Grafting Polymer Brushes by ATRP from Functionalized Poly(ether ether ketone) Microparticles

Abstract

Poly(ether ether ketone) (PEEK) is a semi-crystalline thermoplastic with excellent mechanical and chemical properties. PEEK exhibits a high degree of resistance to thermal, chemical, and bio-degradation. PEEK is used as biomaterial in the field of orthopaedic and dental implants; however, due to its intrinsic hydrophobicity and inert surface, PEEK does not effectively support bone growth. Therefore, new methods to modify PEEK’s surface to improve osseointegration are key to next generation polymer implant materials. Unfortunately, PEEK is a challenging material to both modify and subsequently characterize thus stymieing efforts to improve PEEK osseointegration. In this manuscript, we demonstrate how surface-initiated atom transfer radical polymerization (SI-ATRP) can be used to modify novel PEEK microparticles (PMP). The hard core-soft shell microparticles were synthesized and characterized by DLS, ATR-IR, XPS and TEM, indicating the grafted materials increased solubility and stability in a range of solvents. The discovered surface grafted PMP can be used as compatibilizers for the polymer-tissue interface.

Graphical Abstract

Introduction

Hard tissue implants can be comprised of metal or synthetic polymers. There are three key factors that are considered in hard-tissue implant design 1) ability to support bone growth (osseointegration), 2) radiolucency to support imaging and evaluation of bone on-growth and 3) bone mechanical compatibility by matching tensile strength to that of bone and avoid fracture of the bone adjacent to the implant.^1–3 Metallic implants, primarily comprised of titanium, support bone on-growth but are not radiolucent and have tensile strength far exceeding those of bone.^4,5 Polymer implants, primarily comprised of poly(ether ether ketone) (PEEK), do not support bone on-growth but are radiolucent and have mechanical properties closer to native cortical bone.^6–8 If PEEK’s osseointegration properties could be improved, an important next generation implant material would be realized.

PEEK has many desirable characteristics, including its high resistance to thermal, chemical, and bio-degradation, enabling its use in a variety of applications under harsh mechanical or chemical conditions. These inert properties also give rise to PEEK’s limited capacity to support bone on-growth due to a lack of surface charge or chemical functional groups to promote protein or cell binding. This is contrasted by titanium-based implants which possess a hydrophilic and ionic surface. Such an environment is essential for adhesion, growth, and differentiation of osteoblasts which then leads to successful osseointegration.^9,10 However, titanium has several inherent drawbacks. The radiopacity of the titanium device restricts the postoperative imaging of the implant site making it challenging to assess the bone on-growth process.⁴ Additionally, titanium’s elasticity far exceeds that of human bones- leading to device induced bone fracture.⁵ Lastly, titanium has a density of 4.506 g/cm³, which is approximately three times higher than average human bone density. Comparably, PEEK has an elasticity modulus and density more closely paired to human bone tissue. Since PEEK is an intrinsically hydrophobic inert material, methods to enhance PEEK’s surface properties can lead to improved biological outcomes. Several attempts at improving PEEK’s properties have recently been explored including etching pores and channels into PEEK surface by acid-etching/leaching techniques or coating a layer of metal (i.e. titanium) onto its surface. Such processes enhance the performance of PEEK to a certain degree, but also lead to increased cost and tedious procedures in manufacturing as well as the delamination of the metal coating layer.^11–17 Recently, Averick and coworkers have reported a mild surface modification method by reacting the surface ketone group with hydrophilic hydrazine via oxime chemistry.¹⁸ After this mild modification process, the hydrophilicity of the modified PEEK surface was greatly improved. This resulted in a higher level of both alkaline phosphatase and calcium deposition in mineralization study, indicating an escalated affinity of PEEK surface and osteoblasts. This finding emphasized the importance of a hydrophilic surface towards the promotion of osteoblast adhesion and proliferation, and eventually the success of the osseointegration process.^1,10

Thus, methods to improve PEEK’s osseointegration through chemical derivation and increased surface area are critical to create new PEEK-based implant materials. We postulate surface modified PEEK microparticles can provide the solution to the vexing challenge of PEEKs resistance to osseointegration. We envision that these surface modified PEEK microparticles can be extruded as a new implant material with precision engineered properties. In this manuscript, we report the first of our efforts into the creation of this new implant base material. The field of PEEK microparticles has been underexplored due to PEEKs resistance to dissolution in common solvent, resistance to grinding, and chemical resistance. The existing reports demonstrated poor biocompatibility results of PEEK microparticles^19–21; however, the components that introduced cytotoxicity originated from contaminants during the manufacturing process of PEEK microparticle instead of PEEK itself.²² Herein, we propose a new method of fabricating PEEK microparticle with surface modification by surface initiated atom transfer radical polymerization (SI-ATRP).^23–28 We hypothesize the new PEEK microparticle – grafted polymer complex could potentially exhibit a great improvement as an implant material.

In this work, PEEK pellets were first dissolved in dichloroacetic acid at 185 ˚C and precipitated into a 0.3 wt% poly(vinyl alcohol) solution to obtain PEEK microparticles with a ~1 μm diameter. This size is appropriate for eventual compression into polymer composites for use in implants.²⁹ This was followed by the reduction of carbonyl group to a hydroxyl group and esterification with an ATRP initiator.³⁰ The PEEK microparticles were used as initiators in the grafting-from of a small polymer library. Among the polymer grafted PEEK microparticle samples, the PEEK-g-poly(2-(dimethylamino)ethyl methacrylate) (PEEK-g-PDMAEMA) microparticle demonstrated the highest solubility in methanol, the most stable diameter, and micro-structures as determined by DLS and TEM - pointing to a possible method for processing and fabrication of implant devices.

Results and discussion

PEEK microparticles were fabricated using the micro-precipitation method with necessary procedural adjustments. This method is traditionally employed to prepare polylactide microparticles³¹ and has not yet been adapted for PEEK or other thermoplastic semi-crystalline polymers. Due to stability and chemical resistance, PEEK is essentially insoluble in common organic solvents. Fortunately, PEEK is soluble in chlorinated acetic acids such as dichloroacetic acid (DCA) at 185 °C.³² To prepare PEEK microparticles, PEEK pellets were heated in DCA until complete dissolution and a viscous clear, red-brown solution was obtained. An aqueous solution of 0.3 wt% poly(vinyl alcohol) solution was prepared in 400 mL ultrapure water. The PEEK microparticles (PMP) were then formed through dropwise addition of the PEEK solution into the PVA solution with rapid stirring. Upon addition, the PVA solution became cloudy which corresponded to the appearance of a beige-gray precipitate. The PMP was then separated from the supernatant by centrifugation and dried in a vacuum oven. (Scheme 1) Although the solubility of PMP in methanol was limited, the dissolved PMP could be effectively suspended to allow for particle size measurement by Dynamic Light Scattering (DLS), the PMP particle size was d= 970 nm, which indicated the successful formation of microparticles. (Figure S1)

Scheme 1.

Micro-precipitation process of PEEK microparticles (PMP)

To enable the rapid screening of how surface functionality impacts PEEK stability, we used surface initiated ATRP. SI-ATRP gives access to a large range of possible surface modifications and functionalities. In this manuscript, we studied how polymer grafting effects PMP solvent stability. To achieve this goal, we introduced ATRP initiators onto the surface of PMP using a two-step process. In the first step, the carbonyl groups on the PMP surface were reduced to hydroxyl groups by sodium borohydride in DMSO at 120 °C. Attenuated total reflection – infrared (ATR-IR) spectroscopy shows the appearance of O-H stretch at 3380 cm⁻¹ after reduction. The appearance of the O-H stretch proves the transformation of PMP surface ketone groups into hydroxyl groups; the new material is designated as PMP-OH. The immobilization of ATRP initiator onto PMP-OH was accomplished by reacting 2-bromoisobutyryl bromide (2-BiBr) with hydroxyl groups on PMP-OH, in the presence of the triethylamine catalyst. Upon completion of the esterification, the solid material was separated by gravity filtration, washed, and dried under N₂ flow. ATR-IR was used to evaluate the new material and the presence of a new peak at 1733 cm⁻¹ confirmed the formation of an ester bonds on the PMP surface. The PMP with surface modified ATRP initiators was prepared and is referred to as PMP-iBBr.

After PMP-iBBr was successfully prepared, various polymers were grafted from the PMP surface to evaluate the best conditions for SI-ATRP grafting approach using ARGET, ICAR, and photo-mediated methods. As shown in Table 1, four monomers: methyl methacrylate (MMA), butyl acrylate (BA), 2-hydroxyethyl methacrylate (HEMA) and 2-(dimethylamino)ethyl methacrylate (DMAEMA) were tested. In each reaction trial, 30 mg of PMP-iBBr was added; 2-hydroxyethyl 2-bromoisobutyrate (OH-EBiB) was used as a sacrificial initiator for better control of target DP_n as well as simplifying analysis of polymers by GPC. Results of ARGET ATRP of MMA and BA demonstrated the formation of polymers with predefined M_n and dispersities as low as 1.19 and 1.06 for MMA and BA, respectively (Table 1, entries 1 and 2). The washed-and-dried PMP-g-PMMA complex was measured by ATR-IR spectroscopy, the stark increase at 1727 cm⁻¹ indicated the ester bonds in MMA have extensively covered the surface of PMP (Figure S5). HEMA was polymerized via photoATRP and the conversion reached 98% after 5 hours, with M_n,GPC matching the theoretical value and low dispersity at 1.23 (Table 1, entry 3).

Table 1.

Grafting from PMP-iBBr with different ATRP methods on various monomers ^a)

Entry	Monomer	ATRP method	Reducing Agent	Time (h)	Conv.b) (%)	Mn, Th c)× 10−3	Mn, GPC d)× 10−3	Ɖ d)
1	MMA	ARGET e)	AAc	21	95	19.2	19.9	1.19
2	BA	ARGET e)	AAc	26	53	13.8	13.1	1.06
3	HEMA	Photo f)	UV light	5	98	25.4	28.5	1.23
4	DMAEMA	ICAR g)	AIBN	6	73	23.2	18.7	1.31

^a)30 mg of PMP was added in each entry; [Monomer]₀:[OH-EBiB]₀:[CuBr₂]:[TPMA] = 200:1:0.2:1, V_total = 5 mL, [Monomer] = 20 vol %, solvent = DMF

^b)Conversion determined by ¹H NMR

^c)M_n,Th calculated based on the equation M_n,Th = (M_monomer × [Monomer]₀ × Conv.%)/[OH-EBiB]₀ + M_OH-EBiB

^d)M_n,GPC and Ɖ were determined with free polymers initiated from OH-EBiB by DMF GPC, based on PMMA as calibration standards

^e)[AAc]₀ = 2 × [CuBr₂]

^f)365 nm irradiation (2.85 mW/cm²) at 25 °C.

^g)[AIBN]₀ = 0.25 × [OH-EBiB]₀, T =65 °C

Prior work by our group and others demonstrated the presence of a tertiary amine improved cell growth and adhesion.³³ In this report, we also used PDMAEMA. PDMAEMA coated surface have improved adhesion and proliferation for osteoblasts. PDMAEMA should enhance osseointegration by the introduction of a partial positive charge to the grafted surface promoting interactions with the negatively charged proteoglycans in the extracellular matrix and on cell surfaces at physiological pH (7.4). PDMAEMA was grafted from the PMP surface via ICAR ATRP. The monomer conversion reached 73% after 6 hours with relatively low dispersity at 1.31 and M_{n, GPC} slightly lower than the theoretical value (Table 1, entry 4). This is plausibly due to a small portion of new chains that were generated in-situ by free radicals from AIBN. ATR-IR spectroscopy of PMP-g-PDMAEMA complex displayed the new absorption peaks of sp³ C-H stretch at 2950 cm⁻¹, ester C=O stretch at 1723 cm^−1, and C-N stretch at 1180 cm⁻¹ (Figure 1). The appearances of these peaks indicated the successful grafting of PDMAEMA from the PMP surface.

Figure 1.

Overlay of ATR-IR spectroscopy of PMP-iBBr (black) and PMP-g-PDMAEMA (red).

Thermogravimetric analysis (TGA) of PMP-g-PDMAEMA exhibited a two-step decomposition process. The first sharp drop occurred at 280~350 °C, corresponding to the decomposition of PDMAEMA; while the second drop occurred at ~550 °C which is related to the decomposition of PEEK³⁴ (Figure 2a). Hence, the TGA measurement indicated the covalent bonding between PEEK and PDMAEMA. X-ray photoelectron spectroscopy (XPS) analyses were also employed on PEEK microparticle samples at different stages of modification. From PMP to PMP-g-PDMAEMA, the increased intensity of the peak at 288.6 eV also confirmed the successful immobilization of ATRP initiator, as well as grafting of PDMAEMA.

Figure 2.

a) TGA of PMP-g-PDMAEMA and b) XPS analysis of PMP, PMP-iBBr, and PMP-g-PDMAEMA samples. (C1s scan)

The manufacturing of medical grade implants and composites from microparticles requires stable suspension of materials. Generally, these composites are processed into a membrane through dry-cast process, which requires the volatile solvents such as methanol.^35,36 Thus, the solubility of PMP-g-PDMAEMA in methanol, a common low boiling point and processable solvent, was studied. A stock solution was made of 30 mg washed-and-dried PMP-g-PDMAEMA in 7.5 mL of methanol. The mixture was sonicated for 10 minutes and a stable milky suspension was obtained. 1 mL of original stock solution was diluted ten-fold with methanol. The mixture was sonicated for 10 minutes, followed by centrifugation at 4,000 rpm for 10 minutes. A translucent “initial supernatant” was obtained with minimal beige precipitate. The precipitate was then redissolved with 10 mL methanol and repeated the sonication – centrifugation procedure. Again, the supernatant, that was redissolved for the first time, appeared to be as translucent as the initial supernatant. The same process was then repeated for an additional three times, yielding a total of five supernatant samples labeled as “Initial supernatant” and “Redissolved #1-#4”(Figure 3a). Considering phosphate buffer saline (PBS, pH=7.4) similarity to physiological conditions, same procedure was done using a stock solution of 30 mg washed-and-dried PMP-g-PDMAEMA in 7.5 mL of PBS (Figure S10).

Figure 3:

a) Procedure of redissolving PMP-g-PDMAEMA in methanol; b)The TEM images of PMP-g-PDMAEMA samples prepared by drop-casting a methanol dispersion of PEEK particles onto a Cu-based TEM grid. The grid was dried in air before loaded to the sample chamber of TEM. c) DLS measurements for 0.67±0.02 mg/mL PMP-g-PDMAEMA in methanol at 25 °C

Measured by DLS, the particle size of PMP-g-PDMAEMA in each supernatant in methanol was around 680 nm (Figure 3c). The consistency in particle size revealed the stable structure of the PDMAEMA-grafted PMP. One plausible explanation for the decrease of particle size from 970 nm to 680 nm could be the grafted PDMAEMA had assisted individual PMP to stabilize and induced disaggregation from PMP clusters.

All of the five supernatants were allowed to stand for two weeks and no precipitation was observed. This indicated the PMP-g-PDMAEMA was fully solubilized in methanol at its maximum solubility, at the concentration of 0.67±0.02 mg/mL. The actual images of PMP-g-PDMAEMA were recorded by transmission electron microscopy (TEM). As shown in Figure 3b, although the particles are randomly shaped, their apparent sizes mostly agree with the size determined by DLS.

Grafting density is a critical feature to consider when modifying surfaces for biological applications. This parameter directly affects factors that are crucial in biological applications such as water contact angle³⁷ and cellular immune responses to implants.³⁸ We herein have used BET surface area analysis, and thermal treatment analysis to measure the grafting density of PMP-g-PDMAEMA (Figure S8–S9). The calculated grafting density of as synthesized PMP-g-PDMAEMA is 0.215 chains/nm².

To study the biocompatibility of PEEK microparticles, polymer-grafted PEEK microparticles were cultured with HEK293 cells and their proliferation was assessed over three days. 10 mg/ml PEEK microparticles with different surface modifications were added to HEK293 cells and viability was assessed every 24 hours for three days using CyQUANT assay as shown in Figure 4a. The PEEK microparticles treated cells showed no significant difference compared to the positive control, which was a regular tissue culture plastic plate. This data suggests that microparticles synthesized by the novel fabrication process do not negatively affect cell survival or proliferation. Further, we tested the osteoconductive properties of these PEEK microparticles in the presence of recombinant human bone morphogenetic protein-2 (BMP2). This was assessed by evaluating the upregulation of an early osteogenic differentiation marker – alkaline phosphatase (ALP) in C2C12 cells as described previously.¹⁸ ALP assay is a gold standard in assessing early phase osteoblastic differentiation. The data shown in the Figure 4b & 4c suggests that in presence of BMP2, polymer-grafted PEEK does not suppress osteoblastic differentiation in C2C12 cells. Additionally, PMP did not see traditional limitations to PEEK support of cell on-growth implying that microparticle embedded particles could improve PEEK’s cell response properties. The indirect quantification of ALP staining suggests that the upregulation of ALP in C2C12 cells was not significantly different from C2C12 cells cultured in tissue culture coated plates in presence of BMP2.

Figure 4:

In vitro studies of polymer-grafted PEEK microparticles. a) Biocompatibility assays illustrating the effect of 72 h treatment of HEK293 cells with 10 mg/ml polymer-grafted PEEK microparticles normalized to no treatment control. Data presented is the average of three replicates, ns: not significant vs no treatment control. b)Osteoinductive properties of polymer-grafted PEEK microparticles evaluated in the presence of BMP2 using C2C12 cells. Quantification of ALP expression on different treatments, bars indicate % Mean ± SEM (n=3), ns: not significant vs control. c) Microscopy images of C2C12 cells stained for ALP after 72 hours in cell culture in presence of 10 mg/ml PEEK microparticles and ±100 ng/ml BMP2. Intensity of blue color correlates with intracellular ALP expression, scale bar = 200 μm.

Conclusion

In conclusion, PEEK microparticles were first fabricated by micro-precipitation and grafted with various polymers for further stability. The use of PMP enabled facile characterization of the surface modified PEEK overcoming issues seen in surface modified PEEK materials. Amongst for PMP-polymer hybrid nanoparticles, PDMAEMA was studied in more details due to its reported enhancement towards cell adhesion and differentiation. PMP-g-PDMAEMA has exhibited increased solubility in a volatile solvent like methanol, and narrowly distributed particle sizes determined by DLS and TEM. This type of hard core-soft shell microparticle complex can be potentially used in actual implant devices or composites in the future.

Scheme 2.

Modification of PMP surface with ATRP initiator and grafting polymers via SI-ATRP

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.