1.     Research title

The effect of bisphosphonate-modified poly(L-lactic acid) micro-fibers on the stiffness, bending strength and toughness of calcium phosphate cements.

2.     Purpose of research

The main goal of this research is to research if bisphosphonate modification of poly(L-lactic acid) micro-fibers has a significant effect on the stiffness, bending strength and toughness of calcium phosphate cements.

4.     Abstract

The current generation of synthetic bioceramic bone substitutes such as calcium phosphate cements (CPCs) has poor mechanical properties, which limits the use of these materials to non-load-bearing skeletal sites. These poor mechanical properties can be overcome by adding poly(L)-lactic acid (PLLA) fibers, as these are biodegradable polymers with relatively good strength and biocompatibility. The adhesion between the PLLA fibers and the CPCs is, however, not optimal due to the difference in polarity between these two types of materials. PLLA is a hydrophobic polymer, whereas CPC is a hydrophilic ceramic. Therefore, this study researched the effect of surface modification of PLLA fibers on the mechanical properties of fiber reinforced CPCs. More specifically, the goal of this study was to research if surface modification of poly(L-lactic acid) micro-fibers with calcium-binding bisphosphonate (BP) groups using a polydopamine (PDA)-based immobilisation method had an effect on the stiffness, bending strength and toughness of calcium phosphate cements. To this end, CPCs with or without PLLA fibers were prepared. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP). The fibers were cut mechanically as well as manually. All samples were tested mechanically using a three-point bending test. We hypothesized that the stiffness, bending strength and toughness of calcium phosphate cements reinforced with surface-modified PLLA fibers would be significantly higher than the cements reinforced with unmodified poly(L-lactic acid) microfibers. Our results showed that the fiber modification had no statistically significant effect on either the stiffness, bending strength or toughness of the CPCs reinforced with mechanically or manually cut PLLA fibers. Summarizing, the added value of surface modification of the PLLA fibers using BP coatings was not confirmed in the current study.

6.     Research justification

6.1       Background

6.1.1 Bone defects and bone augmentation

In the oromaxillofacial area bony defects may exist due to infections, tumors, cysts, fractures, periodontal disease or congenital malformation. These defects may occur at different locations such as the marginal bone, intra-alveolar bone and/or at the furcations of teeth. Applying bone substitution materials preserves the alveolar ridge, prevents resorption of bone and provides sufficient bone for immediate or subsequent implant placement.

Different treatments have been developed to increase bone volume; for example bone grafts (autogenous and allogenic block grafts) and barrier membranes for guided bone regeneration. The use of autogenous bone grafts is still the gold standard for the reconstruction of large bone defects. Autologous bone possesses optimal osteoconductive and osteogenic properties. It can be retrieved from several donor sites: crista iliaca, the ribs, the skull, the tibia and several sites in the oral cavity[1].

However, several other materials are available for the use of bone substitution such as animal-derived bone substitutes and synthetic bone substitutes. One widely used xenogenic graft is an animal-derived bone substitute such as Bio-Oss, which is a deproteinized sterilized bovine bone. No local T-cell or B-cell inflammatory reactions have been reported with the use of this inorganic bone. The material has shown to be biocompatible and osteoconductive and is widely used in the medical sector. The disadvantage of the use of these grafts is the extensive procedures in order to obtain and place the materials [2]. This is why synthetical bone substitutes have been on the rise for many years. These substitutes can simplify the process of retrieving the donor bone and make the procedure less invasive.

6.1.2 Materials

Ideally a synthetic bone substitute is biocompatible, supports new bone formation, shows minimal fibrotic reaction and can be remodelled. The first main category of bio-ceramics is the bioactive glass. This is a hard, non-porous, material and has osteointegrative and osteoinductive properties. This glass is based on a glass-forming SiO₂ network. The key compositional features which are responsible for the bioactivity are the low SiO₂ content, the high Na₂O and CaO content and the high CaO/P₂O₅ ratio[3]. A mechanically strong bond is formed between the glass and bone through hydroxyapatite crystals. The main disadvantage of these materials is their poor mechanical strength.

The second main category of bio-ceramics is the calcium phosphate ceramics (CaP ceramics). CaP ceramics are synthetic materials and have been used since the 1980s [4]. These materials do not have osteoinductive or osteogenic properties, but are osteoconductive. The two most widely used calcium phosphate based bio-ceramics are hydroxyapatites (HAP) and β-tricalcium phosphates (β-TCP). Hydroxyapatite (HA) is a very biocompatible ceramic, which is known to be brittle and undergoes a slow resorption process. β-tricalcium phosphate is known to have high osteoconductivity and a fast resorption process [5]. A particularly interesting class of CaP ceramics are the CaP cements; they are injectable pastes, which harden after application into defects.

6.1.3 Calcium phosphate cements (CPC)

Calcium phosphate cements have been researched since the 1980s and are often used as bone substitutes. These materials harden in vivo and have strong similarities with the mineral phase of bone and teeth. They are osteoconductive, which induces growth of bony tissue in order to repair damaged bone. One major advantage is that upon mixing the calcium phosphate powder with a liquid phase, they become mouldable into every shape and size and can easily be injected   in the site of bone damage. Once placed it is desirable that the material is resorbed and replaced by regenerating bone.

Currently, several formulations have been established which can transform into either hydroxyapatite (HA:Ca₁₀(PO₄)₆(OH)₂) or brushite (DCPD: CaHPO₄.2H₂O) cements. In comparison to the brushite cements, the apatitic cements have a higher mechanical strength. It also takes longer before the apatitic cements are dissolved [6]. Another advantage of the CPC is the ability to be injected, which is characterized by the capacity of a paste to be forced through a needle without causing demixing[7]. Surgery is minimally invasive and complex shaped defects will be filled adequately [8]. Unfortunately, the mechanical properties of CPCs are poor in terms of strength, ductility, toughness and fatigue resistance. This limits the use of CPCs to non-stress-bearing sites [9]. This is why one strategy to reinforce these materials is by addition of fibers in the cement matrix.

6.1.4 Mechanism of fiber reinforcing cements

Due to their brittle nature, CPCs typically exhibit sudden fractures without any previous plastic deformation. Therefore the load required to fracture the cement is very low. One way to increase the mechanical strength and fracture toughness is by reinforcing the CPCs with fibers. This idea is taken from civil engineering where they use fibers to reinforce concrete.

In CPCs failure will occur due to the fast propagation of a single crack while in fiber reinforced CPCs the composite will bear increasing loads after the matrix begins to crack. This occurs if the resistance of the pulled-out fibers at first crack is higher than the load at first crack. Since the matrix does not resist any tension at fracture. The fibers carry the entire load taken by the composite ant they will transfer stresses to the matrix through bond stresses when the load on the composite is increased.

A network of multiple cracks will occur by fiber pull-out and fiber bridging, therefore hampering the rapid propagation of the initial crack. This leads to the bending of the material instead of its’ fracture. This process of multiple cracking will progress until either the fibers fail or the accumulated debonding will occur in fiber pull-out.

As shown in figure 1, the increasing strength results in a fast crack and a low amount of plastic deformation with monolithic CPC.  However, the fiber reinforced CPC increases in strength with increasing stress, which makes it a strain hardening composite[6].

Interface is an important aspect in fiber reinforced composites. The interface bonding between the fiber and the matrix needs to be of an adequate strength for the load to take place between the two elements of the composite. A bond that is too weak will lead to fiber pull-out, while a strong bond will determine fiber fracture.

Also, another important parameter is the length and diameter of the fibers, which influences the transfer of the stress from the matrix to the fiber. Fiber length is relevant because Fibers that are too short do not fully utilize their reinforcing capability. This leads to pull-out before the fiber fractures. A critical fiber length and volume is necessary for efficient ductilization and to induce multiple cracking.

6.1.5 Fiber requirements

6.1.5.1 Biological parameters

Biocompatibility relates to the ability of a biomaterial to exhibit its desired function towards a medical therapy, without causing any undesirable systemic or local responses in the recipient of the therapy [10]. Since the fibers for calcium phosphate reinforcement are used in the medical sector their biocompatibility is a crucial aspect that needs to be considered.  Thus, fibers can be degradable or non-degradable. Reinforcing CPCs with degradable fibers follows a different strategy compared to using non-degradable fibers. This first category of CPCs is focused on providing temporary reinforcement. Bone is able to grow into the macropores that are created after the fibers are dissolved. The strength lost by fiber degradation is ideally compensated by the formation of new bone. Poly(L-lactic acid) (PLLA) is a commonly used biodegradable polymer for fiber reinforcement due to its relatively good strength and biocompatibility. Research has shown that porous PLLA has a strong influence on the formation of new bone by degradation and absorption of PLLA [11]. The second category of CPCs reinforced with non-degradable fibers is mainly focused on mechanical reinforcement. A non-resorbable fiber such as polyamide 6.6 (PA6.6) is widely used. These fibers have a high failure strain and a low elastic modulus. The combination of these parameters results in a high work of fracture as a measure for the toughness [9].

6.1.5.2 Physico-chemical parameters

Fiber length, fiber volume fraction, and fiber strength are important parameters that control the mechanical properties of fiber reinforced CPCs [12]. Long fibers are needed to bridge large cracks at non-stress-bearing sites; however the volume fraction of longer fibers can be much smaller in comparison to the volume fraction of shorter fibers. The long fibers decrease the injectability of the mix significantly. In order for fibers to be used effectively for fiber reinforcement, a critical minimum length (lc) can be calculated as shown in the equation below.

          (1)

where lc represents the critical minimum length of the fiber, df is the fiber diameter, f,B is the fiber strength and i is the interfacial shear strength between the fiber and matrix.

The equation shows that the critical length decreases with increasing adhesion of the fiber to the matrix. If the interfacial shear strength increases, the critical length (l_c) will decrease. In summary, when the adhesion of the fiber to the matrix is improved, the fibers can be shorter. A shorter fiber also improves the workability of the cement [6].

6.1.5.3. Adhesion to the matrix

The interfacial shear strength (ISS) is essential for the biomechanical properties of fiber reinforced CPCs. The fiber and matrix work together and therefore the synergy of this two-component system is crucial. By shear deformation the load is transferred at the matrix-fiber interface. The adhesive forces between the fiber surface and the matrix have been known to have an efficient influence on stress transfer between the two phases [9]. When a weak bond exists between the fibers and the matrix, it is possible that the fibers slip out at relatively low loads and lose their fiber-bridging efficacy. In this situation, toughening the system by adding fibers loses its purpose. However, if the bond is too strong, fibers will break before they can dissipate energy through a sliding out mechanism. Polyesters are commonly used as fiber material for reinforcement of calcium phosphate cements. Nevertheless, PLLA is a hydrophobic material, which means that PLLA fibers do not easily adhere to hydrophilic CPCs.

6.1.6 Enhancement of the matrix-fiber interface

Several treatments have been applied in order to improve the matrix-fiber interface. One of these treatments is the mechanical anchoring of the fibers. Anchoring in the matrix can be achieved by fibers with specifically designed morphology. Also roughness of the surface of the fiber can lead to friction and therefore mechanical anchoring.

Another approach of improving the interface is modifying the fiber surface with different functional groups. Surface functionalization by plasma cleans the surface and subsequently generates new polar groups. An etch effect leads to a roughening of the fiber surface and wetting by aqueous media is improved[6].

A third approach will be investigated in the current study, which involves specific chemical binding of the fibers to the matrix. Currently, a novel approach for improving the adhesion of fibers to the cement matrix is researched at the Department of Biomaterials at Radboudumc. This method involves polydopamine-assisted immobilization of bisphosphonate groups onto PLLA polyester fibers to improve the adhesion of PLLA fibers to calcium phosphate cement. Bisphosphonates (BP) are the most commonly prescribed drugs for osteoporosis. They display a specific and very strong adhesion for calcium ions as present in the mineral phase of bone[13]. This might suggest that fibers modified with BPs will have an improved affinity to the CPCs and therefore increase the interfacial shear strength. To immobilize these calcium-binding BP groups onto PLLA fibers, polydopamine immobilization was used in the current study. Polydopamine is advocated as a universal glue due to its adhesive properties [14]. In nature, the proteins that are produced and secreted by mussels have a good adhesive property and dopamine is a compound that retains certain functionalities similar to this class of mussel inspired adhesives, which makes it a great candidate for coating different types of substrates.

6.1.7 Problem

The current generation of synthetic bio-ceramic bone substitutes such as calcium phosphate cements (CPCs) has poor mechanical properties, which limits the use of these materials to non-load-bearing sites. Fiber reinforcement using degradable polyester fibers is proposed to overcome this problem, but conventional polyester fibers such as poly(L-lactic acid) are hydrophobic and do not adhere to calcium phosphate ceramics. Consequently, fiber reinforcement of CPCs using polyester fibers is not yet optimal.

7.2 Purpose of study & hypothesis

7.2.1 Research question

Can the stiffness, bending strength and toughness of fiber reinforced calcium phosphate cements CPCs be improved by coating poly(L-lactic acid) micro-fibers with bisphosphonates using a polydopamine based mobilisation method?

7.2.2 Purpose of study

To study the effect of bisphosphonate modification of poly(L-lactic acid) micro-fibers on the stiffness, bending strength and toughness of calcium phosphate cements.

7.2.3 Hypotheses

1. The stiffness of calcium phosphate cements reinforced with bisphosphonate-modified poly(L-lactic acid) microfibers will be higher than the cements reinforced with unmodified poly(L-lactic acid) micro-fibers.

2. The bending strength of calcium phosphate cements reinforced with bisphosphonate-modified poly(L-lactic acid) microfibers will be higher than the cements reinforced with unmodified poly(L-lactic acid) micro-fibers.

3. The toughness of calcium phosphate cements reinforced with bisphosphonate-modified poly(L-lactic acid) micro-fibers will be higher than the cements reinforced with unmodified poly(L-lactic acid) micro-fibers.

7.3 Study design

The starting material for this study are poly(L-lactic acid) microfibers with a diameter of 11 um and a length of 6. These fibers were cut both manually and mechanically into fiber lengths of 2.0 mm. The mechanically cut fibers were trimmed with a Pulverisette 19 cutting mill while the manually cut fibers were shortened with a carbon steel surgical blade. The next step in the study refers to the chemical modification of the fiber surface. To this end, the fibers are initially coated with polydopamine, followed by conjugation of bisphosphonate (BP) groups (more specifically alendronate groups) onto the fiber surface. Following this, an apatite-forming calcium phosphate cement was mixed with a) unmodified, b) polydopamine-modified poly(L-lactic acid) or c) bisphosphonate-modified poly(L-lactic acid) microfibers. Rectangular CPC samples (4 × 4 × 20 mm) were made consisting of CPCs with or without mechanically cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with either PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP). A sample size of n=10 was applied to all the groups. The stiffness, bending strength and toughness of the samples were measured in a three point bending test in which force versus the displacement was measured. The toughness, flexural modulus and flexural strength of the tested specimens were calculated from these force-displacement curves and analysed statistically using one-way analysis of variance (one-way ANOVA) followed by a Tukey post hoc test.

8. Materials and methods

8.1 Materials

Poly(L-lactic acid) micro-fibers  were obtained from Trevira (Bobingen, Germany) and α-tricalcium phosphate (α-TCP) micro-particles were obtained from CAM Bioceramics (Leiden, The Netherlands). Amino-bisphosphoante (BP) alendronic acid was purchased from AK Scientific. All other chemicals such as ethylenediamine (EDA) were obtained from Sigma Aldrich, unless stated otherwise.

8.2 Polydopamine modification of poly(L-lactic acid) fibers

The mechanically cut fibers were cut into 2.0 mm using a cutting mill (Fritsch, Pulverisette 19). The manually cut fibers were cut into 2.0 mm using a scalpel. For surface modification, each type of fiber was immersed in an aqueous coating solution at room temperature. The reaction solution was obtained by dissolving dopamine hydrochloride (2 mg/mL) in 10 mM Tris buffer (pH 8.5). The solution was added to the fibers and the reaction took place for 24 hours under mild shaking (150 rpm). After 24 h coating time, the loosely bound polydopamine was washed two times with Tris buffer. The solution was added once more to the fibers and the same reaction took place. After 24 h coating time, the loosely bound polydopamine was washed two times with a Tris buffer. All the samples were freeze dried.

8.3 Bisphosphonate modification of poly(L-lactic acid) fibers

The mechanically cut fibers were cut into 2.0 mm using a cutting mill (Fritsch, Pulverisette 19). The manually cut were cut into 2.0 mm using a scalpel. For surface modification, each type of fiber was immersed in an aqueous coating solution at room temperature. The reaction solution was obtained by dissolving dopamine hydrochloride (2 mg/mL) in 10 mM Tris buffer (pH 8.5). The solution was added over the fibers and the reaction takes place for 24 hours, under mild shaking (150 rpm). After 24 h deposition time, the loosely bound polydopamine was washed two times with Tris buffer. A solution of PDA/EDA/BP was prepared by dissolving dopamine hydrochloride (2 mg/mL) and ethylenediamine (111 µL) in 10 mM Tris buffer (pH 8.5). The solution was added to the fibers. By using this technique of PDA/EDA coating, a high density of amine groups was obtained on the substrate surface. Subsequently a 5% glutaraldehyde v/v concentration was added to the fibers. Glutaraldehyde acts as crosslinker for the reaction between the amine groups of alendronic acid and the amine groups present on the surface of fibers. The reaction took place for 24 hours. After 24 h coating time the fibers were washed again with Tris buffer, the supernatant was removed and the fibers were freeze-dried.

8.4 Preparation of CPC/fiber composites

All fibers were weighed in a plastic container. α-tricalcium phosphate (α-TCP) and 4% (w/v) Na₂HPO₄ solution were added to the fibers and manually mixed to disperse the fibers in the cement matrix. After mixing, the blend in paste form was immediately poured into a silicon mold of 4x4x20 mm. The weight fraction of the fibers was 2.5%. The composite mixture in the mold was covered with a glass slide and stored at room temperature for 24 hours. Subsequently, the hardened composite samples were removed from the mold and incubated in phosphate buffered saline (PBS) for 3 days prior to testing the mechanical properties. The fiber-free CPC samples were prepared in the same way without adding the fibers to the mixture.

8.5 Testing of mechanical properties

Rectangular samples were produced (4 × 4 × 20 mm) followed by a three-point bending test [8]. The samples were soaked first in PBS for three days prior to the three-point bending test. The tests were performed on a mechanical test bench using a load cell of 100 N and at a crosshead speed of 1 mm/min for the samples containing fibers and 0.2 mm/min for the fiber-free CPC specimens. The tests were stopped at an extension of 2 mm.

 

Fig. 2 Three-point bending test 

Before testing, each sample was measured in order to get the correct height and width. As seen in figure 2, the sample was placed on two supporting pins a set distance and a third loading pin was lowered from above at a constant rate until the sample fractures. After fracture, the mechanical properties were evaluated from the resulting load–displacement curve:

The flexural strength of a material is the maximum stress before failure occurs.

in which P_max is the maximum load on the load–displacement curve, L is the flexure span, b is the sample width and h is the sample thickness.

The flexural modulus defines the relation between stress (force per unit) and strain (proportional deformation) in the material. The modulus of living bone is quite low and ranges between 15GPa and 20 GPa.

in which m is the slope of the tangent to the initial straight-line portion of the load–displacement curve (N/m), b is the sample width and h is the sample thickness.

The Work of Fracture (WOF) is used to characterize the toughness of a material. The correlated energy dissipation results in a larger Work of Fracture. [6]

in which A is the curve under the load–displacement (the work done by the applied load to deflect and fracture the sample), b is the sample width and h is the sample thickness.

Abbreviation			PLLA w/t%	PDA	BP
CPC-PLLA			2,5% (mechanically cut)	no	no
CPC-PLLA-PDA			2,5% (mechanically cut)	yes	no
CPC-PLLA-BP			2,5% (mechanically cut)	yes	yes
CPC-PLLA			2,5% (manually cut)	no	no
CPC-PLLA-PDA			2,5% (manually cut)	yes	no
CPC-PLLA-BP			2,5% (manually cut)	yes	yes
CPC			0 %	no	no

8.6 Physico-chemical characterization

SEM

The fractured surfaces of the samples were collected to study the morphology after the bending tests. We used scanning electron microscopy (Zeiss), which was operated at an accelerating voltage of 10 kV and a working distance of 15 mm.

8.7 Statistics

Data are presented as mean ± standard deviation. A one-way analysis of variance (ANOVA) followed by a Tukey post hoc test was performed to identify significant differences. Results were considered significant if P < 0.05. Calculations were performed using GraphPad Instat^®.

9. Results

9.1 Load-displacement curves

In figure 3 representative load-displacement curves of the three-point bending tests are shown of CPCs with or without mechanically cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA(CPC-PLLA-PDA) or BP( CPC-PLLA-BP).

The CPC without mechanically cut PLLA fibers displayed a very low amount of plastic deformation. An extension of 0,1 mm was reached, after which the load dropped instantaneously.  For all CPC groups with mechanically cut PLLA fibers the extension reached up to 1 mm and for some samples even up to 2 mm. After the initial linear elastic phase a phase of plastic deformation is seen in which the curve moves up and down repeatedly. With increasing displacement the load decreases until it almost reaches 0 N. The maximum load was at an average of 15 N for each of the CPCs with mechanically cut PLLA fibers.

In figure 4 representative load-displacement curves of the three-point bending tests are shown of CPCs with or without manually cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA(CPC-PLLA-PDA) or BP( CPC-PLLA-BP).

The CPC without mechanically cut PLLA fibers displayed a very low amount of plastic deformation. An extension of 0,1 mm was reached, after which the load dropped to 0 N. For all CPC groups with manually cut PLLA fibers the extension reached up to 1 mm and for some samples even up to 3 mm. After the initial linear elastic phase, a phase of plastic deformation is seen in which the curve moves up and down repeatedly. With increasing displacement the load decreases until it almost reaches 0 N. The phase of plastic deformation starts at a different load for each of the CPCs with manually cut PLLA fibers. CPC-PLLA enters this phase at an average load of 13 N, CPC-PLLA-PDA enters this phase at an average load of 11 N and CPC-PLLA-BP enters this phase at an average load of 8 N.

9.2 Flexural strength

In figure 5 the flexural strength as calculated using formula 2 is shown for CPCs with or without mechanically cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).

Fig. 5 Flexural strength of CPCs with or without mechanically cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).

The average flexural strength of all the experimental groups was about 6 MPa. Differences between all experimental groups were statistically insignificant (p>0,05).

In figure 6 the flexural strength as calculated using formula 2 is shown for CPCs with or without manually cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).

 

Fig. 6 Flexural strength of CPCs with or without manually cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).

The average flexural strength of all the groups is around 6 MPa. Differences between all experimental groups were statistically insignificant (p>0,05).

9.3 Flexural modulus

In figure 7 the flexural modulus as calculated using formula 3 is shown for CPCs with or without mechanically cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).

Fig. 7 Flexural modulus of CPCs with or without mechanically cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP). 

All the groups have an average flexural modulus of around 1300 MPa. Differences between all experimental groups were statistically insignificant (p>0,05).

In figure 8 the flexural modulus as calculated using formula 3 is shown for CPCs with or without manually cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).

Fig. 8  Flexural modulus of CPCs with or without manually cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).  “*” indicates significant difference relative to the two groups connected by the line (p< 0,05) 

The flexural modulus for all the groups was statistically insignificant for all groups except CPC-PLLA-PDA for which a significantly lower stiffness was measured as compared to unreinforced CPC and CPC-PLLA-PDA samples (p<0,05).

9.4 Work of fracture

In figure 9 the Work of Fracture (as a measure for toughness) as calculated using formula 4 is shown for CPCs with or without mechanically cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).

Fig. 9  Work  of Fracture  (J/m2) of CPCs with or without mechanically cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).  “*” indicates significant difference relative to the two groups connected by the line (p< 0,05) 

The Work of Fracture of the CPC specimens without mechanically cut PLLA fibers was significantly lower than the work of fracture for all the groups containing mechanically cut PLLA fibers (p<0,05). Between these three groups the Work of Fracture of the CPC-PLLA specimens was statistically lower than the Work of Fracture of the CPC-PLLA-PDA samples (p<0,05).

In figure 10 the Work of Fracture (as a measure for toughness) as calculated from formula 4 is shown for CPCs with or without manually cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).

Fig. 10 Work of Fracture  (J/m²) of CPCs with or without manually cut PLLA fibers. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP).  “*” indicates significant difference relative to the two groups connected by the line (p< 0,05)

The Work of Fracture of the CPCs without manually cut PLLA fibers was significantly lower than the Work of Fracture of all the groups containing manually cut PLLA fibers (p<0,05). Differences between the CPC specimens with manually cut PLLA fibers groups were statistically insignificant (p>0,05). The Work of Fracture of all these groups are around 700 J/m² in comparison to the CPC group with a much lower Work of Fracture of ~15 J/m².

7.5 SEM Analysis

In figure 11 a scanning micrograph is shown of a sample of CPCs without any PLLA fibers.

Fig. 11 Scanning electron micrograph of CPCs without PLLA fibers.

Nanopores are visible throughout the entire cement. A structure of entangled platelet-shaped hydroxyapatite crystallites was clearly visible, which is characteristic for the microstructure of self-hardening CPCs.

	2,5 % Manually cut	2,5 % Mechanically cut
PLLA	a)	b)
PLLA-PDA	c)	d)
PLLA-BP	e)	f)

Fig. 12 Scanning electron micrographs of both manually and mechanically cut PLLA fibers. a) and b) were left unmodified (PLLA). c) and d) are PLLA fibers coated with PDA (PLLA-PDA). e) and f) are PLLA fibers modified with BP (PLLA-BP).

Figure 12 shows scanning electron micrographs of both manually and mechanically cut fibers. Unmodified PLLA fibers had smooth surfaces and did not display curves or twists. Mechanically cut but unmodified PLLA fibers were twisted and have curves and the surface appearance was irregular. Manually cut PLLA fibers coated with PDA displayed a rougher layer on the fibers, which was caused by the polydopamine coating. The mechanically cut PLLA fibers coated with PDA displayed a thick and rough layer on the fibers, which was caused by the polydopamine coating. The manually cut PLLA fibers that were modified with BP had a smoother layer visible on the fibers. The mechanically cut PLLA fibers modified with BP had a smoother layer visible on the fibers.

	CPC-PLLA-PDA  Manually cut	CPC-PLLA-PDA  Mechanically cut
Magnification: 200 X	a)	c)
Magnification: 1000 X	b)	d)

Fig. 13 Scanning electron micrographs at low (top) and high (bottom) magnifications of fractured CPC samples containing manually cut PLLA fibers coated with PDA (left column) or fractured CPC samples containing mechanically cut PLLA fibers (right column).

Figure 13 shows manually cut PLLA reinforced CPC at a low a) and a high b) magnification and mechanically cut PLLA reinforced CPC at a low c) and a high d) magnification. The sample of manually cut CPC-PLLA-PDA shows holes in the cement, which is the effect of fiber pull-out. The fibers, which are not fully imbedded in the cement, have no curves or twists. The sample of mechanically cut CPC-PLLA-PDA shows holes in the cement a, which is effect of fiber pull-out. The fibers, which are not fully imbedded in the cement, are twisted and curved.

8. Discussion

The current generation of synthetic bioceramic bone substitutes such as calcium phosphate cements (CPCs) has poor mechanical properties, which limits the use of these materials to non-load-bearing skeletal sites. These poor mechanical properties can be overcome by adding poly(L)-lactic acid (PLLA) fibers, as these are biodegradable polymers with relatively good strength and biocompatibility. The adhesion between the PLLA fibers and the CPCs is, however, not optimal due to the difference in polarity between these two types of materials. PLLA is a hydrophobic polymer, whereas CPC is a hydrophilic ceramic. Therefore, this study researched the effect of surface modification of PLLA fibers on the mechanical properties of fiber reinforced CPCs. More specifically, the goal of this study was to research if surface modification of poly(L-lactic acid) micro-fibers with calcium-binding bisphosphonate (BP) groups using a polydopamine (PDA)-based immobilisation method had an effect on the stiffness, bending strength and toughness of calcium phosphate cements. To this end, CPCs with or without PLLA fibers were prepared. The fibers were left unmodified (CPC-PLLA) or coated with PDA (CPC-PLLA-PDA) or BP (CPC-PLLA-BP). The fibers were cut mechanically as well as manually. All samples were tested mechanically using a three-point bending test. We hypothesized that the stiffness, bending strength and toughness of calcium phosphate cements reinforced with surface-modified PLLA fibers would be significantly higher than the cements reinforced with unmodified poly(L-lactic acid) microfibers. Our results showed that the fiber modification had no statistically significant effect on either the stiffness, bending strength or toughness of the CPCs reinforced with mechanically or manually cut PLLA fibers. Summarizing, the added value of surface modification of the PLLA fibers using BP coatings was not confirmed in the current study.

Fiber-free calcium phosphate cements

In figure 11 the scanning micrograph is visible of a CPC without any PLLA fibers. Nanopores are visible, which is the main cause for the brittleness of CPC. As observed in figure 3 and 4 the CPCs without PLLA fibers indeed fractured instantaneously after a load of approximately 15 N and afterwards there was no plastic deformation.

Fiber modification

In figure 12 both the manually and the mechanically cut PLLA fibers are visible. Before any modification was performed the manually cut PLLA fibers show smooth surfaces and no curves or twists. However, the mechanically cut PLLA fibers show uneven surfaces and many curves and twists. Apparently, the mechanical cutting affected the surface morphology of the PLLA fibers. After application of the polydopamine coating a visible difference can be seen in figure 12 as well. The mechanically cut fibers show a thicker and rougher layer of the fibers in comparison to the manually cut PLLA fibers coated with PDA. This phenomenon can be attributed to the fact that the polydopamine coating adheres more easily to the mechanically cut PLLA fibers than to the manually cut PLLA fibers with a smoother outer layer.

In the process of preparing the samples there was a big difference visible between the manually and the mechanically cut PLLA fibers. By cutting the PLLA fibers mechanically, fiber aggregates were formed, even more so after modifying them. These aggregates eventually hindered homogeneous dispersion of the PLLA, which led to samples in which the PLLA fibers were not homogeneously distributed. This problem was observed to a lesser extent for the manually cut PLLA fibers. These PLLA fibers were more easily dispersed in comparison to the mechanically cut PLLA fibers.

Force-displacement curves

Figure 3 and 4 show that instead of breaking instantly due to fast propagation of a single crack, the fiber reinforced CPCs displayed tough behaviour characterized by repeated fluctuations in the load-displacement curves where each fluctuation corresponds to a pull-out event of a PLLA fiber. The images in figure 13 are in line with the above-mentioned fiber pull-out phenomena since several fibers were pulled from the CPC matrix. This process of fiber pull-out hampered the rapid propagation of the initial crack, which eventually led to bending of the material instead of instantaneous fracture characteristic for brittle behaviour. This explains the fluctuation in the representative load-displacement curves shown in figure 3 and figure 4.

Mechanical properties of fiber-reinforced cements

When comparing the flexural strength of all the groups no significant differences were observed as shown in figure 5 and 6. The addition of mechanically cut PLLA fibers with or without modifications or manually cut PLLA fibers with or without modifications to the CPCs seems to have no additional value to the strength of the samples. The flexural modulus also showed no significant differences between all of the groups as shown in figure 7. This might be explained by the relative weakness of PLLA fibers, since fibers for reinforcement should be ideally strong and stiff in order to obtain a high reinforcement efficacy resulting into high stiffness and bending strength. The poor dispersion of the PLLA fibers in the matrix could also play a significant role in the low stiffness and bending strength. However, it is difficult to show the dispersion of these PLLA fibers in unfractured PLLA-reinforced samples using scanning electron microscopy. It might be visible in figure 13, in which parts of the CPC in which there are cutters of fibers and other parts that almost have no fibers at all. Fiber agglomeration in CPC was also observed in other studies such as Xu et al. [15]. An alternative method to visualize the dispersion of the fibers before fracturing is the use of nanocomputed tomography (nanoCT), which is recommended for further research.

In figure 8, however, a significant difference is shown between the CPCs without any PLLA fibers and the CPCs with manually cut PLLA fibers coated with PDA. This suggests that the stiffness decreased after adding these fibers to the CPCs. The polydopamine coating might reduce the adhesion of the fibers to the CPC and therefore result into less effective stress transfer. In addition, it should be realized that the PLLA fibers were already hydrophilic after application of a proprietary undisclosed surface coating by manufacturer Trevira. The SEM analysis showed that the manually cut fibers were covered with a thinner layer of polydopamine in comparison to the mechanically cut fibers, as seen in figure 12. It could be possible that the polydopamine can be applied more easily on the mechanically cut fibers as these were curved and twisted. The manually cut fibers showed very smooth surfaces and no curves or twists. The polydopamine might have adhered more weakly due to this smoothness. There is also no experimental proof that the polydopamine coating and BP modification actually makes the surface of the fibers more hydrophilic.  Investigating if the addition of polydopamine coating and the BP modification have a significant improving effect when using a fiber that has not already been coated is a possibility.

Generally, our results indicate that the stiffness and bending strength of calcium phosphate cements reinforced with bisphosphonate-modified poly(L-lactic acid) microfibers was equal to cements reinforced with unmodified PLLA fibers. Consequently, hypotheses 1 and 2 (section 7.2.3) were refuted, and it was concluded that there was no effect of modification of the PLLA fibers on the stiffness and bending strength of the resulting fiber-reinforced composites.

However, both CPCs reinforced with mechanically cut PLLA fibers as well as CPCs reinforced with manually cut PLLA fibers were significantly stiffer, stronger and tougher than CPCs without the addition of PLLA fibers. The WoF was significantly higher, around 800 J/m², in comparison to the WoF of the CPCs with the addition of mechanically cut PLLA fibers. The WoF of the CPCs with manually cut PLLA fibers was approximately 400 J/m² higher than the CPCs with the mechanically cut PLLA fibers. By cutting the PLLA fibers mechanically the surface of these fibers is affected which was observed in figure 12. It suggests that the modifications only have a significant effect if the properties of the fibers are damaged. Consequently, the damaged mechanically cut fibers benefited from the surface modifications, whereas the manually cut PLLA fibers did not benefit from the modifications. The PLLA fibers increase the toughness of the CPCs without any modification beforehand, even more than the PLLA fibers with the PDA coating or bisphosphonate modifications.

An additional significant difference is shown in figure 9. Previous studies have also shown that adding fibers to CPCs increases the WoF[6][7][9][12]. Between the CPC-PLLA and the CPC-PLLA-PDA an additional significant increase is shown. The polydopamine layer on the fibers might have made the adhesion between fibers and the CPC better. The fibers could have been more hydrophilic by the polydopamine coating. Therefore, the adhesion to the hydrophilic CPCs could possibly be better than uncoated PLLA fibers.

The third hypothesis (section 7.2.3) can be refuted as well since the toughness of calcium phosphate cements reinforced with bisphosphonate-modified poly(L-lactic acid) micro-fibers was not statistically different from the cements reinforced with unmodified poly(L-lactic acid) micro-fibers. The WoF for the CPC-PLLA-PDA was, however, significantly higher than the WoF of the CPC-PLLA samples. The bisphosphonate-modification might not have worked because of the lack of interaction with the matrix. It might also be possible that the bisphosphonates did not adhere enough to the polydopamine coating in order to improve the toughness. Testing the adhesion of polydopamine coatings onto PLLA fibers is, however, extremely difficult from an experimental point of view since the PDA coatings are thinner than 100 nm [16].

A study by Zuo et al. investigated CPCs reinforced with PLLA fibers [8]. This study showed a WoF of 180 J/m^2, which was lower than the WoF presented in this study(~700-900 J/m²). A fiber content of 3 w/t% was applied, which was close to the 2,5 w/t% as applied in our study. However, they used electrospun PLLA nanofibers instead of manually cut PLLA micro-fibers. By electrospinning the PLLA fibers the lower WoF might be explained. Instead of reinforcing the CPCs it might be weakening the samples. Furthermore it is quite difficult to make an accurate comparison between other studies, as the PLLA fibers in this study have not been researched a lot. Still, there are several previous studies performed with PLGA fibers such as polyglactin fibers (biodegradable copolymers of lactide and glycolide in a lactide:glycolide ratio of 10:90) which reveal an average WoF of between 2000-4000 J/m² [6]. For example, in this study, the CPCs with the addition of manually cut PLLA fibers show a WoF of  ~700-900 J/m². A different study shows a WoF of 2500 – 5000 J/m² with CPCs reinforced with polyglactin fibers [12]. The differences that can be observed between the studies were the concentration of fibers and different lengths of fibers. The above-mentioned study used 6 w/t% and 3 mm fibers in comparison to 2,5 w/t% and 2 mm fibers which was used in our study. They also used a different kind of fiber, i.e., polyglactin instead of PLLA. Another study with polyglactin and an additive of methylcellulose showed a WoF of 300 J/m². The fiber concentration (2,5 w/t%) was the same as our study, but the fibers were cut at a length of 8 mm [17]. In comparison the PLLA fibers seem to have a less pronounced influence on the WoF of CPCs. Different lengths of fibers might explain this. Generally the studies with higher fiber contents showed higher WoF.

In any future study it is suggested to avoid the use of mechanically cut PLLA fibers as the cutting process damages the fibers too much. The method of immobilizing the bisphosphonate groups to the PLLA fibers should be investigated extensively, as this method is quite promising. Future research will have to focus on compatibilization of the fiber-matrix composition, the degradation behaviour of these components and making these composites strong enough for load bearing sites in order for them to be applied clinically.

9. Conclusion

In conclusion it was shown that:

1. PLLA fibers were damaged by cutting them mechanically using a cutting mill.
2. The surface modification of PLLA fibers did not increase the stiffness, bending strength or toughness of calcium phosphate cements reinforced with bisphosphonate-modified poly(L-lactic acid) micro-fibers in comparison to the cements reinforced with unmodified poly(L-lactic acid) micro-fibers.

9. References

1. Hannink, G. and J.J. Arts, Bioresorbability, porosity and mechanical strength of bone substitutes: what is optimal for bone regeneration? Injury, 2011. 42 Suppl 2: p. S22-5.

2. Piattelli, M., et al., Bone reactions to anorganic bovine bone (Bio-Oss) used in sinus augmentation procedures: a histologic long-term report of 20 cases in humans. Int J Oral Maxillofac Implants, 1999. 14(6): p. 835-40.

3. Rahaman, M.N., et al., Bioactive glass in tissue engineering. Acta Biomater, 2011. 7(6): p. 2355-73.

4. Bohner, M., U. Gbureck, and J.E. Barralet, Technological issues for the development of more efficient calcium phosphate bone cements: a critical assessment. Biomaterials, 2005. 26(33): p. 6423-9.

5. Boccaccini, A.R. and P.X. Ma, Tissue engineering using ceramics and polymers. Second edition. ed. Woodhead Publishing series in biomaterials. 2014, Cambridge, UK ; Boston: Elsevier/WP, Woodhead Publishing is an imprint of Elsevier. xxix, 698 pages.

6. Kruger, R. and J. Groll, Fiber reinforced calcium phosphate cements — on the way to degradable load bearing bone substitutes? Biomaterials, 2012. 33(25): p. 5887-900.

7. Perez, R.A., H.W. Kim, and M.P. Ginebra, Polymeric additives to enhance the functional properties of calcium phosphate cements. J Tissue Eng, 2012. 3(1): p. 2041731412439555.

8. Zuo, Y., et al., Incorporation of biodegradable electrospun fibers into calcium phosphate cement for bone regeneration. Acta Biomater, 2010. 6(4): p. 1238-47.

9. Canal, C. and M.P. Ginebra, Fibre-reinforced calcium phosphate cements: a review. J Mech Behav Biomed Mater, 2011. 4(8): p. 1658-71.

10. Williams, D.F., On the mechanisms of biocompatibility. Biomaterials, 2008. 29(20): p. 2941-53.

11. Hamada, Y., et al., The preparation of PLLA/calcium phosphate hybrid composite and its evaluation of biocompatibility. Dent Mater J, 2012. 31(6): p. 1087-96.

12. Xu, H.H., F.C. Eichmiller, and A.A. Giuseppetti, Reinforcement of a self-setting calcium phosphate cement with different fibers. J Biomed Mater Res, 2000. 52(1): p. 107-14.

13. An, J., et al., Influence of polymeric additives on the cohesion and mechanical properties of calcium phosphate cements. J Mater Sci Mater Med, 2016. 27(3): p. 58.

14. Rim, N.G., et al., Mussel-inspired surface modification of poly(L-lactide) electrospun fibers for modulation of osteogenic differentiation of human mesenchymal stem cells. Colloids Surf B Biointerfaces, 2012. 91: p. 189-97.

15. Xu, H.H., et al., Strong and macroporous calcium phosphate cement: Effects of porosity and fiber reinforcement on mechanical properties. J Biomed Mater Res, 2001. 57(3): p. 457-66.

16. Lee, H., et al., Mussel-inspired surface chemistry for multifunctional coatings. Science, 2007. 318(5849): p. 426-30.

17. Xu, H.H., et al., Injectable and macroporous calcium phosphate cement scaffold. Biomaterials, 2006. 27(24): p. 4279-87.