] has shown that a variety of different cells ences at just 6 weeks post implant [34]. In fact, as bone and proteins are affected by the inclusion of positive surface ingrowth appears to occur at an earlier stage, our model may charges into PC polymers. However, it is not clear how have suffered from the potential for newly-synthesised bone selective this effect is in a multi-cellular in vivo environment. to undergo subsequent bone resorption due to mechanical For example, fibroblast cells may preferentially attach to the stress or excessive inflammation. PC surface over osteoblast cells. Furthermore, both osteo- blasts and fibroblasts share the same precursors in the bone marrow and these surfaces may be inducing fibroblast dif- 5 Conclusions ferentiation over osteoblasts, resulting in a fibro-encapsulated sheath. In addition, pro-inflammatory cells may preferentially Modifying phosphorylcholine surfaces by the addition of bind to the charged PC surfaces and induce cell death and the cationic charge alters the response of cells to these surfaces production of LAM [27]. A further experiment, which and could allow the delivery of drugs, including antibiotics, examines the effects of PC and modified PC surfaces in a from PC implant coatings. However, our studies show that multicellular environment, would be an interesting way to whilst increasing positive surface charge allowed increased assess which cells are preferentially attaching to the polymer osteoblast adhesion to the PC surface and stimulated surfaces. Alternatively, LAM maybe caused by continuous increased cell differentiation and the production of calcium micromotion of the implant due to the failure of early bone phosphate deposits in vitro, when implanted in bone charge growth at the bone-implant interface, however this seems modified PC materials stimulated the production of fibrous unlikely as such a response might be expected equally with tissue and areas of loosely associated matrix around the PC0, PC6 and SS implants; a trend that was not observed in implant. Despite this, non-charged phosphorylcholine the data. The in vivo results for hydroxy-apatite coated polymers were tolerated in the osseous environment and implants clearly accords with the clinical utility of this could be used to coat orthopaedic devices, where bone material; however, the relatively poor outcome for PC0, bonding is not required, for the delivery of antibiotics in PC6 and PC20 indicate that these materials would not be order to reduce surgical site infection. suitable as coatings for bone bonding applications. Acknowledgements The authors would specifically like to thank Phosphorylcholine-based coatings may however show Biocompatibles UK Ltd for supplying the polymers and for financial greater stability in vivo and could be useful as drug delivery and academic support, without which this research would not be vehicles in situations where bone bonding is not required or possible. Financial assistance from the Engineering and Physical be used where inhibiting tissue bonding would be useful for Science Research Council and Biotechnology and Biological Sciences Research Council UK is also acknowledged. example with fracture fixation plates. We do not know whether PC coating was lost from the surface of the pins due Compliance with ethical standards to mechanical damage during insertion. This could be investigated in cadaveric models by inserting coated pins into Conflict of interest The authors declare that they have no competing the tibia, breaking open the bone and determining the uni- interests. formity of coating on the retrieved pins. The results presented in this paper describe the response of bone cells and bone to PC and modified PC surfaces and also demonstrate that in vitro experiments are not always References accurate predictors of the more complex in vivo response. It 1. Hench LL. Biomaterials: a forecast for the future. Biomaterials. was hypothesised that due to the increased proliferation and 1998;19:1419–23. differentiation of HOB cells on PC20, and the evidence that 2. Vander Sloten J. Materials selection and design for orthopaedic these cells were undergoing mineralisation (Figs. 3–6), that implants with improved long-term performance. Biomaterials. PC20 coated implants would stimulate an increase in bone- 1998;19:1455–9. 3. Hench LL. The challenge of orthopaedic materials. Curr Orthop. bonding compared with the lower and non-charged PC 2000;14:7–15. materials, whereby osteoblasts are actively attracted to the 4. Rose FRAJ, Oreffo ROC. Bone tissue engineering: hope v’s hype. surface of the polymer and mineralise to produce a tight Biochem Biophys Res Com. 2002;292(1):1–7. implant-bone interface. However, in this limited scope 5. Ishihara K, Nomura H, Mihara T, Kurita K, Iwasaki Y, Naka- bayashi N. Why do phospholipids polymers reduce platelet in vivo experiment, it was shown that this was not the case. adsorption? J Biomed Mater Res. 1998;39:323–30. Ideally, a full time-course experiment would be needed to 6. Iwasaki Y, Mikami A, Kuria K, Yui N, Ishihara K, Nakabayashi properly address the evolution of the biological response; N. Reduction of surface-induced platelet activation on phospho- here we were limited to one time point and chose 14 weeks lipid polymers. J Biomed Mater Res. 1997;36(4):508–15. 144 Page 12 of 13 J Mater Sci: Mater Med (2017) 28:144 7. Goreish HH, Lewis AL, Rose S, Lloyd AW. The effect of 23. Singh G, Reichard T, Hameister R, Awiszus F, Schenk K, phosphorylcholine-coated materials on the inflammatory response Feuerstein B, Roessner A, Lohmann C. Ballooning osteolysis in and fibrous capsule formation: in vitro and in vivo observations. J 71 failed total ankle arthroplastie: is hydroxyapatite a risk factor? Biomed Mater Res. 2004;68(A):1–9. Acta Orth. 2016;87(4):401–5. 8. Gobeil F, Juneau C, Plante S. Thrombus formation on guide wires 24. Voelker DR. Bridging gaps in phospholipid transport. Trends during routine PTCA procedures. A scanning electron micro- Biochem Sci. 2005;30(7):396–404. scopic evaluation. Can J Cardiol. 2002;18:263–9. 25. Wu LNY, Genge BR, Wuthier RE. Analysis and molecular 9. Whelan DM, van der Giessen WJ, Krabbendam SC, van Vilet EA, modelling of the formation, structure and activity of the Verdouw OD, Serruys PW, van Beuekom HMM, Biocompat- phosphatidylserine-calcium-phosphate complex associated with ibility of phosphorylcholine coated stents in normal porcine cor- biominerlization. J Biol Chem. 2008;283(7):3827–38. onary arteries.Heart. 2000;83(3):338–45 26. Merolli A, Santin M. Role of phosphatidyl-serine in bone repair 10. De Somer F, van Belleghem Y, Caes F, Francois K, Arnout J, and its technological exploitation. Molecules. 2009;14:5367–81. Bossuyt X, Taeymans Y, van Nooten G. Phosphorylcholine 27. Gong YK, Luo L, Petit A, Zukor DJ, Huk OL, Antoniou J, coatings offers natural platelet preservation during cardio- Winnik FM, Mwale F. Adhesion of human U937 macrophages to pulmonary bypass. Perfusion. 2002;1:39–44. phosphorylcholine-coated surfaces. J Biomed Mater Res A. 11. Lemp MA, Caffery B, Lebow K, Lembach R, Park J, Foulks G, 2005;71(1):1–9. Hall B, Bowers RJW, McGarvey S, Young G, Omafilcon A. 28. Shelton RM. Protein adsorption at the interface between charged (Proclear) soft contact lenses in dry eye population. CLAO J. polymer substrata and migrating osteoblasts. Biomaterials. 1999;25(1):40–7. 1988;9:24–9. 12. Kyomoto M, Moro T, Konno T, Takadama H, Yamawaki N, 29. Bet MR. Cell adhesion and cytoxity studies over polyanionic Takatori Y, Nakamura K, Ishihara K. Enhanced wear resistance of collagen surfaces with variable charge and wettability. Bioma- modified cross-linked polyethylene by grafting with poly (2- terials. 2003;24:131–7. methacryloyloxyethyl phosphorylcholine). J Biomed Mater Res 30. Lee JH, Khang SS, Lee JW, Lee HB. Interaction of proteins and A. 2007;82(1):10–7. cells on functional group gradient surfaces. Molec Symp. 13. Takatori Y, Moro T, Ishihara K, Kamogawa M, ODa H, 1997;118:571–6. Umeyama T, Kim YT, Kyomoto M, Tanaka T, Kawaguchi H, 31. Kishida A. Cell behaviour on polymeric substances with non-ionic Tanaka S. Clinical and radiographic outcomes of total hip repla- and ionic monomers. Biomaterials. 1991;12:786–92. cement with poly(2-methacryloyloxyethyl phosphorylcholine)- 32. Rose SF, Lewis AL, Hanlon GW, Lloyd AW. Biological grafted highly cross-linked polyethylene liners: three-year results responses to cationically charged phosphorylcholine-based mate- of a prospective consecutive series. Mod Rheumatol. 2015;25 rials in vitro. Biomaterials. 2004;25:5125–35. (2):286–91. 33. Palmer RP, Lewis AL, Kirkwood LC, Rose SF, Lloyd AW, Vick 14. Kyomoto M, Iwasaki Y, Moro T, Konno T, Miyaji F, Kawaguchi TA, Stratford PW. Biological evaluation and drug delivery H, Takatori Y, Nakamura K, Ishihara K. High lubricious surface applications of cationically modified phospholipid polymers. of cobalt-chromium-molybdenum alloy prepared by grafting poly Biomaterials. 2004;19:4785–96. (2-methacryloyloxyethyl phosphorylcholine). Biomaterials. 34. Susin C, Qahash M, Hall J, Sennerby L, Wikesjö UM. Histolo- 2007;28(20):3121–30. gical and biomechanical evaluation of phosphorylcholine-coated 15. Kyomoto M, Moro T, Iwasaki Y, Miyaji F, Kawaguchi H, titanium implants. J Clin Peridontol. 2008;35(3):270–75. Takatori Y, Nakamura K, Ishihara K. Superlubricious surface 35. Christenson RH. Biochemical markers of bone metabolism: an mimicking articular cartilage by grafting poly(2-methacryloylox- overview. Clin Biochem. 1997;30(8):573–93. yethyl phosphorylcholine) on orthopaedic metal bearings. J 36. Tortora GJ, Grabowski S, editors. Principles of Anatomy and Biomed Mater Res A. 2009;91(3):730–41. Physiology. 8th ed. California: John Wiley & Sons Inc; 1996. p. 16. Kyomoto M, Moro T, Yamane S, Hashimoto M, Takatori Y, Ish- 143–50. ihara K. Poly (ether-ether-ketone) orthopaedic bearing surface 37. Mark SC, Odgren PR. Structure and development of the skeleton. modified by self-initiated surface grafting of poly(2-methacryloy- In: Bilezikium JP, Raisz LG, Rodam GA, editors. Principles of loxyethyl phosphorylcholine). Biomaterials. 2013;34(32):7829–39. Bone Biology. 2nd ed. New York: Academic Press; 2000. p. 17. Hench LL, Wilson J, editors. Clinical performance of skeletal 3–15. prostheses. London: Chapman & Hall; 1996. p. 12–7. 38. Lewis AL, Berwick J, Davies MC, Robberts CJ, Wang J-H, Small 18. Morrey BF, editor. Biological, material and mechanical con- S, Dunn A, O’Bryan V, Redman RP, Jones SA. Synthesis and siderations of joint replacement. New York, NY: Raven Press; characterisation of cationically modified phospholipid polymers. 1990. p. 22–47. Biomaterials. 2004;25:3099–108. 19. Sun LM, Berndt CC, Gross KA, Kncuk A. Material fundamentals 39. Sabokbar A, Millett PJ, Myer N, Rushton N. A rapid, quantitative and clinical performance of plasma-sprayed hydroxyapatite coat- assay for measuring alkaline phosphatase activity in osteoblastic ings: a review. J Biomed Mater Res A. 2001;58(5):570–92. cells in vitro. Bone Mineral. 1994;27:57–67. 20. Sandiford N, Doctor C, Rajaratnam SS, Ahmed S, East DJ, Miles 40. Murshed M, Harmey D, Millan JL, McKee MD, Karsenty G. K, Butler-Manuel A, Shepperd JAN. Primary total hip replace- Unique co-expression in osteoblasts of broadly expressed genes ment with a Furlong fully hydroxyapatite-coated titanium alloy accounts for the spatial restriction of ECM mineralisation to bone. femoral component: results at a minimum follow up of 20 years. Genes Dev. 2005;19:1093–104. Bone Joint J. 2013;95B(4):467–71. 41. Allen M, Brett F, Millett P, Rushton N. The effects of particulate 21. Gross KA, Walsh W, Swarts E. Analysis of retrieved polyethylene at a weight bearing bone-implant interface: a study hydroxyapatite-coated prostheses. J Thermal Spray Tech. 2004;13 in rats. J Bone Joint Surg [B]. 1996;82-B:32–7. (2):190–9. 42. Clarke SA, Brooks RA, Rushton N. Testing bone substitutes in a 22. Tonino A, Oosterbos C, Rahmy A, Therin M, Doyle C. small animal model of revision arthroplasty. J Mater Sci Mater Hydroxyapatite-coated acetabular components – Histological and Med. 2002;13:829–36. histomorphometric analysis of six cups retrieved at autopsy 43. Clarke SA, Brooks RA, Lee PTH, Rushton N. The effect of between three and seven years after successful implantation. J osteogenic growth factors on bone growth into a ceramic filled Bone Joint Surg [Am]. 2001;83A:817–25. defect around an implant. J Ortho Res. 2004;22:1016–24. J Mater Sci: Mater Med (2017) 28:144 Page 13 of 13 144 44. Tang L, Eaton J. Mechanism of acute response to biomaterials. 53. Martin RI, Brown PW. Aqueous formation of hydroxyapatite. J Cells Mat. 1994;4(4):429–36. Biomed Mater Res. 1997;35(3):299–308. 45. Anderson JM. The cellular cascades of wound healing. In: Davies 54. Neo M, Kotani S, Nakamura T, Yamamuro T, Ohtsuki C, Kokubo JE, editor. Bone Eng. Toronto: em2 Incorporated; 2000. p. 81–93. T, Bando Y. A comparative study of ultrastructure of the interface 46. Audrade JD, Hlady V. Protein adsorption and materials bio- between four kinds of surface-active ceramic and bone. J Biomed compatibility: a tutorial review and suggested hypothesis. Ad Pol Mater Res. 1992;26:1419–32. Sci. 1987;79:1–59. 55. Neo M, Nakamura T, Ohtsuki C, Kokubo T, Yamamuro T. 47. Jenny CR, Anderson JM. Adsorbed serum proteins responsible for Apatite formation on three kinds of bioactive materials at an early surface dependent human macrophage behaviour. J Biomed stage in vivo: a comparative study by TEM. J Biomed Mater Res. Mater. 2000;49(4):435–47. 1993;27:999–1006. 48. Lewis AL. Phosphorylcholine materials and their use in the pre- 56. Kokubo T, Kim HM, Kawashita M, Nakamura T. What kinds of vention of biofouling. Col Surf B Bioint. 2000;18:261–75. materials exhibit bone-bonding? In: Davies JE, editor. Bone 49. Lewis AL, Driver M. Blending in with the body. J Chem Ed. Engineering. Toronto: em2 Incorporated; 2000. p. 190–194. 2002;79(3):323–6. 57. Dicarlo EF, Bullogh PG. The biological responses to orthopaedic 50. Howlett CR, Evans MDM, Walsh WR, Johnson G, Steele JG. implants and their wear debris. Clin Mater. 1992;9:235–6. Mechanisms of initial attachment of cells derived from human 58. Moroni A, Orienti L, Stea S, Visentin M. Improvement of the bone to commonly used prosthetic materials during cell-culture. bone-pin interface with hydroxyapatite coating: an in vivo long Biomaterials. 1994;15(3):213–22. term study. J Ortho Trauma. 1996;10(4):236–42. 51. Billiard J, Moran RA. Transcriptional profiling of human osteo- 59. Martin EA, editor. Concise colour medical dictionary. 2nd ed. blast differentiation. J Cellular Biochem. 2003;89(2):389–400. Oxford: Oxford University Pres; 1998. 52. Stein GS, Lian JB. Molecular mechanisms mediating prolifera- 60. Dagtekin G, Schiffer R, Klein B, Jahnen-Dechent W, Zwadlo- tion, differentiation interrelationships during progressive devel- Klarwasser G. Modulation of angiogenic functions in human opment of the osteoblast phenotype. Endocrine Rev. 1993;14 macrophages by biomaterials. Biomaterials. 2003;24:3395–401. (4):424–42. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Materials Science: Materials in Medicine Springer Journals

The effect of cationically-modified phosphorylcholine polymers on human osteoblasts in vitro and their effect on bone formation in vivo

Download PDF
13 pages

Loading next page...
1
 
/lp/springer_journal/the-effect-of-cationically-modified-phosphorylcholine-polymers-on-nzcxoeOKC2

References (75)

Publisher
Springer Journals
Copyright
Copyright © 2017 by The Author(s)
Subject
Materials Science; Biomaterials; Biomedical Engineering; Regenerative Medicine/Tissue Engineering; Polymer Sciences; Ceramics, Glass, Composites, Natural Materials; Surfaces and Interfaces, Thin Films
ISSN
0957-4530
eISSN
1573-4838
DOI
10.1007/s10856-017-5958-8
pmid
28819908
Publisher site
See Article on Publisher Site

Abstract

J Mater Sci: Mater Med (2017) 28:144 DOI 10.1007/s10856-017-5958-8 BIOCOMPATIBILITY STUDIES Original Research The effect of cationically-modified phosphorylcholine polymers on human osteoblasts in vitro and their effect on bone formation in vivo 1 1 2 2 ● ● ● ● Jonathan M. Lawton Mariam Habib Bingkui Ma Roger A. Brooks 1 3 2 1 ● ● ● Serena M. Best Andrew L. Lewis Neil Rushton William Bonfield Received: 30 January 2017 / Accepted: 3 August 2017 / Published online: 17 August 2017 © Springer Science+Business Media, LLC 2017 Abstract The effect of introducing cationic charge into biotolerant, stimulating the production of fibrous tissue and phosphorylcholine (PC)-based polymers has been investi- areas of loosely associated matrix (LAM) around the gated in this study with a view to using these materials as implant. Their development, as formulated in this study, as coatings to improve bone formation and osseointegration at bone interfacing implant coatings is therefore not warranted. the bone-implant interface. PC-based polymers, which have been used in a variety of medical devices to improve bio- Graphical abstract compatibility, are associated with low protein adsorption resulting in reduced complement activation, inflammatory response and cell adhesion. However, in some applications, such as orthopaedics, good integration between the implant and bone is needed to allow the distribution of loading stresses and a bioactive response is required. It has pre- viously been shown that the incorporation of cationic charge into PC-based polymers may increase protein adsorption that stimulates subsequent cell adhesion. In this paper, the effect of cationic charge in PC-based polymers on human osteoblasts (HObs) in vitro and the effect of these polymers on bone formation in the rat tibia was assessed. Increasing PC positive surface charge increased HOb cell adhesion and stimulated increased cell differentiation and the production of calcium phosphate deposits. However, when implanted in bone these materials were at best Jonathan M. Lawton and Mariam Habib contributed equally to this work * Andrew L. Lewis 1 Introduction andrew.lewis@btgplc.com Advances in modern medicine, such as the introduction of Department of Materials Science and Metallurgy, Cambridge penicillin, antiseptics, and vaccinations, have significantly Centre for Medical Materials, University of Cambridge, New Museum Site, Cambridge CB2 3QZ, UK increased human life expectancy [1]. The ageing population in the West (in the UK and US alone there are over 100 Orthopaedic Research Unit, University of Cambridge, Addenbrookes Hospital, Hills Road, Cambridge CB2 2QQ, UK million individuals over the age of 50) [2], coupled with the deterioration of bone stock with age (particularly for woman Biocompatibles UK Ltd, Chapman House, Farnham Business Park, Weydon Lane, Farnham, Surrey GU9 8QL, UK of post-menopausal age) [3], mean that the clinical need for 144 Page 2 of 13 J Mater Sci: Mater Med (2017) 28:144 tissue repair and replacement has never been greater. Bone biomineralization [24–26]. This suggested the potential of tissue loss, predominantly due to degenerative diseases PC polymer to also interact with calcium and form a stable (such as osteoporosis), osteosarcoma and trauma [4], has led interface with bone mineral. However, osteoblast activity at to the design of a variety of surgical techniques, medical the interface would be required for osteointegration and devices and specialised materials. These range from total unmodified PC polymer has reduced cellular interaction. hip and knee replacements to fracture fixation devices and Attempts have been made to modify PC polymers to pro- bone stock replacements. duce materials with inherent biocompatibility but also with Phosphorylcholine (PC) materials are bio-inspired poly- additional components to invoke specific biological inter- mers that mimic the extracellular surface of red blood cells, actions. In one study, a PC-containing copolymer of N- containing an exact chemical copy of the predominant isopropylacrylamide and N-(n-octadecyl) acrylamide was zwitterionic phospholipid headgroup found in the cell lipid shown to preferentially adhere U937 macrophages, which membrane. Unlike most biomaterials, the well-hydrated and subsequently increased expression of TNF-α in response to 2+ 3+ neutrally-charged PC surface allows for the interaction of Co and Co ions [27]. In another approach, researchers proteins without inducing shape changes in the protein’s have also shown that the introduction of surface charge into three-dimensional structure and thus reduce irreversible biomaterials increases cell spreading, adhesion, growth and protein adsorption [5]. Furthermore, this decrease in protein proliferation [28–31]. Rose et al [32] demonstrated that the adsorption results in decreased blood clotting [6], cellular introduction of a cationic moiety (choline methacrylate adhesion, and in a reduction in the inflammatory response chloride salt – CMA) into PC polymers also significantly and fibrous capsule formation [7]. Such properties have increased the amount of protein adsorption and subsequent resulted in PC materials being used for a variety of bio- cell attachment, where the amount of protein adsorption and medical applications where a passive interaction between cell attachment was found to be protein, cell and CMA the material and the body is required. Examples include monomer content specific. The authors [32] concluded that coating blood contact devices, such as coronary guide wires the increase in PC bioactivity was caused by the increase in [8] and stents [9], extracorporeal circuits [10], and for the polymer surface charge, whereby the predominantly contact lenses [11]. PC materials have also found utility in negatively charged biomolecules interact electrostatically the orthopaedic field, as superlubricious, low-wear surfaces with the positively charged surfaces. Moreover, Palmer et al grafted onto polyethylene acetabular liners [12], for which [33] have shown that cationically- modified PC materials there are now 3 year data from an 80 patient study [13], can also act as drug delivery vehicles, where negatively colbalt-chromium-molybdenum metal alloy bearings [14, charged therapeutic biomolecules interact reversibly with 15] and poly(ether-ether ketone) orthopaedic bearing sur- PC’s positive charge and are released in a controlled man- faces [16]. ner. In a comparative study of PC and cationically-modified However, in some clinical applications a “bioinert” PC coatings on titanium porous oxide surface implants response is not the most appropriate. For example, ortho- placed in trabecular and cortical bone of rabbits [34], paedic implants, such as artificial hips, require a degree of although there was no significant differences in bone den- interaction between the femoral stem and the surrounding sity between the coatings or uncoated control, there was a bone tissue in order to enhance load transfer and to decrease higher bone-implant contact for the cationically-charged PC micromotion and stress shielding of the implant, which can coating and uncoated control compared to the PC coating lead to implant loosening and premature failure [17, 18]. (p < 0.05) at the 6 week explant point. This somewhat Furthermore, increasing the speed and quantity of bone- indicates that the surface charge is able to orchestrate cel- implant integration allows early loading, resulting in lular interaction, overcoming the inherent anti-adhesiveness quicker patient mobility, improved patient well-being, of the PC material, but only histometric and biomechanical shorter hospital stays, and a reduction in healthcare costs analysis was performed and hence no mechanism proposed. [1]. An important development in this regard was plasma- The aim of current work presented herein was to investigate sprayed hydroxyapatite coatings which produce improved the effects of high and low charge levels of cationically- bone bonding of implants to bone and increased longevity charged PC materials on human osteoblast cells (HObs). of some joint replacements [19, 20]. However, the risks Osteoblasts are skeletal cells responsible for bone formation associated with these relatively thick coatings is delamina- [35]. They are derived from messenchymal stem cells found tion and there is continued concern about the resorption of in the bone marrow [36] via osteoprogenitor cells and they HA by interfacial remodelling in some applications [21– secrete extracellular matrix (type I collagen) and non- 23]. The phosphorylcholine side chain of the PC-polymer collagenous material that is involved in the nucleation and contains the phosphate moiety present in phosphatidylcho- formation of bone mineral [37]. Hence the cells are likely line and phosphatidylserine, the latter being membrane lipid being influenced either directly or indirectly by the implant component that is able to bind calcium and is implicated in surface characteristics. We describe in vitro experiments in J Mater Sci: Mater Med (2017) 28:144 Page 3 of 13 144 order to shed light on potential mechanisms of interaction 2 h at room temperature, the polymers were cured onto the between biomaterial and tissue and investigate the in vivo plates and flasks at 70 °C for 72 h. The coated culture dishes longevity of any beneficial effect in a 14 week rat tibia were then sterilised under UV light for 2 h. implantation model. 2.1.2 Cell culture 2 Materials and methods Human osteoblasts (HObs) (PromoCell, Heidelberg, Ger- many) isolated from hip bone were grown in 75 cm tissue 2.1 In vitro experiments culture flasks in McCoy’s 5a growth medium (Invitrogen, Paisley, UK) containing 10% foetal bovine serum (FBS), 2.1.1 Materials 1% glutamine and 30 μg/ml vitamin C (standard culture conditions) (Sigma, Poole, Dorset, UK). After cryogenic The PC polymers used in this experiment have previously recovery and 24 h incubation, cell adherence was checked been characterised [38]. They are methacrylate-based and the medium replaced; thereafter the medium was polymers synthesised using free radical polymerisation of replaced every 2 days. At 60–80% confluence, the cells 2-methacryloxyethylphosphorylcholine (MPC), lauryl- were washed three times with Hank’s Balanced Salt Solu- methacrylate (LMA), hydroxypropylmethacrylate (HPMA) tion (HBSS) (Invitrogen) and incubated (RT for 5 min) and trimethoxysilylmethacrylate (TMSMA). Positive before trypsin/ethylenediaminetetraacetic acid (EDTA) charge was introduced by including choline methacrylate (Sigma) was used to detach the cells from the flask. Com- (CMA). A schematic representation of this polymer is plete cellular detachment was confirmed by phase-contrast shown in Fig. 1 and the constituents (weight percentages) microscopy. After 5 min, McCoy’s growth medium was and their function are detailed in Table 1. In addition to the added to neutralise the trypsin/EDTA. The cell suspension other polymers, PC5 was used in the in vitro experiments was then centrifuged at 220 × g for 4 min at room tem- and PC6 was used to coat the in vivo implants due to the perature before the supernatant was aspirated and the cell availability of these two substrates when these two experi- pellet re-suspended in 1 mL of McCoy’s medium and ments were carried out. The difference in response to the counted using a haemocytometer. small difference in cationic charge between these two PC polymers is unlikely to be significant. The polymers were 2.1.3 Cell number supplied by Biocompatibles UK Ltd (Farnham, Surrey, UK). HOb cells were cultured on non-coated, PC5 and PC20 Each PC polymer was dissolved in ethanol at a con- coated T25 flasks for 28 days in medium (as previously centration of 10 mg/ml and the solution filtered through a described), incubated at 37 °C in 5% CO humidity. Cell 0.2 μm filter. Tissue culture plates and flasks were then number was determined at 6 h, 1, 2, 4, 7, 14 and 28 days TM filled with the polymer-ethanol solution and left for 5 min at using Vialight HS proliferation/cytotoxicity kit (Lonza, room temperature before being emptied. After air-drying for UK). In this assay, luciferase enzyme reacts with adenosine trisphosphate (ATP) (found in all metabolically active cells) to emit light and its intensity is indicative of cell number. At each time point the culture media was aspirated and the cells washed 3 times with phosphate buffer saline solution (PBS) (Sigma). For each flask 1 mL 0.5% v/v Triton X-100 (VWR, Lutterworth, UK) in PBS was added and two freeze thaw cycles (15 min at −70 °C and 37 °C) were used to lyse the cell suspension. 180 μL of the lysate was then trans- ferred in duplicate to a white-walled 96-well plate and 20 μL of ATP monitoring reagent added to each well. The ATP concentration was then immediately read (TopCount. TM NXT , Packard BioScience Co, Meriden, USA). Using a standard curve (light intensities of pre-determined HOb cell numbers of 0, 2500, 5000, 10000, 20000 and 40000 on uncoated well plates) the cell numbers on the substrates was determined. Cell number was also determined for cells on PC0 at 6 h, 1 and 2 days. Three separate flasks were used Fig. 1 A schematic chemical representation of a cationically charged phosphorylcholine polymer for each substrate and time point. 144 Page 4 of 13 J Mater Sci: Mater Med (2017) 28:144 Table 1 Constituents of the Monomer content (wt%) phosphorylcholine materials (weight percent) Monomer PC20 PC6 PC5 PC0 Component function MPC v 22 29 28 29 Phosphorylcholine head group—charge neutrality—non- thrombogenic LMA w 41 50 50 51 Lauryl group—alkyl chain component—hydrophobic—substrate adsorption HPMA x 12 12 12 15 Aids crosslinking and film formation TSMA y 5 4 5 5 Methoxysilyl crosslinker—mechanical stability. Aids substrate adhesion TMA z 20 6 5 0 Cationic charge 2.1.4 Cell differentiation characterise the elemental constituents of mineral exudates on the cells and ECM (extracellular matrix). HOb cells were cultured on non-coated, PC5 and PC20 coated T25 flasks for 28 days in McCoy’s medium. Cell 2.2 In vivo experiment differentiation was determined by identifying the enzyme alkaline phosphatase (ALP), a widely recognised [39] 2.2.1 Materials enzyme marker of osteoblast differentiation associated with skeletal mineralisation [40]. 0.5 M 2-amino-2-methyl-1- Thirty surgical grade stainless steel (316 L) pins (R J Lay- propanol (AMP) substrate buffer (Sigma) was prepared in land, Rayleigh, UK) were cleaned using dry Emory cloth distilled water (pH 10) and supplemented with 2 mM paper (P1200, T444 Norton, Tufbak Durite) and sonicated magnesium chloride (MgCl ) and 9 mM p-nitrophenol for 5 min in both acetone and ethanol. The pins were phosphate (p-NPP) (Sigma). In an alkaline solution, ALP equally assigned to one of five groups and coated accord- catalyses p-nitrophenyl phosphate to p-nitrophenol, ingly. Six pins were uncoated (negative control). Six pins appearing yellow in colour. Cells were lysed at 1, 2, 4, 7, 14 were plasma-spray coated with 60–100 μm of hydro- and 28 days in 3 flasks per substrate per timepoint (as xyapatite (Plasma Biotal, Tideswell, UK) (positive control). described for ATP measurement) and 50 μL of each lysate Three groups of six were then double dip coated at 4 mm/ in duplicate and 50 μL of a range of p-nitrophenol con- sec in 5 mg/ml polymer concentrations in ethanol of PC0, centrations, diluted to produce a standard curve were PC6 or PC20 respectively (as previously described in Table transferred to a 96-well plate. Subsequently, 50 μLofAMP 1), resulting in polymer coatings approximately 30–50 nm buffer was added to each well and the plate incubated at thick. The polymer coated pins were then cured at 70 °C for 37 °C for 15 min. Absorbance was immediately read at 405 4 h. All pins were designed to have equal dimensions prior nm using an EL800 Universal Microplate reader (Bio-tek to implantation (Fig. 2a) and were sterilised using gamma Instruments, Inc., Winooski, USA). A standard curve of irradiation (25 kGy, Isotron, UK). absorbance as function of p-nitrophenol concentration was generated and used to determine the total ALP content of 2.2.2 Surgery each test sample. Optical Imaging and Energy Dispersive X-ray (EDX) The surgical method for implantation has previously been Characterisation of Cells HOb cells were cultured (seeding described by Allen et al [41] and others [42, 43]. All pro- 3 2 density of 12 × 10 /cm ) in T25 flasks coated with PC0, cedures underwent ethical review and were carried out in PC5 or PC20 and on a non-coated tissue culture plastic accordance with the regulations as set out in the Animal control for 48 h and 28 days. Optical imaging using a Leitz (Scientific Procedures) Act 1986. Thirty mature Sprague Labovert phase contrast microscope was carried out on all Dawley rats, weighing between 300–350 g (Harlam UK substrates at 48 h. For SEM, the cells on the substrates at 48 Ltd, Bicester, UK) were randomly allocated to one of the h and 28 days, were washed with PBS, fixed in 4% paraf- five groups. The number of animals used per group was ormaldehyde (Sigma) and washed three times with distilled based upon previous studies which showed significant water. A square piece, approximately 20 × 20 mm, of each results with groups of 6. One pin was implanted into the flask was cut and mounted onto an SEM stub using double- right tibia of each animal (Fig. 2b). sided carbon tape and sputter coated with carbon. A JEOL All drugs were obtained from National Veterinary Ser- XL30 scanning electron microscope (SEM) at 5 kV vices, Stoke-on-Trent, UK. Briefly, anaesthesisa was was used to image the samples and EDX was used to induced using a gaseous mixture of oxygen and halothane J Mater Sci: Mater Med (2017) 28:144 Page 5 of 13 144 Fig. 2 a Schematic representation of the stainless steel pins used for implantation; b a radiograph indicating the placement of the surgical pin in the rat tibia; c diagrammatic map of a laterel sectioned pin. The figure indicates the locations where images for histological analysis were captured (1–5) and the local distinction between the interface and interfacial area (4%), at a rate of 6 litres/min in an anaesthetic chamber. de-fatted using continuously topped-up acetone under Midazolam (3 mg/kg) was given by intra-peritoneal injec- vacuum for 7 days. The tibias were then embedded in poly tion and the rats maintained with halothane 4% delivered (methyl methacrylate) (PMMA). Methyl methacrylate through an anaesthetic mask. The right hind limb was (Sigma) containing 2.5% w/v benzoyl peroxide (catalyst) shaved, peri-operative analgesia administered sub- and 2.5% v/v dibutyl phthalate (plasticiser) was infiltrated cutaniously (Rimadyl—carprofen, 5 mg/ml) and the rat into the specimens at 4 °C under vaccuum for 7 days. The placed supine. A 1.5 cm incision was made lateral to the methacrylate solution was then replaced with fresh solution patella and a lateral capsulotomy performed allowing and polymerised. Once polymerised, 300–500 μm thick medial dislocation of the patella exposing the tibial plateau. longtitudinal sections (sagital plane of the pin and tibia) A 1.5 mm diameter hand drill was used to drill through the were cut using a diamond saw (100 CA blade 0.1 mm/sec, centre of the tibial plateau to a depth of 10 mm and counter Accuton 5, Struers, Glasgow, UK) and glued (Cyno 40, sunk (using a 4 mm counter-bore) to allow the head of the Delta, Leeds, UK) onto frosted glass microscpe slides implant to be flush with articulating surfaces. The pin was before being ground and polished to 100 μm thick using a then press-fit into the tibia, the patella reduced and the graded series of silicon carbide papers (Struers, Denmark). incision closed. Immediately after surgery the rats received 150 mg/kg of antibiotic (Synulox—coamoxyclavulanic 2.2.4 Microscopy and histological analysis acid) intramuscularly and a sub-cutaneous injection of 0.15 mg/kg analgesia (temgesic: buprenorphine). The rats were Slides were washed in liquid detergent and water for 1 min, allowed to recover in IVC caging systems (Techniplast UK rinsed in water for 3 min and stained with toluidine blue Ltd, Northants, UK) and after 5 h a second injection of (pH 9, 56 °C, 30 min). Five fields (Fig. 2c) for each slide temgesic (0.1 mg/kg) was administered. were digitally photographed using a Leitz Dialux 20 light microscope (X100) with attached camera and stored on a 2.2.3 Tissue processing computer (Aquis Image Acquisition software—Synoptics, Cambridge, UK). The rats were sacrificed at 14 weeks according to Home The head of the pin was ignored due to relatively poor Office (UK) schedule 1. The right tibia was removed and bone apposition in this region. Such a response may be due fixed in ice-cold paraformaldehyde in 0.1 M phosphate to stress shielding (poor stress distribution through the pin) buffer (pH 7.4) with 0.1% w/v sucrose and 0.05% v/v resulting in bone resorption rather than due to the problems gluteraldehyde at 4 °C. After 5 h the tibias were washed associated with surgery and experimental technique and has three times in PBS and dehydrated in ascending con- been observed previously [33]. The image was then edited centrations of methylated spirit (50–100%) before being in Adobe Photoshop 6.0 and cropped to allow a distance of 144 Page 6 of 13 J Mater Sci: Mater Med (2017) 28:144 50 μm from the pin surface into the surrounding bone and tissue plastic control. Interestingly, the number of cells bone marrow leaving an image of the bone-implant inter- decreased on PC5 with increasing time, up to 28 days. After face (Figs. 2c and 8). Four tissue types, bone, marrow, day 1, significant differences (P < 0.001) existed between fibrous tissue and loosely associated matrix (LAM), were PC5 and PC20 and control surfaces for all time points. This then identified and colour coded. The coloured images were indicates that HOb cells proliferated well on PC20, how- then processed into binary images. Scion Image Analysis ever, proliferation on control surfaces was significantly (Scion Corporation, USA) was used to measure the areas of greater at all time points (P < 0.001). the four tissue types and these were expressed as a per- Figure 4b shows the effect of time and charged PC centage of the total area for each field position on each substrate on cell differentiation. The ALP activity per individual pin (interfacial area). The percentage of each thousand cells decreased on PC20 over the first 3 days of tissue was also calculated along the length of the interface culture and on the control substrate between day 1 and 2. between the pin and the bone (interface). The five positions Thereafter ALP activity did not change significantly and along the pin were then summed and averaged. there was no significant difference between ALP activity on PC20 and control substrates after day 2. Alkaline phos- phatase activity and the number of cells on PC5 was too low 2.2.5 Statistical analysis to give an accurate measure of normalised activity For both the in vivo and in vitro experiments the data was compared using a one-way ANOVA. Identified trends were further investigated using either specific t-testing or Bon- feronii analysis to highlight significant differences. The distribution of the data is represented using standard error. 3 Results 3.1 In vitro experiments Figure 3 indicates the number of HOb cells on PC coated and control surfaces over 48 h. At each time point, tissue plastic control surfaces exhibited the largest number of cells. Significant differences (P < 0.001) were observed between the control and PC0 and between PC20 and PC0 surfaces at all three time points. However, significant differences (P < 0.001) between PC20 and PC5 were only seen at 24 and 48 h. Figure 4a indicates the number of HOb cells on charge- modified PC coated and control surfaces over 28 days. The number of HOb cells increased with time on PC20 and Fig. 4 a The effect of cationically-modified PC on the number of HOb cells in vitro over a 28-day period in culture as compared to control Fig. 3 The effect of PC and cationically-modified PC on the number surfaces. The number of cells on control surfaces was significantly of HOb cells in vitro after 6, 24 and 48 h in culture as compared to greater (p < 0.001) than on PC20 and PC5 at all time points and on control surfaces. *p < 0.001 compared to other materials at the PC20 than PC5 after day1; b corresponding effect of cationically- equivalent time point. +p < 0.001 compared to PC20 and control at modified PC on HOb cell ALP production in vitro over a 28-day the equivalent time points period as compared to control surfaces J Mater Sci: Mater Med (2017) 28:144 Page 7 of 13 144 Fig. 5 Photomicrographs of HOb Cells on a PC0, b PC5, c PC20 and d control surfaces at 48 h in vitro. (Scale bars = 100 μm) Figure 4a shows the effect of PC charge content on HOb tissue than both bone and marrow and LAM. For SS there cell proliferation and spreading at 48 h and supports the data was also significantly more bone and marrow than LAM. presented in Fig. 3. PC20 and the control surfaces showed HA had the highest bone and marrow apposition compared the largest number of cells, and on these surfaces the HOb to the other four surfaces and also less fibrous tissue (p < cells have developed fully-spread morphologies (Fig. 5). 0.05). In a comparison of the PC materials, PC0 had sig- PC0 supported minimal cell adhesion and, although there nificantly (p < 0.05) more fibrous tissue than PC6 but there were more cells on PC5, the cells on both these substrates were no significant differences between PC0 and PC20 or remained rounded. The photomicrographs in Fig. 5 also between PC0 and SS. Furthermore, there was no significant indicate that the cells on tissue plastic controls and on difference between any of the PC groups in relation to the PC20 showed the presence of mineral deposits. formation of LAM. Figure 6 shows SEM with elemental analysis and sup- Figure 7b indicates the amount (percentage) of the dif- ports Fig. 5, confirming both the presence of mineral ferent tissue types in the bone-implant interfacial area deposits on PC20 and control surfaces at 48 h and that the defined in Fig. 2c. The data complements that described in number of these mineral deposits increased with time (up to Fig. 7a, however, in this instance bone and marrow, sig- 28 days). Figure 6 also shows that the mineral deposits on nificantly (p < 0.05) dominate over fibrous tissue and LAM both PC20 and control surfaces at 28 days contained both formation for all surface coatings. For PC6, PC0 and SS calcium and phosphorus. there was significantly (p < 0.05) more fibrous tissue than LAM. In a similar manner to the implant interface, Fig. 7b 3.2 In vivo results indicates that HA had, significantly (p < 0.05), the most bone and marrow apposition and the least fibrous tissue formation in comparison to the other four surfaces. PC0 had Figure 7a indicates the amount (percentage) of different types of tissue at the bone-implant interface. For HA there significantly (p < 0.05) more fibrous tissue than PC6. Analysis of these data shows that there is a trend for a linear was significantly (p < 0.05) more bone and marrow than both fibrous tissue and LAM. Both cationically charged PC increase in LAM with increasing CMA content (R = 0.9) materials produced a similar response, with no significant in the implant interfacial area; a weak linear dependency of differences found between any of the tissue types. For both LAM with CMA content was also found at the implant PC0 and SS there was significantly (p < 0.05) more fibrous interface (R = 0.7). Typical histological micrographs of the 144 Page 8 of 13 J Mater Sci: Mater Med (2017) 28:144 Fig. 6 SEM images showing osteoblast cells on a PC20 and b control at 48 h and c PC20 and d control at 28 days. An increase in the amount of mineral deposits (MD) were seen on both PC20 (c) and control (d) surfaces at 28 days. Corresponding elemental analysis is shown in the inset for each image and demonstrates that these mineral deposits contain calcium and phosphorus bone-implant interface are shown in Fig. 8. These show a thin layer of fibrous tissue between the implant and bone (Fig. 8a–c) with the exception of HA where the bone is directly apposed to the ceramic (Fig. 8d). The LAM is primarily seen adjacent to the PC coated implants (Fig. 8b, c). 4 Discussion 4.1 In vitro response It has been well documented that non-charge modified phosphorylcholine surfaces (PC0) significantly reduce cel- lular adhesion compared to tissue culture plastic controls [7, 9, 32]. Within 1 min of implantation most biomaterial sur- faces are coated with a protein film [44], which in an in vivo environment can comprise over 200 different proteins [45]. Proteins adsorb onto surfaces to minimise surface free energy. This can occur through protein conformational Fig. 7 The effect of PC and cationically-modified PC on the percen- change, partial dehydration of the protein and surface, and tage of bone and marrow, fibrous tissue and loosely associated matrix the redistribution of surface charges [46]. Subsequent cel- (LAM) compared to hydroxyapatite (HA) and stainless steel (SS) lular interactions are largely governed by the nature and positive and negative controls at a the bone-implant interface. * Sig- conformation of these adsorbed proteins whereby the nificantly more (p < 0.05) than the other two tissues on that material. +Signficantly different amounts of tissue on HA (p < 0.05) than the adsorbed proteins can act as ligands for cell-surface recep- same tissue on the other 4 materials. b In the bone-implant interfacial tors [47]. PC0 surfaces interact with proteins without area inducing conformational shape changes in their structure J Mater Sci: Mater Med (2017) 28:144 Page 9 of 13 144 Fig. 8 Typical experimental micrographs indicating different tissue types in the bone-implant interfacial area, where M is marrow, B is bone, FT is fibrous tissue and LAM is loosely associated matrix, for a stainless steel; b, c PC20; d hydroxyapatite (HA) resulting in a decrease in protein adsorption. It is thought Figure 4a shows that HOb cells attached to PC5 after 24 that their highly hydrophilic phosphorylcholine zwitterionic h are in a non-proliferate state. This is in stark contrast to head group (MPC monomer, Fig. 1) results in the formation the number of cells proliferating on both tissue culture of a hydration layer that allows proteins to interact with the plastic and PC20 up to 28 days. Generally, initial cell surface reversibly whereby the hydrated layer limits the adhesion occurs in the first 30–120 min of cell-surface dehydration step involved in adsorption [48]. This zwitter- contact followed by a stronger attachment involving the ionic group may also decrease protein adsorption through secretion and development of ECM, cellular migration and charge neutrality. The subsequent decrease in protein growth [50]. The larger number of proliferating osteoblasts, adsorption significantly reduces cellular adhesion, including at 48 h, on PC20 and the control in comparison to PC5 can monocytes, macrophages, fibroblasts and human granulo- also be seen visually in Fig. 5. This suggests that there is cytes [49]. Such a response was also observed in HOb cells only a weak interaction formed between the PC5 surface (Fig. 3), whereby there are significantly (p < 0.001) less and the osteoblast cells, reducing stimulation of the cells to HOb cells on PC0 at 6, 24 and 48 h compared to PC5, PC20 produce ECM. The lack of ECM production is likely to and tissue culture plastic. The lack of cell adhesion to PC0 cause the subsequent loss of cells by apoptosis, which is can also be seen visually in Fig. 5. seen at the later time points (Fig. 4a). The weaker interac- Figure 3 indicates that over early time periods (up to 48 h) tion between PC5 and the osteoblast cells is likely to relate increasing CMA content increases the number of adherent to the charge sensitive adsorption of proteins. For example, HOB cells. This is in agreement with the findings of Rose PC5 may not contain enough positive charge to adsorb et al [32] who showed that increasing CMA content in PC vitronectin or fibronectin, adhesive glycoproteins that are polymers increased cell adherence. The increase in cell known to play a key role in the anchorage of osteoblasts attachment is likely to be caused by either the non-specific [50]. electrostatic interaction of the predominantly negative cells The results for ALP production by the cells shown in and the positively charged surface or through an increase in Fig. 4b have been adjusted for the change in cell number in the number of cellular adhesive proteins adsorbed onto the the cultures shown in Fig. 4a. ALP is an early marker of surface of the PC polymer caused by the increase in CMA osteoblast differentiation [51] and is only up-regulated by content. It is possible that both these mechanisms are run- osteoblast cells after proliferation has finished and prior to ning concurrently. Rose et al [32] also showed that the mineralisation of the ECM. Once mineralisation starts, ALP inclusion of CMA content in PC polymers resulted in an production ceases [52]. On PC20 and control substrates the increase in protein adsorption compared to their non- osteoblasts continue proliferating throughout the culture charged counterparts and that specific proteins have a period, as seen in Fig. 4a, and do not reach confluence, minimal CMA or charge content requirement. Such which would enable the onset of differentiation throughout dependencies may help explain the results observed in Fig. the culture and increase ALP production. ALP production 4a that indicate the effect of PC cationic charge on HOb cell per thousand cells decreases over the first 2 days in these proliferation up to 28 days. cultures as the newly divided cells are not producing ALP. 144 Page 10 of 13 J Mater Sci: Mater Med (2017) 28:144 Thereafter it is likely that groups of cells are showing PC0 show significantly (p < 0.05) more fibrous tissue than increased ALP but not the proliferating cells in the culture; any other tissue type (Fig. 7). This is at odds with work by thus, the ALP activity per thousand cells does not show a Goreish et al. [7] who showed that intramuscular PC0 marked change. On PC5 the ALP production remains implants in the rabbit were surrounded by 40% less constant over the initial 2 days as cell number does not inflammatory cells than polyethylene controls, concluding increase after which the very low cell number and ALP that this was due to PC0’s well hydrated surface and passive production prevent an accurate measure of ALP production interaction with the surrounding tissue. Implant location per thousand cells. (bone is rich in pre-matrix producing cells that can differ- Evidence of mineral formation on PC20 is supported by entiate into fibroblasts), the nature of the controls (poly- Fig. 5, which shows that on both PC20 and tissue plastic ethylene is known to elicit a high fibrous response and control surfaces HOb cells have developed fully spread stainless steel (SS) 316 is reasonably “bioinert” in an oss- morphologies that contain deposits of calcium phosphate eous environment), as well as the different animal studies (Fig. 6), which increased with time, up to 28 days. How- used may explain the observed differences. ever, the EDX spectra (insets of Fig. 6a-d) showed that the The reasons for the effect of PC CMA content on bone calcium phosphate (Ca:P) ratio of both PC20 and control and marrow apposition is also debatable. Cationic charge was approximately 0.6, considerably lower than hydro- decreases fibrous tissue formation with a consequential xyapatite (the mineral phase of bone) which has a Ca:P ratio increase in bone and marrow apposition suggesting that of between 1.5 and 1.67 [53]. A lower Ca:P ratio is indi- increasing surface charge increases osseous-implant inte- cative of calcium deficient mineral. Such a calcium deficient gration (Fig. 7a). However, this increase is not linear with HA maybe the due to the presence of the cationic surface increasing CMA content and both charged polymers (PC6 2+ charge of the PC that may be preventing Ca deposition and PC20) show similar amounts of both bone and marrow into the mineral. and fibrous tissue (Fig. 7). Furthermore, no significant dif- ference in bone and marrow and fibrous tissue exist between 4.2 In vivo response the negative control (SS) and the two charged PC coatings. One might expect that increasing surface charge would Figure 7a, b shows the amount (percentage) of different increase fibrous tissue formation, due to surface charge tissue types at the bone-implant interface and interfacial aggravates the foreign body response. However, Rose et al. area respectively. The results indicate that the HA coating [32] showed that although the presence of surface charge in PC increases the number of attached fibroblast cells, unlike creates excellent bone apposition against the implant, showing significantly (p < 0.05) more bone and marrow other cells (eg monocytes and granulocytes), the number of implant apposition compared with the other four substrates. fibroblasts on all charged materials PC materials with up to The degree of this apposition can be seen in Fig. 8d and is 30% CMA was maintained at a constant level. in agreement with the literature [54, 55]. Hydroxyapatite An explanation for the decrease in both bone and marrow (Ca (PO )OH) based coatings create an interactive bond apposition and fibrous tissue formation with increasing PC 5 4 between the bone and implant that is characterised by the CMA content is the increase in LAM with increasing sur- presence of a thin calcium phosphate (apatite) layer, which face charge (Fig. 7a). LAM is shown visually in Fig. 8c, d forms early in the implantation process. Kokubo et al. [56] and can be characterised as a fluid-containing capsule argue that this intermediate apatite layer allows for the comprised of a few cells and loosely aggregated matrix preferential proliferation and differentiation of osteoblasts surrounded by a fibrous membrane. No reference in the on its surface, stimulating the formation of new bone. literature exists as to the formation of such tissue at the In contrast, the phosphorylcholine and the stainless steel implant-bone interface, though such tissue may be descri- implants are apposed with fibrous tissue and LAM (Figs. 7 bed as lesions; zones of tissue which have impaired function and 8a–c). Fibrous encapsulation of so-called “bio-tolerant” as a result of damage or disease [59]. The correlation (R )of polymeric and metallic implants at the bone-implant inter- increasing LAM formation with CMA content increases face has been well documented and is a consequence of the from 0.7 to 0.9 when examining the interfacial area (Fig. foreign-body reaction [57, 58]. After the immediate protein 7b) of the implant as apposed to the interface (Fig. 7b) and adsorption that follows implantation, local tissue damage is due to LAM extending from the surface of the implant causes the recruitment of white blood cells, including and occupying space (Fig. 8b, c). LAM formation maybe macrophages, which subsequently bind to the surface of the caused by the preferential attachment of inflammatory cells implant. However, these phagocytic cells cannot ingest the such as monocytes that may become over stimulated in the implant and fuse to form giant cells, recruiting fibroblasts, production of tumour necrosis factor (TNF) -α, a chemo- which subsequently secrete collagen matrix resulting in the tactic pro-inflammatory cytokine that can cause the forma- fibrous encapsulation of the implant [45]. Both SS and tion of necrotic tissue [60]. In addition, the increase in J Mater Sci: Mater Med (2017) 28:144 Page 11 of 13 144 charge may also be acting on cells in such a way that it post-implantation to allow for a period of fuller integration results in local cell apoptosis. with the implant, as previous work had seen some differ- Rose et al. [32] has shown that a variety of different cells ences at just 6 weeks post implant [34]. In fact, as bone and proteins are affected by the inclusion of positive surface ingrowth appears to occur at an earlier stage, our model may charges into PC polymers. However, it is not clear how have suffered from the potential for newly-synthesised bone selective this effect is in a multi-cellular in vivo environment. to undergo subsequent bone resorption due to mechanical For example, fibroblast cells may preferentially attach to the stress or excessive inflammation. PC surface over osteoblast cells. Furthermore, both osteo- blasts and fibroblasts share the same precursors in the bone marrow and these surfaces may be inducing fibroblast dif- 5 Conclusions ferentiation over osteoblasts, resulting in a fibro-encapsulated sheath. In addition, pro-inflammatory cells may preferentially Modifying phosphorylcholine surfaces by the addition of bind to the charged PC surfaces and induce cell death and the cationic charge alters the response of cells to these surfaces production of LAM [27]. A further experiment, which and could allow the delivery of drugs, including antibiotics, examines the effects of PC and modified PC surfaces in a from PC implant coatings. However, our studies show that multicellular environment, would be an interesting way to whilst increasing positive surface charge allowed increased assess which cells are preferentially attaching to the polymer osteoblast adhesion to the PC surface and stimulated surfaces. Alternatively, LAM maybe caused by continuous increased cell differentiation and the production of calcium micromotion of the implant due to the failure of early bone phosphate deposits in vitro, when implanted in bone charge growth at the bone-implant interface, however this seems modified PC materials stimulated the production of fibrous unlikely as such a response might be expected equally with tissue and areas of loosely associated matrix around the PC0, PC6 and SS implants; a trend that was not observed in implant. Despite this, non-charged phosphorylcholine the data. The in vivo results for hydroxy-apatite coated polymers were tolerated in the osseous environment and implants clearly accords with the clinical utility of this could be used to coat orthopaedic devices, where bone material; however, the relatively poor outcome for PC0, bonding is not required, for the delivery of antibiotics in PC6 and PC20 indicate that these materials would not be order to reduce surgical site infection. suitable as coatings for bone bonding applications. Acknowledgements The authors would specifically like to thank Phosphorylcholine-based coatings may however show Biocompatibles UK Ltd for supplying the polymers and for financial greater stability in vivo and could be useful as drug delivery and academic support, without which this research would not be vehicles in situations where bone bonding is not required or possible. Financial assistance from the Engineering and Physical be used where inhibiting tissue bonding would be useful for Science Research Council and Biotechnology and Biological Sciences Research Council UK is also acknowledged. example with fracture fixation plates. We do not know whether PC coating was lost from the surface of the pins due Compliance with ethical standards to mechanical damage during insertion. This could be investigated in cadaveric models by inserting coated pins into Conflict of interest The authors declare that they have no competing the tibia, breaking open the bone and determining the uni- interests. formity of coating on the retrieved pins. The results presented in this paper describe the response of bone cells and bone to PC and modified PC surfaces and also demonstrate that in vitro experiments are not always References accurate predictors of the more complex in vivo response. It 1. Hench LL. Biomaterials: a forecast for the future. Biomaterials. was hypothesised that due to the increased proliferation and 1998;19:1419–23. differentiation of HOB cells on PC20, and the evidence that 2. Vander Sloten J. Materials selection and design for orthopaedic these cells were undergoing mineralisation (Figs. 3–6), that implants with improved long-term performance. Biomaterials. PC20 coated implants would stimulate an increase in bone- 1998;19:1455–9. 3. Hench LL. The challenge of orthopaedic materials. Curr Orthop. bonding compared with the lower and non-charged PC 2000;14:7–15. materials, whereby osteoblasts are actively attracted to the 4. Rose FRAJ, Oreffo ROC. Bone tissue engineering: hope v’s hype. surface of the polymer and mineralise to produce a tight Biochem Biophys Res Com. 2002;292(1):1–7. implant-bone interface. However, in this limited scope 5. Ishihara K, Nomura H, Mihara T, Kurita K, Iwasaki Y, Naka- bayashi N. Why do phospholipids polymers reduce platelet in vivo experiment, it was shown that this was not the case. adsorption? J Biomed Mater Res. 1998;39:323–30. Ideally, a full time-course experiment would be needed to 6. Iwasaki Y, Mikami A, Kuria K, Yui N, Ishihara K, Nakabayashi properly address the evolution of the biological response; N. Reduction of surface-induced platelet activation on phospho- here we were limited to one time point and chose 14 weeks lipid polymers. J Biomed Mater Res. 1997;36(4):508–15. 144 Page 12 of 13 J Mater Sci: Mater Med (2017) 28:144 7. Goreish HH, Lewis AL, Rose S, Lloyd AW. The effect of 23. Singh G, Reichard T, Hameister R, Awiszus F, Schenk K, phosphorylcholine-coated materials on the inflammatory response Feuerstein B, Roessner A, Lohmann C. Ballooning osteolysis in and fibrous capsule formation: in vitro and in vivo observations. J 71 failed total ankle arthroplastie: is hydroxyapatite a risk factor? Biomed Mater Res. 2004;68(A):1–9. Acta Orth. 2016;87(4):401–5. 8. Gobeil F, Juneau C, Plante S. Thrombus formation on guide wires 24. Voelker DR. Bridging gaps in phospholipid transport. Trends during routine PTCA procedures. A scanning electron micro- Biochem Sci. 2005;30(7):396–404. scopic evaluation. Can J Cardiol. 2002;18:263–9. 25. Wu LNY, Genge BR, Wuthier RE. Analysis and molecular 9. Whelan DM, van der Giessen WJ, Krabbendam SC, van Vilet EA, modelling of the formation, structure and activity of the Verdouw OD, Serruys PW, van Beuekom HMM, Biocompat- phosphatidylserine-calcium-phosphate complex associated with ibility of phosphorylcholine coated stents in normal porcine cor- biominerlization. J Biol Chem. 2008;283(7):3827–38. onary arteries.Heart. 2000;83(3):338–45 26. Merolli A, Santin M. Role of phosphatidyl-serine in bone repair 10. De Somer F, van Belleghem Y, Caes F, Francois K, Arnout J, and its technological exploitation. Molecules. 2009;14:5367–81. Bossuyt X, Taeymans Y, van Nooten G. Phosphorylcholine 27. Gong YK, Luo L, Petit A, Zukor DJ, Huk OL, Antoniou J, coatings offers natural platelet preservation during cardio- Winnik FM, Mwale F. Adhesion of human U937 macrophages to pulmonary bypass. Perfusion. 2002;1:39–44. phosphorylcholine-coated surfaces. J Biomed Mater Res A. 11. Lemp MA, Caffery B, Lebow K, Lembach R, Park J, Foulks G, 2005;71(1):1–9. Hall B, Bowers RJW, McGarvey S, Young G, Omafilcon A. 28. Shelton RM. Protein adsorption at the interface between charged (Proclear) soft contact lenses in dry eye population. CLAO J. polymer substrata and migrating osteoblasts. Biomaterials. 1999;25(1):40–7. 1988;9:24–9. 12. Kyomoto M, Moro T, Konno T, Takadama H, Yamawaki N, 29. Bet MR. Cell adhesion and cytoxity studies over polyanionic Takatori Y, Nakamura K, Ishihara K. Enhanced wear resistance of collagen surfaces with variable charge and wettability. Bioma- modified cross-linked polyethylene by grafting with poly (2- terials. 2003;24:131–7. methacryloyloxyethyl phosphorylcholine). J Biomed Mater Res 30. Lee JH, Khang SS, Lee JW, Lee HB. Interaction of proteins and A. 2007;82(1):10–7. cells on functional group gradient surfaces. Molec Symp. 13. Takatori Y, Moro T, Ishihara K, Kamogawa M, ODa H, 1997;118:571–6. Umeyama T, Kim YT, Kyomoto M, Tanaka T, Kawaguchi H, 31. Kishida A. Cell behaviour on polymeric substances with non-ionic Tanaka S. Clinical and radiographic outcomes of total hip repla- and ionic monomers. Biomaterials. 1991;12:786–92. cement with poly(2-methacryloyloxyethyl phosphorylcholine)- 32. Rose SF, Lewis AL, Hanlon GW, Lloyd AW. Biological grafted highly cross-linked polyethylene liners: three-year results responses to cationically charged phosphorylcholine-based mate- of a prospective consecutive series. Mod Rheumatol. 2015;25 rials in vitro. Biomaterials. 2004;25:5125–35. (2):286–91. 33. Palmer RP, Lewis AL, Kirkwood LC, Rose SF, Lloyd AW, Vick 14. Kyomoto M, Iwasaki Y, Moro T, Konno T, Miyaji F, Kawaguchi TA, Stratford PW. Biological evaluation and drug delivery H, Takatori Y, Nakamura K, Ishihara K. High lubricious surface applications of cationically modified phospholipid polymers. of cobalt-chromium-molybdenum alloy prepared by grafting poly Biomaterials. 2004;19:4785–96. (2-methacryloyloxyethyl phosphorylcholine). Biomaterials. 34. Susin C, Qahash M, Hall J, Sennerby L, Wikesjö UM. Histolo- 2007;28(20):3121–30. gical and biomechanical evaluation of phosphorylcholine-coated 15. Kyomoto M, Moro T, Iwasaki Y, Miyaji F, Kawaguchi H, titanium implants. J Clin Peridontol. 2008;35(3):270–75. Takatori Y, Nakamura K, Ishihara K. Superlubricious surface 35. Christenson RH. Biochemical markers of bone metabolism: an mimicking articular cartilage by grafting poly(2-methacryloylox- overview. Clin Biochem. 1997;30(8):573–93. yethyl phosphorylcholine) on orthopaedic metal bearings. J 36. Tortora GJ, Grabowski S, editors. Principles of Anatomy and Biomed Mater Res A. 2009;91(3):730–41. Physiology. 8th ed. California: John Wiley & Sons Inc; 1996. p. 16. Kyomoto M, Moro T, Yamane S, Hashimoto M, Takatori Y, Ish- 143–50. ihara K. Poly (ether-ether-ketone) orthopaedic bearing surface 37. Mark SC, Odgren PR. Structure and development of the skeleton. modified by self-initiated surface grafting of poly(2-methacryloy- In: Bilezikium JP, Raisz LG, Rodam GA, editors. Principles of loxyethyl phosphorylcholine). Biomaterials. 2013;34(32):7829–39. Bone Biology. 2nd ed. New York: Academic Press; 2000. p. 17. Hench LL, Wilson J, editors. Clinical performance of skeletal 3–15. prostheses. London: Chapman & Hall; 1996. p. 12–7. 38. Lewis AL, Berwick J, Davies MC, Robberts CJ, Wang J-H, Small 18. Morrey BF, editor. Biological, material and mechanical con- S, Dunn A, O’Bryan V, Redman RP, Jones SA. Synthesis and siderations of joint replacement. New York, NY: Raven Press; characterisation of cationically modified phospholipid polymers. 1990. p. 22–47. Biomaterials. 2004;25:3099–108. 19. Sun LM, Berndt CC, Gross KA, Kncuk A. Material fundamentals 39. Sabokbar A, Millett PJ, Myer N, Rushton N. A rapid, quantitative and clinical performance of plasma-sprayed hydroxyapatite coat- assay for measuring alkaline phosphatase activity in osteoblastic ings: a review. J Biomed Mater Res A. 2001;58(5):570–92. cells in vitro. Bone Mineral. 1994;27:57–67. 20. Sandiford N, Doctor C, Rajaratnam SS, Ahmed S, East DJ, Miles 40. Murshed M, Harmey D, Millan JL, McKee MD, Karsenty G. K, Butler-Manuel A, Shepperd JAN. Primary total hip replace- Unique co-expression in osteoblasts of broadly expressed genes ment with a Furlong fully hydroxyapatite-coated titanium alloy accounts for the spatial restriction of ECM mineralisation to bone. femoral component: results at a minimum follow up of 20 years. Genes Dev. 2005;19:1093–104. Bone Joint J. 2013;95B(4):467–71. 41. Allen M, Brett F, Millett P, Rushton N. The effects of particulate 21. Gross KA, Walsh W, Swarts E. Analysis of retrieved polyethylene at a weight bearing bone-implant interface: a study hydroxyapatite-coated prostheses. J Thermal Spray Tech. 2004;13 in rats. J Bone Joint Surg [B]. 1996;82-B:32–7. (2):190–9. 42. Clarke SA, Brooks RA, Rushton N. Testing bone substitutes in a 22. Tonino A, Oosterbos C, Rahmy A, Therin M, Doyle C. small animal model of revision arthroplasty. J Mater Sci Mater Hydroxyapatite-coated acetabular components – Histological and Med. 2002;13:829–36. histomorphometric analysis of six cups retrieved at autopsy 43. Clarke SA, Brooks RA, Lee PTH, Rushton N. The effect of between three and seven years after successful implantation. J osteogenic growth factors on bone growth into a ceramic filled Bone Joint Surg [Am]. 2001;83A:817–25. defect around an implant. J Ortho Res. 2004;22:1016–24. J Mater Sci: Mater Med (2017) 28:144 Page 13 of 13 144 44. Tang L, Eaton J. Mechanism of acute response to biomaterials. 53. Martin RI, Brown PW. Aqueous formation of hydroxyapatite. J Cells Mat. 1994;4(4):429–36. Biomed Mater Res. 1997;35(3):299–308. 45. Anderson JM. The cellular cascades of wound healing. In: Davies 54. Neo M, Kotani S, Nakamura T, Yamamuro T, Ohtsuki C, Kokubo JE, editor. Bone Eng. Toronto: em2 Incorporated; 2000. p. 81–93. T, Bando Y. A comparative study of ultrastructure of the interface 46. Audrade JD, Hlady V. Protein adsorption and materials bio- between four kinds of surface-active ceramic and bone. J Biomed compatibility: a tutorial review and suggested hypothesis. Ad Pol Mater Res. 1992;26:1419–32. Sci. 1987;79:1–59. 55. Neo M, Nakamura T, Ohtsuki C, Kokubo T, Yamamuro T. 47. Jenny CR, Anderson JM. Adsorbed serum proteins responsible for Apatite formation on three kinds of bioactive materials at an early surface dependent human macrophage behaviour. J Biomed stage in vivo: a comparative study by TEM. J Biomed Mater Res. Mater. 2000;49(4):435–47. 1993;27:999–1006. 48. Lewis AL. Phosphorylcholine materials and their use in the pre- 56. Kokubo T, Kim HM, Kawashita M, Nakamura T. What kinds of vention of biofouling. Col Surf B Bioint. 2000;18:261–75. materials exhibit bone-bonding? In: Davies JE, editor. Bone 49. Lewis AL, Driver M. Blending in with the body. J Chem Ed. Engineering. Toronto: em2 Incorporated; 2000. p. 190–194. 2002;79(3):323–6. 57. Dicarlo EF, Bullogh PG. The biological responses to orthopaedic 50. Howlett CR, Evans MDM, Walsh WR, Johnson G, Steele JG. implants and their wear debris. Clin Mater. 1992;9:235–6. Mechanisms of initial attachment of cells derived from human 58. Moroni A, Orienti L, Stea S, Visentin M. Improvement of the bone to commonly used prosthetic materials during cell-culture. bone-pin interface with hydroxyapatite coating: an in vivo long Biomaterials. 1994;15(3):213–22. term study. J Ortho Trauma. 1996;10(4):236–42. 51. Billiard J, Moran RA. Transcriptional profiling of human osteo- 59. Martin EA, editor. Concise colour medical dictionary. 2nd ed. blast differentiation. J Cellular Biochem. 2003;89(2):389–400. Oxford: Oxford University Pres; 1998. 52. Stein GS, Lian JB. Molecular mechanisms mediating prolifera- 60. Dagtekin G, Schiffer R, Klein B, Jahnen-Dechent W, Zwadlo- tion, differentiation interrelationships during progressive devel- Klarwasser G. Modulation of angiogenic functions in human opment of the osteoblast phenotype. Endocrine Rev. 1993;14 macrophages by biomaterials. Biomaterials. 2003;24:3395–401. (4):424–42.

Journal

Journal of Materials Science: Materials in Medicine – Springer Journals

Published: Aug 17, 2017

There are no references for this article.