- Email: [email protected]
Biologically mediated resorption of brushite cement in vitro
ARTICLE IN PRESS
Biomaterials 27 (2006) 2178–2185 www.elsevier.com/locate/biomaterials
Biologically mediated resorption of brushite cement in vitro Liam M. Grovera,, Uwe Gbureckb, Adrian J. Wrightc, Maryjane Tremaynec, Jake E. Barraleta a Faculty of Dentistry, McGill University, 3640 University Street, Montre´al, Que´bec, Canada H3A 2B2 Department of Functional Materials in Medicine and Dentistry, University of Wurzburg, Wurzburg D-97070, Germany c School of Chemistry, University of Birmingham, Birmingham B15 2TT, UK
b
Received 12 August 2005; accepted 6 November 2005
Abstract A new calcium phosphate cement is reported, which sets to form a matrix consisting of brushite, dicalcium pyrophosphate dihydrate and an amorphous phase following the mixture of b-tricalcium phosphate with an aqueous pyrophosphoric acid solution. This reactant combination set within a clinically relevant time-frame (approximately 10 min) and exhibited a higher compressive strength (25 MPa) than previously reported brushite cements. The in vitro degradation of the b-tricalcium phosphate–pyrophosphoric acid cement was tested in both phosphate buffered saline and bovine serum. The pyrophosphate ion containing cement reported here was found not to be hydrolysed to form hydroxyapatite in vitro like b-tricalcium phosphate–orthophosphoric acid solution cements. This finding is significant since the formation of hydroxyapatite by hydrolysis is thought to retard in vivo degradation of brushite cements. When aged in bovine serum, the cement lost considerably more mass than when aged in phosphate buffered saline, indicating that proteins, most likely phosphatase enzymes played an important role in the degradation. As pyrophosphate ions are thought to be the source of orthophosphate ions during bone mineralisation, this new class of bone cement offers a route to new degradable synthetic bone grafting materials. r 2005 Elsevier Ltd. All rights reserved. Keywords: Calcium phosphate cement; Alkaline phosphatase; Brushite; Degradation; In vitro test; Pyrophosphate
1. Introduction In vivo studies investigating the biological reaction to and degradation of brushite cements have reported complete or extensive resorption [1], in addition to fragmentation [2] or long-term stability of the cement [3]. Crystallographic and spectroscopic analyses of retrieved brushite cement implants have shown that a marked reduction in the rate of resorption occurs following the formation of hydroxyapatite in the cement, the presence of which is thought to be caused by hydrolysis of the brushite [4]. A previous in vitro ageing study by our group has shown that the rate of mass loss from brushite cement and the formation of hydroxyapatite are affected strongly by ageing medium refreshment rate, the volume of liquid to which the cement is exposed and the presence of protein in Corresponding authors. Tel.: +1 514 398 7203; fax: +1 514 398 8900.
E-mail address: [email protected] (L.M. Grover). 0142-9612/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2005.11.012
the ageing media (which inhibited hydrolysis) [5]. As potential resorption is the major advantage that brushitebased calcium phosphate cements have over hydroxyapatite cements, attempts have been made to prevent brushite hydrolysis in the cement. One approach was the addition of magnesium ions to the cement mix [6]. Magnesium is a potent inhibitor of hydroxyapatite crystallisation since it adsorbs to the surface of newly forming hydroxyapatite crystal nuclei and blocks active growth sites [7]. In an in vitro study, a magnesium salt (8.5 wt% MgHPO4 3H2O) was shown to prevent hydroxyapatite formation when a brushite cement was aged in Hank’s solution [8]. Following the implantation of a similar cement formulation containing 5 wt% MgHPO4 3H2O in the distal and proximal humerus of Swiss Alpine sheep, however, hydroxyapatite formation was noted after 4 months [9]. Another apatite crystallisation inhibitor [10], the pyrophosphate (P2O4 7 ) ion has previously been added to brushite cements at very low concentrations as a means of
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
2. Methods and materials The b-tricalcium phosphate (b-TCP)–pyrophosphoric acid solution cement was produced from the combination of b-TCP (Plasma-Biotal, Derbyshire, UK) with 1 mL of a liquid component consisting of 540 mg pyrophosphoric acid (Rhodia, West-Midlands, UK) and 720 mg double distilled water. To determine the range of powder to liquid ratios to which the cement could be mixed, the mass of b-TCP added to the liquid phase was varied between 1.25 and 2.75 g at 250 mg increments (powder to liquid ratios of 1.25–2.75 g/mL). To mix the cement paste, the b-TCP was divided into four equal portions. The b-TCP was then added to the liquid phase a quarter at a time which was mixed using a non-corrodible spatula for a period of 5 s. Once all of the b-TCP was combined with the liquid, the cement paste was kneaded for a further 20 s. The setting times of the cements ðn ¼ 3Þ were determined using the Gilmore needles technique. Samples for compression testing were fabricated by placing manipulated cement slurry into a PTFE split mould forming hardened cement cylinders of diameter 6 mm and height 12 mm. Once set, the samples were removed from the mould and stored in 100% relative humidity at 3771 1C for a period of 24 h. Before testing, geometrical measurements of the cement cylinders were made in triplicate and the samples were weighed. Samples were mounted on the testing machine (5544, Instron, Bucks, UK) so that the long axes of the cement cylinders were perpendicular to the lower anvil. A compressive force was then applied to the upper surface of the cement samples at a constant crosshead displacement rate of 1 mm/min until failure occurred. Mean strength was determined from the average of 10 measurements. The applied load was measured using a 2 kN load cell (5544, Instron, Buckinghamshire, UK). After testing in compression, cement fragments were retrieved, weighed and dried in a vacuum desiccator at a temperature of 3771 1C. The cement fragments were then ground to powder using a pestle and mortar. The true density of the powder was determined using a helium pycnometer (Accupyc 1330, Micromeritics, Bedfordshire, UK). The volume of each sample ðn ¼ 4Þ was measured 10 times following 10 purges of the measurement chamber with helium. The relative porosity of the cement was calculated from apparent and true density measurements. To determine the effects of ageing on the b-TCP–pyrophosphoric acid cement cylinders, samples mixed to a powder to liquid ratio of 1.75 g/mL were aged in PBS (100 mM; Sigma-Aldrich, Dorset, UK) at a liquid to cement volume ratio of 60. The PBS was refreshed on a daily basis to remove any dissolution products. To find the quantity of mass lost from the cement samples with time, samples were removed daily from the ageing medium and weighed. After periods of 3, 14 and 28 days of ageing, cement cylinders were removed from the PBS and tested in compression. Compressive strength values were the mean of 10 measurements. The pore size distribution of the b-TCP–pyrophosphoric acid solution cement prior to and following 28 days of ageing was measured by using a mercury intrusion porosimeter (9420, Micromeritics, Bedfordshire, UK). Since proteins are known to influence the dissolution and crystallisation of calcium phosphate compounds and as such may influence the in vivo
degradation rate of the b-TCP–pyrophosphoric acid cement, cylinders mixed to a powder to liquid ratio of 1.75 g/mL were aged in bovine serum (Sigma-Aldrich, Dorset, UK) containing sodium azide (Sigma-Aldrich, Dorset, UK) at a concentration of 0.1 wt%. The cylinders were aged at a liquid to cement volume ratio of 60 and the bovine serum was refreshed on a daily basis, when three of the samples were weighed. The compositions of the cement samples were determined by means of Rietveld refinement phase analysis of X-ray diffraction data after periods of 3, 7, 14, 28 and 90 days of ageing. Fourier transform infrared spectra of the samples were collected from a KBr (Sigma-Aldrich, Gillingham, UK) pellet at a spectral resolution of 4 cm1. Prior to ageing and following 3, 7, 14, 28 and 90 days of ageing, the composition of the cement cylinders was determined by means of Rietveld refinement analysis of X-ray diffraction data. X-ray powder diffraction data were collected using Cu Ka1 radiation from a X-ray powder diffractometer (D5005, Siemens, Karlsruhe, Germany), over the range 2y ¼ 20–401 with a step size of 0.021 and a normalised count time of 1 s/ step. Quantitative phase compositions of the materials were calculated by means of total Rietveld refinement analysis with the TOPAS software (Bruker AXS, Karlsruhe, Germany) using the structural models of CaHPO4 2H2O (brushite)[12], b-Ca3(PO4)2 (b-TCP) [13] and Ca2(P2O7) 2H2O (DCPP)[14] from the literature together with a Chebychev fourth-order background model [see Supplementary information for an example of a Rietveld profile]. Analysis of the data indicated the presence of an amorphous phase and in order to confirm and quantify this, powdered samples were mixed with 20 wt% crystalline anatase (TiO2). Diffraction data were collected from the anatase/cement powder mixtures and the resultant patterns were refined using the Rietveld method. The proportion of amorphous material was calculated from the apparent excess in anatase wt% due to non-crystalline phases not being included in refinement calculations.
3. Results Initial experimentation revealed that workable cement pastes could be mixed at powder to liquid ratios in the range of 1.25–2.75 g/mL. Increasing the powder to liquid mixing ratios of the cement paste to 41.25 g/mL reduced both the initial and final setting times exhibited by the cement formed with pyrophosphoric acid solution (Fig. 1). When mixed to a powder to liquid ratio of 1.25 g/mL the initial setting time of the cement formed with pyrophosphoric acid solution was 2171 min and the final setting
50 45 40 Setting time (mins)
retarding the setting rate of brushite cements. Our group has recently reported the formation of new cements consisting of brushite and calcium pyrophosphate by the addition of a pyrophosphoric acid solution, (rather than orthophosphoric (PO3 4 ) acid solution) to b-tricalcium phosphate [11]. In the present study the influence of powder to liquid ratio on the setting times and compressive strength of a b-TCP–pyrophosphoric acid cement was determined. The cement formulation which exhibited the best handling properties was selected and aged in vitro for periods of up to 90 days. The influence of ageing the cement in PBS and serum on compressive strength, porosity and the phase composition of the cement were determined.
2179
35 30 25 20 15 10 5 0
1.25
1.5
1.75
2
2.25
2.5
2.75
Powder to liquid ratio (g/ml)
Fig. 1. The initial and final setting times of cement formed from the mixture of b-TCP and pyrohosphoric acid solution mixed to powder to liquid ratios in the range of 1.25–2.75 g/mL.
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
2180
time was 4471 min and when mixed to a powder to liquid mixing ratio of 2.75 g/mL the initial and final setting times of the cements were less than 1 min. While it was possible to mix a setting cement paste at powder to liquid mixing ratios of 2.50 and 2.75 g/mL, these pastes were dry and flaky as established visually. Powder to liquid ratio had a marked influence on the compressive strength of the resulting cements (Fig. 2). An approximate six-fold increase in compressive strength from 471 to 2572 MPa was observed when powder to liquid mixing ratio was increased from 1.25 to 2.25 g/mL. Increasing the powder to liquid mixing ratio to 2.50 g/mL and higher was detrimental to compressive strength; the cement mixed to a powder to liquid mixing ratio of 2.75 g/ mL exhibited a compressive strength of only 1873 MPa. Fig. 2 illustrates that the increase in compressive strength that was observed when powder to liquid ratio was increased from 1.25 to 1.75 g/mL was associated with a reduction in the relative porosity of the set cement from 4171% to 2171%. Although the cements mixed to higher powder to liquid ratios exhibited higher compressive strengths, no further significant reduction in porosity was apparent. Table 1 indicates that increasing powder to liquid ratio to 41.75 g/mL reduced the degree of reaction. The quantity of unreacted b-TCP increased as the powder to liquid ratio increased, thus strength (Fig. 2) and degree of reaction were not directly correlated.
Period of immersion (days)
Compressive strength (MPa)
Relative porosity (%)
0 3 14 28
1472 1574 1173 873
2171 2271 2671 3671
0.03
40 35 30
20
25 15 20 15
10
10
Relative Porosity (%)
25
5 5 1.25
1.5
1.75 2 2.25 2.5 Powder to Liquid Ratio (g/ml)
2.75
0
Fig. 2. The effects of powder to liquid ratio on the compressive strength and relative porosities of cements formed from the mixture of b-TCP with pyrohosphoric acid solution.
Incremental intrusion volume (ml/g)
Compressive Strength Relative Porosity Compressive Strength (MPa)
Table 2 The compressive strengths and relative porosities of b-TCP–pyrophosphoric acid cement mixed to a powder to liquid ratio of 1.75 g/mL, following different periods of ageing in daily refreshed PBS at an LCVR of 60
45
30
0
When aged in PBS, the mean compressive strength of the cement formed with pyrophosphoric acid solution deteriorated with time from 1472 MPa prior to ageing to 873 after 28 days of ageing (Table 2). The reduction in the compressive strength exhibited by the cement was associated with an increase in relative porosity from 2171% to 3671% after 28 days of ageing (Table 2). Mercury intrusion porosimetry showed that prior to immersion in PBS, the majority of pores in the cement were of diameter between 100 nm and 10 mm (Fig. 3). Ageing of the cement in daily refreshed PBS for a period of 28 days resulted in a considerable increase in the total pore volume of the
Prior to immersion After 28 d immersion 0.02
0.01
0 0.001
0.01
0.1 1 10 Pore diameter (µm)
100
Fig. 3. Pore size distributions of b-TCP–pyrophosphoric acid solution cement (mixed to a powder to liquid mixing ratio of 1.75 g/mL) prior to ageing and after 28 days of ageing in daily refreshed PBS at a LCVR of 60.
Table 1 The influence of powder to liquid ratio on the composition of the set b-TCP–pyrophosphoric acid cement P:L (g/mL)
1.25 1.75 2.25 2.75
1000
Rietveld Rwp (%)
Composition (wt%) Brushite
b-TCP
DCPP
8172 8472 5872 4072
1472 1272 3972 5372
572 472 372 772
6.87 6.80 5.61 6.76
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
cement (Table 2, Fig. 3). The aged cement contained a distribution of pores bimodal about 50 nm or 1 mm in diameter (Fig. 3). The mass loss profile for the cement formed with pyrophosphoric acid solution is shown in Fig. 4. For the first 28 days the cement formed using pyrophosphoric acid solution lost mass at a rate of 0.06% per day when stored in PBS refreshed daily, (Fig. 4A), so that after a period of 28 days of ageing the cement had lost a total of 1.8% of its original mass. After 28 days of ageing there was a slight increase in the rate at which the cement lost mass, with the cement losing mass at a rate of 0.25% per day so that after a period of 90 days of ageing the cement lost a total of 19% of its original mass. Ageing using identical parameters (liquid volume and sample size) but in bovine serum resulted in the cement losing mass at a linear rate of 0.6% per day for the entire experiment, so that after a period of 90 days of ageing the cement had lost a total of 57% of its original mass (Fig. 4B), more than three times the mass lost
Relative Residual Mass (%)
100
80 Pyro- cement Ortho- Cement[5] 60
40 0
20
(A)
40 60 Ageing Time (days)
80
100
Relative Residual Mass (%)
100 Pyro- cement Ortho- Cement[5]
when the cement formed using pyrophosphoric acid solution was aged in PBS for the same period. The crystalline compositions of the cement aged in PBS or bovine serum determined from Rietveld refinement are shown in Tables 3a and b, respectively. Given the contrast in mass loss profiles, it might have been expected that the phase compositions of the cement formed using pyrophosphoric acid solution would differ when aged in PBS and when aged in bovine serum, however, surprisingly throughout the experiment there was little difference in the phase compositions of the cement during ageing. In both media, following immersion in the ageing medium there was a small reduction in the proportion of brushite and dicalcium pyrophosphate dihydrate in the cement coupled with an increase in the proportion of the cement contributed by bTCP. Most phase composition change occurred after only 3 days immersion. Fourier transform infrared spectra demonstrated that prior to ageing, there were peaks 1 present on the spectra characteristic of HPO2 4 (872 cm ) 3 1 and PO4 (1070–1140 cm ) groups (Fig. 5), a small peak was also present at a wavenumber of 725 cm1 characteristic of the stretching vibration of the P–O–P bond in dicalcium pyrophosphate dihydrate. As the ageing period in both PBS and bovine serum was increased, there was almost no variation in the FTIR spectra (Fig. 5), indicating little chemical change occurred in the cement. Rietveld refinement of X-ray diffraction patterns of the powdered cement combined with 20 wt% crystalline anatase enabled the calculation of the amorphous fraction in the cements. Before ageing the cement consisted of 28 wt% amorphous material (Table 4). With ageing the amorphous fraction of the cement was reduced to 24 wt% after 28 days of ageing in PBS and 16 wt% after 90 days of ageing in serum. These data enabled a recalculation of Rietveld refinement results of crystalline phase composition (Table 3). It can be seen that the crystalline composition of the cement hardly changed after 90 days ageing in serum, despite a considerable loss of the amorphous component (Fig. 6). 4. Discussion
80
60
40 0
(B)
2181
20
40 60 Ageing Time (days)
80
100
Fig. 4. (A) The mass loss profile for the b-TCP–pyrophosphoric acid solution cement (powder to liquid ratio of 1.75 g/mL), aged in daily refreshed PBS at a liquid to cement volume ratio of 60, compared with data for brushite cement containing orthophosphate only. (B) The mass loss profile for the b-TCP–pyrophosphoric acid solution cement (powder to liquid ratio of 1.75 g/mL), aged in daily refreshed bovine serum at a liquid to cement volume ratio of 60, compared with data for brushite cement containing orthophosphate only.
Cement strength may be influenced by a number of factors including relative porosity, critical flaw size, crystal morphology, degree of conversion and homogeneity of the cement matrix [15,16]. Of these factors, varying powder to liquid ratio is the most likely to have had the largest affect on degree of conversion and relative porosity. Since the relationship between porosity and strength is inverse logarithmic, a small reduction in relative porosity might be expected to result in a large gain in compressive strength as was observed in this study (Fig. 2). However, contrary to expectations, it was found that the optimum compressive strength exhibited by the b-TCP–pyrophosphoric acid cement was not that in which the reactants had converted most completely (Table 1). When the b-TCP–pyrophosphoric acid cement mixed to a powder to liquid ratio of 1.75 g/mL was aged in PBS, the
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
2182
Table 3 The crystalline compositions of the b-TCP–pyrophosphoric acid cements mixed to a powder to liquid ratio of 1.75 g/mL and aged in daily refreshed (a) PBS or (b) bovine serum at an LCVR of 60 for up to 90 days (a) Time (days)
PBS Composition (wt%)
0 3 7 14 28 90 (b) Time (days)
Brushite
b-TCP
DCPP
Rietveld P (%)
8472 7672 7372 7172 7572 7272 Bovine serum
1272 2172 2672 2772 2472 2772
472 372 o2 o2 o2 o2
6.80 5.14 6.98 6.22 4.84 6.19
Brushite
b-TCP
DCPP
Rietveld Rwp (%)
8472 7872 7272 8072 7972 7572
1272 2072 2772 1972 2072 2472
472 272 o2 o2 o2 o2
6.80 5.18 7.15 4.14 4.79 6.19
Composition (wt%)
0 3 7 14 28 90
65 1126
1070 1004
Unaged Aged 28 d PBS Aged 90d BS
989
60 55
872 725
50 45 40
Absorbance (AU)
1140
35 30
1200
1100
1000
900
800
25 700
Wavenumber (cm-1)
Fig. 5. Fourier transform infrared spectra of b-TCP–pyrophosphoric acid solution cement prior to ageing and after 28 days of ageing in PBS and 90 days of ageing in bovine serum.
compressive strength exhibited by the cement reduced significantly ðpo0:05Þ with time from 1472 MPa prior to ageing to 873 MPa after 28 days of ageing (Table 2). The reduction in compressive strength was likely to have been caused by dissolution of the brushite and calcium pyrophosphate components of the cement resulting in an increase in the volume of the cement contributed by relative porosity from 2171% to 3671% (Table 2). A similar reduction in compressive strength was shown to occur in a brushite cement formed from b-TCP and monocalcium phosphate monohydrate following ageing in water [17]. The calculated calcium to phosphorous molar ratio (Ca/P) of the crystalline components (Table 3) in the set
b-TCP–pyrophosphoric acid cement (Ca/P ¼ 1.07) was different to the Ca/P actually used in the cement mix (Ca/P ¼ 0.98). This disparity suggested that a cement component not detectable using X-ray diffraction had a Ca/P ratio o1. Knowing the proportion of amorphous material in the set cement (Table 4) and the volume of excess water (from porosity values in Table 2) the Ca/P of this amorphous fraction could be calculated as being 0.72. Subsequent measurements in our laboratory have shown amorphous calcium pyrophosphates to have Ca/P ratios in the order of 0.7–0.8, which ties in with the values measured here. Amorphous calcium pyrophosphate and amorphous calcium phosphate however are unstable in aqueous conditions [18,19] and so would not be expected to persist after long periods of ageing as indicated in Table 4. However pyrophosphate ions are known to stabilise amorphous calcium orthophosphates in vitro [19] and such ortho–pyrophosphate complexes may well have formed in this cement. Another possibility is that pyrophosphate ions formed a coating across the surface of the brushite crystals. It has previously been reported that such a layer may be formed on the surface of other calcium orthophosphate crystals and that this layer of pyrophosphate ions reduced solubility rates [20]. In a previous study by our group [5], a brushite cement formed from the combination of b-TCP with orthophosphoric acid (2 M at a powder to liquid ratio of 1.75 g/mL), rather than pyrophosphoric acid, was aged in an identical model to that used here. Hydroxyapatite was detected in the set cement 14 days after immersion in PBS and this was shown to result in a reduction in the rate at which mass was lost from the cement (Fig. 4a). Other studies have reported
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
2183
Table 4 The proportion of the set cement comprised by an amorphous product as determined by Rietveld refinement of X-ray diffraction data Time (days)
Ageing medium
0 28 90
None PBS Serum
Amorphous fraction (wt%)
Corrected crystalline composition (wt%)
2872 2472 1672
Brushite
b-TCP
DCPP
Rietveld Rwp (%)
6072 5772 6372
972 1872 2072
372 42 42
10.12 10.44 12.98
Mass (as a percentage of the original)
Also shown, the corrected crystalline phase compositions taken from Table 3.
100 90 80 Hydroxyapatite detected
70 60
PBS
50
Serum
40 30 0
5
10
15 Time (days)
20
25
30
Fig. 6. A mass loss profile for a previously reported b-TCP—orthophosphoric acid solution brushite cement (powder to liquid ratio 1.75 g/ml) aged in daily refreshed PBS or serum at an LCVR of 60 for 28 days, taken from Grover et al. (2003) [5]. Indicated on the mass loss profile is the time period at which hydroxyapatite was first detected in the matrix of the set cement.
the formation of hydroxyapatite in a brushite cement as soon as 72 h after initial immersion in PBS, with the brushite converting completely within a period of 19 days of ageing [21]. This was not the case in the present study, the brushite component of the cement did not hydrolyse to form apatite for the duration of the experiment (Table 3a) when aged in PBS. As a consequence, the cement lost mass throughout the experiment at a consistent rate resulting in a more linear mass loss profile (Fig. 4a). Thus it appeared that the pyrophosphate ions in the cement matrix may have inhibited hydroxyapatite formation for the duration of the study. As it is the formation of hydroxyapatite in the matrices of brushite cements that has been shown to cause long-term stability of the grafts, it is possible that the stabilisation of the cement by pyrophosphate ions may allow complete resorption of the b-TCP–pyrophosphoric acid cement in vivo. Other workers have added inhibitors of hydroxyapatite formation to brushite cement mixes in the hope of preventing brushite hydrolysis. Rosseau et al. [8] added a magnesium salt (8.5 wt% MgHPO4 3H2O) to a brushite cement mix and showed that it prevented hydroxyapatite formation when aged in Hank’s solution (a medium containing no protein), however when cement consisting of a similar quantity of MgHPO4 3H2O
(5 wt%) was implanted in sheep, hydroxyapatite formation was shown to occur after 4 months of implantation [9]. The difference between the in vitro and in vivo findings may have been because of the fact that while magnesium is a well documented inhibitor of hydroxyapatite formation in vitro, it is a co-factor required for the activation of alkaline phosphatase [22]. Alkaline phosphatase is an enzyme known to hydrolyse phospho-ester bonds of a variety of substrates including inorganic pyrophosphate ions causing localised biomineralisation [23,24]. There was a stark contrast between the total mass loss from the b-TCP–pyrophosphoric acid cement when aged for 90 days in bovine serum (57 wt%) or PBS (16 wt%). A similar difference was reported following the ageing of bTCP–orthophosphoric acid cement in PBS or serum for 28 days (Fig. 4b) and was thought to be caused by the inhibition of hydroxyapatite formation in the cement aged in serum allowing brushite dissolution and subsequent fragmentation of the cement [5]. Since no compositional differences between the b-TCP–pyrophosphoric acid solution cements aged in serum or PBS were observed in the work reported here, the difference in weight loss in this new pyrophosphate containing brushite cement could be attributed to the difference in the solubility of the cement matrix in different ageing media. The fact that the brushite and dicalcium pyrophosphate dihydrate phases dissolved more rapidly in serum than in PBS is surprising since bovine serum contains 50 and 90 ppm Ca2+ and similar levels of these ions have been reported to reduce the rate that brushite is dissolved by up to 98% of the rate that brushite dissolves in medium containing no Ca2+ (such as PBS) [25]. It appeared probable, that another serum component accelerated the rate of dissolution of the brushite and calcium pyrophosphate phases present in the b-TCP–pyrophosphoric acid cement. Serum contains a number of enzymes with pyrophosphatase activity which may have accelerated the dissolution of calcium pyrophosphates, thus increasing the rate of cement degradation compared to buffered enzyme free conditions (Fig. 4a). Enzymes present in serum that exhibit pyrophosphatase activity include tissue non-specific alkaline phosphatase and inorganic pyrophosphatases [23]. Pritzker and co-workers have demonstrated that yeast derived alkaline phosphatase [26] and tissue non-specific alkaline phosphatase produced by chondrocytes [27]
ARTICLE IN PRESS 2184
L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
accelerate the dissolution of calcium pyrophosphate crystals in vitro. Using alkaline phosphatase labelled with fluorescein isothiocyanate [28], they demonstrated that the presence of etch pits indicative of dissolution were associated with the binding position of the alkaline phosphatase, which was shown to bind selectively to the (010) faces of the dicalcium pyrophosphate dihydrate crystals. Calcium polyphosphate ((PO3)n)-based materials, (mainly glasses), have been evaluated as biomaterials previously [e.g. 29]. Like pyrophosphates they are potent crystallisation inhibitors when in a soluble form and can be hydrolysed to form an orthophosphate group and a shorter chain of length n1. In contrast, cleavage of the P–O–P bond in a pyrophosphate ion results in the production of two orthophosphate ions. Hence the action of alkaline phosphatase on pyrophosphate ions in vivo is two-fold: removal of the crystallisation inhibitor and production of only orthophosphate ions resulting in mineralisation [23,24]. This study shows that degradation of pyrophosphate containing brushite cements was accelerated by a serum constituent and not by inhibition of insoluble apatite formation as is the case in purely orthophosphate-based brushite cements [5]. To our knowledge, this is the first report of biologically mediated degradation of an inorganic biomaterial, rather than simple dissolution as is the case with gypsum and a-TCP-based bioceramics.
5. Conclusion We have demonstrated that by using pyrophosphoric acid in a cement paste, it is possible to form a matrix consisting of crystalline and amorphous components. The pyrophosphate ions may have been associated with the amorphous components or have been absorbed to the surface of the brushite crystals. The presence of the amorphous phase and/or the pyrophosphate ions prevented the formation of hydroxyapatite in the cement matrix during in vitro ageing, as has previously been shown to occur in purely orthophosphate-based brushite cements, resulting in a linear mass loss for the duration of the experiment. Having demonstrated the potential of biologically mediated control of an inorganic biomaterial’s degradation for the first time, further work will seek to precisely identify key parameters for predictable manipulation of this phenomenon.
Acknowledgments The authors acknowledge the financial support of the EPSRC (LMG) and the provision of a CASE studentship by Smith and Nephew Group Research Centre, York, UK. MJT is grateful to the Royal Society for the award of a University Research Fellowship.
Appendix A. Supplementary materials Supplementary data associated with this article can be found in the on-line version at doi:10.1016/j.biomaterials.2005.11.012.
References [1] Ohura K, Bohner M, Hardouin P, Lemaitre J, Pasquier G, Flautre B. Resorption of, and bone formation from, new b-tricalcium phosphate–monocalcium phosphate cements: an in vivo study. J Biomed Mater Res 1996;30:193–200. [2] Fraysinnet P, Gineste L, Conte P, Fages J, Rouquet N. Short-term implantation effects of a DCPD based calcium phosphate cement. Biomaterials 1998;19:971–7. [3] Flautre B, Delecourt C, Blary MC, Van Landuyt P, Lemaitre J, Hardouin P. Volume effect on biological properties of a calcium phosphate hydraulic cement: experimental study in sheep. Bone 1999;2:35S–9S. [4] Penel G, Leroy N, Van Landuyt P, Flautre B, Hardouin P, Lemaitre J, et al. Raman microspectrometry studies of brushite cement: in vivo evaluation in a sheep model. Bone 1999;25:81S–4S. [5] Grover LM, Knowles JC, Fleming GJP, Barralet JE. In-vitro ageing of a brushite calcium phosphate cement. Biomaterials 2003;24: 4133–41. [6] Bohner M, Matter S. Brushite hydraulic cement stabilized with a magnesium salt. US Patent number 6,733,582, 2004. [7] Amjad Z, Koutsoukos PG, Nancollas GH. The crystallisation of hydroxyapatite and fluorapatite in the presence of magnesium ions. J Colloid Interf Sci 1984;101:250–6. [8] Rosseau S, Lemaitre J, Bohner M, Frei C. Long term ageing of brushite cements in physiological conditions: an in-vitro study. In: Proceedings of the GRIBOI, Shanghai, 2002. [9] Bohner M, Theiss F, Apelt D, Hirsiger W, Houriet R, Rizzoli G, et al. Compositional changes of a dicalcium phosphate dihydrate cement after implantation in sheep. Biomaterials 2003;24:3463–74. [10] Eidelman N, Brown WE, Meyer JL. Selective inhibition of crystal growth on octacalcium phosphate and nonstoichiometric hydroxyapatite by pyrophosphate at physiological concentrations. J Cryst Growth 1991;113:643–52. [11] Grover LM, Gbureck U, Wright AJ, Barralet JE. Cement formulations in the calcium phosphate H2O–H3PO4–H4P2O7 system. J Am Ceram Soc 2005;88:3096–103. [12] Curry NA, Jones DW. Crystal structure of brushite, calcium hydrogen orthophosphate dihydrate: a neutron diffraction investigation. J Chem Soc A 1971:3725–9. [13] Dickens B, Schroeder LW, Brown WE. Crystallographic studies of the role of Mg as a stabilising impurity in b-Ca3(PO4)2 I. The crystal structure of pure b-Ca3(PO4)2. J Solid State Chem 1974;10:232–48. [14] Mandel NS. The crystal structure of calcium pyrophosphate dihydrate. Acta Crystallog B 1975;31:1730–4. [15] Ishikawa K, Asaoka K. Estimation of ideal mechanical strength and critical porosity of calcium phosphate cement. J Biomed Mater Res 1995;29:1537–43. [16] Ishikawa K, Takagi S, Chow LC, Suzuki K. Reaction of calcium phosphates with different amounts of tetracalcium phosphate and dicalcium phosphate anhydrous. J Biomed Mater Res 1999;46: 504–10. [17] Mirtchi AA, Lemaitre J. Calcium phosphate cements: study of the btricalcium phosphate–monocalcium phosphate system. Biomaterials 1989;10:475–80. [18] Brown EH, Lehr JR, Smith JP, Frazier AW. Preparation and characterisation of some calcium pyrophosphates. Agri Food Chem 1963;11:214–22.
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185 [19] Elliott JC. Structure and chemistry of the apatites and other calcium orthophosphates. Amsterdam: Elsevier; 1994. p. 55–58. [20] Fleisch H, Maerki J, Russell RGG. Effect of pyrophosphates on dissolution of hydroxyapatite and its possible importance in calcium homeostasis. Proc Soc Exp Biol Med 1966;122:317–20. [21] Constantz BR, Barr BM, Ison IC, Fulmer MT, Baker J, McKinney L, et al. Histological, chemical and crystallographic analysis of four calcium phosphate cements in different rabbit osseus sites. J Biomed Mater Res 1998;43:451–61. [22] Terkeltaub RA. Inorganic pyrophosphate generation and disposition in pathophysiology. Am J Physiol Cell Physiol 2001;281:C1–C11. [23] Johnson KA, Hessle L, Vaingankar S, Wennberg C, Mauro S, Narisawa S, et al. Osteoblast tissue-nonspecific alkaline phosphatase antagonises and regulates PC-1. Am J Physiol Regul Integr Comp Physiol 2000;279:R1365–77. [24] Murshed M, Harmey D, Millan JL, McKee MD, Karsenty G. Unique coexpression in osteoblasts of broadly expressed genes
[25] [26]
[27]
[28]
[29]
2185
accounts for the spatial restriction of ECM mineralization to bone. Genes Dev 2005;19:1093–104. Xie J, Riley C, Chittur K. Effect of albumin on brushite transformation to hydroxyapatite. J Biomed Mater Res 2001;57:357–65. Xu Y, Cruz TF, Pritzker KPH. Alkaline-phosphatase dissolves calcium pyrophosphate dihydrate crystals. J Rheumatol 1991;18: 1606–10. Xu Y, Pritzker KPH, Cruz TF. Characterization of chondrocyte alkaline-phosphatase as a potential mediator in the dissolution of calcium pyrophosphate dihydrate crystals. J Rheumatol 1994;21:912–9. Shinozaki T, Xu Y, Cruz TF, Pritzker KPH. Calcium pyrophosphate dihydrate (CPPD) crystal dissolution by alkaline-phosphatase— interaction of alkaline-phosphatase on CPPD crystals. J Rheumatol 1995;22:117–23. Grynpas MD, Pillia RM, Kandel RA, Renlund R, Filiaggi M, Dumitrui M. Porous calcium polyphosphate scaffolds for bone substitute applications in vivo studies. Biomaterials 2002;23:2063–70.
Biomaterials 27 (2006) 2178–2185 www.elsevier.com/locate/biomaterials
Biologically mediated resorption of brushite cement in vitro Liam M. Grovera,, Uwe Gbureckb, Adrian J. Wrightc, Maryjane Tremaynec, Jake E. Barraleta a Faculty of Dentistry, McGill University, 3640 University Street, Montre´al, Que´bec, Canada H3A 2B2 Department of Functional Materials in Medicine and Dentistry, University of Wurzburg, Wurzburg D-97070, Germany c School of Chemistry, University of Birmingham, Birmingham B15 2TT, UK
b
Received 12 August 2005; accepted 6 November 2005
Abstract A new calcium phosphate cement is reported, which sets to form a matrix consisting of brushite, dicalcium pyrophosphate dihydrate and an amorphous phase following the mixture of b-tricalcium phosphate with an aqueous pyrophosphoric acid solution. This reactant combination set within a clinically relevant time-frame (approximately 10 min) and exhibited a higher compressive strength (25 MPa) than previously reported brushite cements. The in vitro degradation of the b-tricalcium phosphate–pyrophosphoric acid cement was tested in both phosphate buffered saline and bovine serum. The pyrophosphate ion containing cement reported here was found not to be hydrolysed to form hydroxyapatite in vitro like b-tricalcium phosphate–orthophosphoric acid solution cements. This finding is significant since the formation of hydroxyapatite by hydrolysis is thought to retard in vivo degradation of brushite cements. When aged in bovine serum, the cement lost considerably more mass than when aged in phosphate buffered saline, indicating that proteins, most likely phosphatase enzymes played an important role in the degradation. As pyrophosphate ions are thought to be the source of orthophosphate ions during bone mineralisation, this new class of bone cement offers a route to new degradable synthetic bone grafting materials. r 2005 Elsevier Ltd. All rights reserved. Keywords: Calcium phosphate cement; Alkaline phosphatase; Brushite; Degradation; In vitro test; Pyrophosphate
1. Introduction In vivo studies investigating the biological reaction to and degradation of brushite cements have reported complete or extensive resorption [1], in addition to fragmentation [2] or long-term stability of the cement [3]. Crystallographic and spectroscopic analyses of retrieved brushite cement implants have shown that a marked reduction in the rate of resorption occurs following the formation of hydroxyapatite in the cement, the presence of which is thought to be caused by hydrolysis of the brushite [4]. A previous in vitro ageing study by our group has shown that the rate of mass loss from brushite cement and the formation of hydroxyapatite are affected strongly by ageing medium refreshment rate, the volume of liquid to which the cement is exposed and the presence of protein in Corresponding authors. Tel.: +1 514 398 7203; fax: +1 514 398 8900.
E-mail address: [email protected] (L.M. Grover). 0142-9612/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2005.11.012
the ageing media (which inhibited hydrolysis) [5]. As potential resorption is the major advantage that brushitebased calcium phosphate cements have over hydroxyapatite cements, attempts have been made to prevent brushite hydrolysis in the cement. One approach was the addition of magnesium ions to the cement mix [6]. Magnesium is a potent inhibitor of hydroxyapatite crystallisation since it adsorbs to the surface of newly forming hydroxyapatite crystal nuclei and blocks active growth sites [7]. In an in vitro study, a magnesium salt (8.5 wt% MgHPO4 3H2O) was shown to prevent hydroxyapatite formation when a brushite cement was aged in Hank’s solution [8]. Following the implantation of a similar cement formulation containing 5 wt% MgHPO4 3H2O in the distal and proximal humerus of Swiss Alpine sheep, however, hydroxyapatite formation was noted after 4 months [9]. Another apatite crystallisation inhibitor [10], the pyrophosphate (P2O4 7 ) ion has previously been added to brushite cements at very low concentrations as a means of
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
2. Methods and materials The b-tricalcium phosphate (b-TCP)–pyrophosphoric acid solution cement was produced from the combination of b-TCP (Plasma-Biotal, Derbyshire, UK) with 1 mL of a liquid component consisting of 540 mg pyrophosphoric acid (Rhodia, West-Midlands, UK) and 720 mg double distilled water. To determine the range of powder to liquid ratios to which the cement could be mixed, the mass of b-TCP added to the liquid phase was varied between 1.25 and 2.75 g at 250 mg increments (powder to liquid ratios of 1.25–2.75 g/mL). To mix the cement paste, the b-TCP was divided into four equal portions. The b-TCP was then added to the liquid phase a quarter at a time which was mixed using a non-corrodible spatula for a period of 5 s. Once all of the b-TCP was combined with the liquid, the cement paste was kneaded for a further 20 s. The setting times of the cements ðn ¼ 3Þ were determined using the Gilmore needles technique. Samples for compression testing were fabricated by placing manipulated cement slurry into a PTFE split mould forming hardened cement cylinders of diameter 6 mm and height 12 mm. Once set, the samples were removed from the mould and stored in 100% relative humidity at 3771 1C for a period of 24 h. Before testing, geometrical measurements of the cement cylinders were made in triplicate and the samples were weighed. Samples were mounted on the testing machine (5544, Instron, Bucks, UK) so that the long axes of the cement cylinders were perpendicular to the lower anvil. A compressive force was then applied to the upper surface of the cement samples at a constant crosshead displacement rate of 1 mm/min until failure occurred. Mean strength was determined from the average of 10 measurements. The applied load was measured using a 2 kN load cell (5544, Instron, Buckinghamshire, UK). After testing in compression, cement fragments were retrieved, weighed and dried in a vacuum desiccator at a temperature of 3771 1C. The cement fragments were then ground to powder using a pestle and mortar. The true density of the powder was determined using a helium pycnometer (Accupyc 1330, Micromeritics, Bedfordshire, UK). The volume of each sample ðn ¼ 4Þ was measured 10 times following 10 purges of the measurement chamber with helium. The relative porosity of the cement was calculated from apparent and true density measurements. To determine the effects of ageing on the b-TCP–pyrophosphoric acid cement cylinders, samples mixed to a powder to liquid ratio of 1.75 g/mL were aged in PBS (100 mM; Sigma-Aldrich, Dorset, UK) at a liquid to cement volume ratio of 60. The PBS was refreshed on a daily basis to remove any dissolution products. To find the quantity of mass lost from the cement samples with time, samples were removed daily from the ageing medium and weighed. After periods of 3, 14 and 28 days of ageing, cement cylinders were removed from the PBS and tested in compression. Compressive strength values were the mean of 10 measurements. The pore size distribution of the b-TCP–pyrophosphoric acid solution cement prior to and following 28 days of ageing was measured by using a mercury intrusion porosimeter (9420, Micromeritics, Bedfordshire, UK). Since proteins are known to influence the dissolution and crystallisation of calcium phosphate compounds and as such may influence the in vivo
degradation rate of the b-TCP–pyrophosphoric acid cement, cylinders mixed to a powder to liquid ratio of 1.75 g/mL were aged in bovine serum (Sigma-Aldrich, Dorset, UK) containing sodium azide (Sigma-Aldrich, Dorset, UK) at a concentration of 0.1 wt%. The cylinders were aged at a liquid to cement volume ratio of 60 and the bovine serum was refreshed on a daily basis, when three of the samples were weighed. The compositions of the cement samples were determined by means of Rietveld refinement phase analysis of X-ray diffraction data after periods of 3, 7, 14, 28 and 90 days of ageing. Fourier transform infrared spectra of the samples were collected from a KBr (Sigma-Aldrich, Gillingham, UK) pellet at a spectral resolution of 4 cm1. Prior to ageing and following 3, 7, 14, 28 and 90 days of ageing, the composition of the cement cylinders was determined by means of Rietveld refinement analysis of X-ray diffraction data. X-ray powder diffraction data were collected using Cu Ka1 radiation from a X-ray powder diffractometer (D5005, Siemens, Karlsruhe, Germany), over the range 2y ¼ 20–401 with a step size of 0.021 and a normalised count time of 1 s/ step. Quantitative phase compositions of the materials were calculated by means of total Rietveld refinement analysis with the TOPAS software (Bruker AXS, Karlsruhe, Germany) using the structural models of CaHPO4 2H2O (brushite)[12], b-Ca3(PO4)2 (b-TCP) [13] and Ca2(P2O7) 2H2O (DCPP)[14] from the literature together with a Chebychev fourth-order background model [see Supplementary information for an example of a Rietveld profile]. Analysis of the data indicated the presence of an amorphous phase and in order to confirm and quantify this, powdered samples were mixed with 20 wt% crystalline anatase (TiO2). Diffraction data were collected from the anatase/cement powder mixtures and the resultant patterns were refined using the Rietveld method. The proportion of amorphous material was calculated from the apparent excess in anatase wt% due to non-crystalline phases not being included in refinement calculations.
3. Results Initial experimentation revealed that workable cement pastes could be mixed at powder to liquid ratios in the range of 1.25–2.75 g/mL. Increasing the powder to liquid mixing ratios of the cement paste to 41.25 g/mL reduced both the initial and final setting times exhibited by the cement formed with pyrophosphoric acid solution (Fig. 1). When mixed to a powder to liquid ratio of 1.25 g/mL the initial setting time of the cement formed with pyrophosphoric acid solution was 2171 min and the final setting
50 45 40 Setting time (mins)
retarding the setting rate of brushite cements. Our group has recently reported the formation of new cements consisting of brushite and calcium pyrophosphate by the addition of a pyrophosphoric acid solution, (rather than orthophosphoric (PO3 4 ) acid solution) to b-tricalcium phosphate [11]. In the present study the influence of powder to liquid ratio on the setting times and compressive strength of a b-TCP–pyrophosphoric acid cement was determined. The cement formulation which exhibited the best handling properties was selected and aged in vitro for periods of up to 90 days. The influence of ageing the cement in PBS and serum on compressive strength, porosity and the phase composition of the cement were determined.
2179
35 30 25 20 15 10 5 0
1.25
1.5
1.75
2
2.25
2.5
2.75
Powder to liquid ratio (g/ml)
Fig. 1. The initial and final setting times of cement formed from the mixture of b-TCP and pyrohosphoric acid solution mixed to powder to liquid ratios in the range of 1.25–2.75 g/mL.
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
2180
time was 4471 min and when mixed to a powder to liquid mixing ratio of 2.75 g/mL the initial and final setting times of the cements were less than 1 min. While it was possible to mix a setting cement paste at powder to liquid mixing ratios of 2.50 and 2.75 g/mL, these pastes were dry and flaky as established visually. Powder to liquid ratio had a marked influence on the compressive strength of the resulting cements (Fig. 2). An approximate six-fold increase in compressive strength from 471 to 2572 MPa was observed when powder to liquid mixing ratio was increased from 1.25 to 2.25 g/mL. Increasing the powder to liquid mixing ratio to 2.50 g/mL and higher was detrimental to compressive strength; the cement mixed to a powder to liquid mixing ratio of 2.75 g/ mL exhibited a compressive strength of only 1873 MPa. Fig. 2 illustrates that the increase in compressive strength that was observed when powder to liquid ratio was increased from 1.25 to 1.75 g/mL was associated with a reduction in the relative porosity of the set cement from 4171% to 2171%. Although the cements mixed to higher powder to liquid ratios exhibited higher compressive strengths, no further significant reduction in porosity was apparent. Table 1 indicates that increasing powder to liquid ratio to 41.75 g/mL reduced the degree of reaction. The quantity of unreacted b-TCP increased as the powder to liquid ratio increased, thus strength (Fig. 2) and degree of reaction were not directly correlated.
Period of immersion (days)
Compressive strength (MPa)
Relative porosity (%)
0 3 14 28
1472 1574 1173 873
2171 2271 2671 3671
0.03
40 35 30
20
25 15 20 15
10
10
Relative Porosity (%)
25
5 5 1.25
1.5
1.75 2 2.25 2.5 Powder to Liquid Ratio (g/ml)
2.75
0
Fig. 2. The effects of powder to liquid ratio on the compressive strength and relative porosities of cements formed from the mixture of b-TCP with pyrohosphoric acid solution.
Incremental intrusion volume (ml/g)
Compressive Strength Relative Porosity Compressive Strength (MPa)
Table 2 The compressive strengths and relative porosities of b-TCP–pyrophosphoric acid cement mixed to a powder to liquid ratio of 1.75 g/mL, following different periods of ageing in daily refreshed PBS at an LCVR of 60
45
30
0
When aged in PBS, the mean compressive strength of the cement formed with pyrophosphoric acid solution deteriorated with time from 1472 MPa prior to ageing to 873 after 28 days of ageing (Table 2). The reduction in the compressive strength exhibited by the cement was associated with an increase in relative porosity from 2171% to 3671% after 28 days of ageing (Table 2). Mercury intrusion porosimetry showed that prior to immersion in PBS, the majority of pores in the cement were of diameter between 100 nm and 10 mm (Fig. 3). Ageing of the cement in daily refreshed PBS for a period of 28 days resulted in a considerable increase in the total pore volume of the
Prior to immersion After 28 d immersion 0.02
0.01
0 0.001
0.01
0.1 1 10 Pore diameter (µm)
100
Fig. 3. Pore size distributions of b-TCP–pyrophosphoric acid solution cement (mixed to a powder to liquid mixing ratio of 1.75 g/mL) prior to ageing and after 28 days of ageing in daily refreshed PBS at a LCVR of 60.
Table 1 The influence of powder to liquid ratio on the composition of the set b-TCP–pyrophosphoric acid cement P:L (g/mL)
1.25 1.75 2.25 2.75
1000
Rietveld Rwp (%)
Composition (wt%) Brushite
b-TCP
DCPP
8172 8472 5872 4072
1472 1272 3972 5372
572 472 372 772
6.87 6.80 5.61 6.76
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
cement (Table 2, Fig. 3). The aged cement contained a distribution of pores bimodal about 50 nm or 1 mm in diameter (Fig. 3). The mass loss profile for the cement formed with pyrophosphoric acid solution is shown in Fig. 4. For the first 28 days the cement formed using pyrophosphoric acid solution lost mass at a rate of 0.06% per day when stored in PBS refreshed daily, (Fig. 4A), so that after a period of 28 days of ageing the cement had lost a total of 1.8% of its original mass. After 28 days of ageing there was a slight increase in the rate at which the cement lost mass, with the cement losing mass at a rate of 0.25% per day so that after a period of 90 days of ageing the cement lost a total of 19% of its original mass. Ageing using identical parameters (liquid volume and sample size) but in bovine serum resulted in the cement losing mass at a linear rate of 0.6% per day for the entire experiment, so that after a period of 90 days of ageing the cement had lost a total of 57% of its original mass (Fig. 4B), more than three times the mass lost
Relative Residual Mass (%)
100
80 Pyro- cement Ortho- Cement[5] 60
40 0
20
(A)
40 60 Ageing Time (days)
80
100
Relative Residual Mass (%)
100 Pyro- cement Ortho- Cement[5]
when the cement formed using pyrophosphoric acid solution was aged in PBS for the same period. The crystalline compositions of the cement aged in PBS or bovine serum determined from Rietveld refinement are shown in Tables 3a and b, respectively. Given the contrast in mass loss profiles, it might have been expected that the phase compositions of the cement formed using pyrophosphoric acid solution would differ when aged in PBS and when aged in bovine serum, however, surprisingly throughout the experiment there was little difference in the phase compositions of the cement during ageing. In both media, following immersion in the ageing medium there was a small reduction in the proportion of brushite and dicalcium pyrophosphate dihydrate in the cement coupled with an increase in the proportion of the cement contributed by bTCP. Most phase composition change occurred after only 3 days immersion. Fourier transform infrared spectra demonstrated that prior to ageing, there were peaks 1 present on the spectra characteristic of HPO2 4 (872 cm ) 3 1 and PO4 (1070–1140 cm ) groups (Fig. 5), a small peak was also present at a wavenumber of 725 cm1 characteristic of the stretching vibration of the P–O–P bond in dicalcium pyrophosphate dihydrate. As the ageing period in both PBS and bovine serum was increased, there was almost no variation in the FTIR spectra (Fig. 5), indicating little chemical change occurred in the cement. Rietveld refinement of X-ray diffraction patterns of the powdered cement combined with 20 wt% crystalline anatase enabled the calculation of the amorphous fraction in the cements. Before ageing the cement consisted of 28 wt% amorphous material (Table 4). With ageing the amorphous fraction of the cement was reduced to 24 wt% after 28 days of ageing in PBS and 16 wt% after 90 days of ageing in serum. These data enabled a recalculation of Rietveld refinement results of crystalline phase composition (Table 3). It can be seen that the crystalline composition of the cement hardly changed after 90 days ageing in serum, despite a considerable loss of the amorphous component (Fig. 6). 4. Discussion
80
60
40 0
(B)
2181
20
40 60 Ageing Time (days)
80
100
Fig. 4. (A) The mass loss profile for the b-TCP–pyrophosphoric acid solution cement (powder to liquid ratio of 1.75 g/mL), aged in daily refreshed PBS at a liquid to cement volume ratio of 60, compared with data for brushite cement containing orthophosphate only. (B) The mass loss profile for the b-TCP–pyrophosphoric acid solution cement (powder to liquid ratio of 1.75 g/mL), aged in daily refreshed bovine serum at a liquid to cement volume ratio of 60, compared with data for brushite cement containing orthophosphate only.
Cement strength may be influenced by a number of factors including relative porosity, critical flaw size, crystal morphology, degree of conversion and homogeneity of the cement matrix [15,16]. Of these factors, varying powder to liquid ratio is the most likely to have had the largest affect on degree of conversion and relative porosity. Since the relationship between porosity and strength is inverse logarithmic, a small reduction in relative porosity might be expected to result in a large gain in compressive strength as was observed in this study (Fig. 2). However, contrary to expectations, it was found that the optimum compressive strength exhibited by the b-TCP–pyrophosphoric acid cement was not that in which the reactants had converted most completely (Table 1). When the b-TCP–pyrophosphoric acid cement mixed to a powder to liquid ratio of 1.75 g/mL was aged in PBS, the
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
2182
Table 3 The crystalline compositions of the b-TCP–pyrophosphoric acid cements mixed to a powder to liquid ratio of 1.75 g/mL and aged in daily refreshed (a) PBS or (b) bovine serum at an LCVR of 60 for up to 90 days (a) Time (days)
PBS Composition (wt%)
0 3 7 14 28 90 (b) Time (days)
Brushite
b-TCP
DCPP
Rietveld P (%)
8472 7672 7372 7172 7572 7272 Bovine serum
1272 2172 2672 2772 2472 2772
472 372 o2 o2 o2 o2
6.80 5.14 6.98 6.22 4.84 6.19
Brushite
b-TCP
DCPP
Rietveld Rwp (%)
8472 7872 7272 8072 7972 7572
1272 2072 2772 1972 2072 2472
472 272 o2 o2 o2 o2
6.80 5.18 7.15 4.14 4.79 6.19
Composition (wt%)
0 3 7 14 28 90
65 1126
1070 1004
Unaged Aged 28 d PBS Aged 90d BS
989
60 55
872 725
50 45 40
Absorbance (AU)
1140
35 30
1200
1100
1000
900
800
25 700
Wavenumber (cm-1)
Fig. 5. Fourier transform infrared spectra of b-TCP–pyrophosphoric acid solution cement prior to ageing and after 28 days of ageing in PBS and 90 days of ageing in bovine serum.
compressive strength exhibited by the cement reduced significantly ðpo0:05Þ with time from 1472 MPa prior to ageing to 873 MPa after 28 days of ageing (Table 2). The reduction in compressive strength was likely to have been caused by dissolution of the brushite and calcium pyrophosphate components of the cement resulting in an increase in the volume of the cement contributed by relative porosity from 2171% to 3671% (Table 2). A similar reduction in compressive strength was shown to occur in a brushite cement formed from b-TCP and monocalcium phosphate monohydrate following ageing in water [17]. The calculated calcium to phosphorous molar ratio (Ca/P) of the crystalline components (Table 3) in the set
b-TCP–pyrophosphoric acid cement (Ca/P ¼ 1.07) was different to the Ca/P actually used in the cement mix (Ca/P ¼ 0.98). This disparity suggested that a cement component not detectable using X-ray diffraction had a Ca/P ratio o1. Knowing the proportion of amorphous material in the set cement (Table 4) and the volume of excess water (from porosity values in Table 2) the Ca/P of this amorphous fraction could be calculated as being 0.72. Subsequent measurements in our laboratory have shown amorphous calcium pyrophosphates to have Ca/P ratios in the order of 0.7–0.8, which ties in with the values measured here. Amorphous calcium pyrophosphate and amorphous calcium phosphate however are unstable in aqueous conditions [18,19] and so would not be expected to persist after long periods of ageing as indicated in Table 4. However pyrophosphate ions are known to stabilise amorphous calcium orthophosphates in vitro [19] and such ortho–pyrophosphate complexes may well have formed in this cement. Another possibility is that pyrophosphate ions formed a coating across the surface of the brushite crystals. It has previously been reported that such a layer may be formed on the surface of other calcium orthophosphate crystals and that this layer of pyrophosphate ions reduced solubility rates [20]. In a previous study by our group [5], a brushite cement formed from the combination of b-TCP with orthophosphoric acid (2 M at a powder to liquid ratio of 1.75 g/mL), rather than pyrophosphoric acid, was aged in an identical model to that used here. Hydroxyapatite was detected in the set cement 14 days after immersion in PBS and this was shown to result in a reduction in the rate at which mass was lost from the cement (Fig. 4a). Other studies have reported
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
2183
Table 4 The proportion of the set cement comprised by an amorphous product as determined by Rietveld refinement of X-ray diffraction data Time (days)
Ageing medium
0 28 90
None PBS Serum
Amorphous fraction (wt%)
Corrected crystalline composition (wt%)
2872 2472 1672
Brushite
b-TCP
DCPP
Rietveld Rwp (%)
6072 5772 6372
972 1872 2072
372 42 42
10.12 10.44 12.98
Mass (as a percentage of the original)
Also shown, the corrected crystalline phase compositions taken from Table 3.
100 90 80 Hydroxyapatite detected
70 60
PBS
50
Serum
40 30 0
5
10
15 Time (days)
20
25
30
Fig. 6. A mass loss profile for a previously reported b-TCP—orthophosphoric acid solution brushite cement (powder to liquid ratio 1.75 g/ml) aged in daily refreshed PBS or serum at an LCVR of 60 for 28 days, taken from Grover et al. (2003) [5]. Indicated on the mass loss profile is the time period at which hydroxyapatite was first detected in the matrix of the set cement.
the formation of hydroxyapatite in a brushite cement as soon as 72 h after initial immersion in PBS, with the brushite converting completely within a period of 19 days of ageing [21]. This was not the case in the present study, the brushite component of the cement did not hydrolyse to form apatite for the duration of the experiment (Table 3a) when aged in PBS. As a consequence, the cement lost mass throughout the experiment at a consistent rate resulting in a more linear mass loss profile (Fig. 4a). Thus it appeared that the pyrophosphate ions in the cement matrix may have inhibited hydroxyapatite formation for the duration of the study. As it is the formation of hydroxyapatite in the matrices of brushite cements that has been shown to cause long-term stability of the grafts, it is possible that the stabilisation of the cement by pyrophosphate ions may allow complete resorption of the b-TCP–pyrophosphoric acid cement in vivo. Other workers have added inhibitors of hydroxyapatite formation to brushite cement mixes in the hope of preventing brushite hydrolysis. Rosseau et al. [8] added a magnesium salt (8.5 wt% MgHPO4 3H2O) to a brushite cement mix and showed that it prevented hydroxyapatite formation when aged in Hank’s solution (a medium containing no protein), however when cement consisting of a similar quantity of MgHPO4 3H2O
(5 wt%) was implanted in sheep, hydroxyapatite formation was shown to occur after 4 months of implantation [9]. The difference between the in vitro and in vivo findings may have been because of the fact that while magnesium is a well documented inhibitor of hydroxyapatite formation in vitro, it is a co-factor required for the activation of alkaline phosphatase [22]. Alkaline phosphatase is an enzyme known to hydrolyse phospho-ester bonds of a variety of substrates including inorganic pyrophosphate ions causing localised biomineralisation [23,24]. There was a stark contrast between the total mass loss from the b-TCP–pyrophosphoric acid cement when aged for 90 days in bovine serum (57 wt%) or PBS (16 wt%). A similar difference was reported following the ageing of bTCP–orthophosphoric acid cement in PBS or serum for 28 days (Fig. 4b) and was thought to be caused by the inhibition of hydroxyapatite formation in the cement aged in serum allowing brushite dissolution and subsequent fragmentation of the cement [5]. Since no compositional differences between the b-TCP–pyrophosphoric acid solution cements aged in serum or PBS were observed in the work reported here, the difference in weight loss in this new pyrophosphate containing brushite cement could be attributed to the difference in the solubility of the cement matrix in different ageing media. The fact that the brushite and dicalcium pyrophosphate dihydrate phases dissolved more rapidly in serum than in PBS is surprising since bovine serum contains 50 and 90 ppm Ca2+ and similar levels of these ions have been reported to reduce the rate that brushite is dissolved by up to 98% of the rate that brushite dissolves in medium containing no Ca2+ (such as PBS) [25]. It appeared probable, that another serum component accelerated the rate of dissolution of the brushite and calcium pyrophosphate phases present in the b-TCP–pyrophosphoric acid cement. Serum contains a number of enzymes with pyrophosphatase activity which may have accelerated the dissolution of calcium pyrophosphates, thus increasing the rate of cement degradation compared to buffered enzyme free conditions (Fig. 4a). Enzymes present in serum that exhibit pyrophosphatase activity include tissue non-specific alkaline phosphatase and inorganic pyrophosphatases [23]. Pritzker and co-workers have demonstrated that yeast derived alkaline phosphatase [26] and tissue non-specific alkaline phosphatase produced by chondrocytes [27]
ARTICLE IN PRESS 2184
L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185
accelerate the dissolution of calcium pyrophosphate crystals in vitro. Using alkaline phosphatase labelled with fluorescein isothiocyanate [28], they demonstrated that the presence of etch pits indicative of dissolution were associated with the binding position of the alkaline phosphatase, which was shown to bind selectively to the (010) faces of the dicalcium pyrophosphate dihydrate crystals. Calcium polyphosphate ((PO3)n)-based materials, (mainly glasses), have been evaluated as biomaterials previously [e.g. 29]. Like pyrophosphates they are potent crystallisation inhibitors when in a soluble form and can be hydrolysed to form an orthophosphate group and a shorter chain of length n1. In contrast, cleavage of the P–O–P bond in a pyrophosphate ion results in the production of two orthophosphate ions. Hence the action of alkaline phosphatase on pyrophosphate ions in vivo is two-fold: removal of the crystallisation inhibitor and production of only orthophosphate ions resulting in mineralisation [23,24]. This study shows that degradation of pyrophosphate containing brushite cements was accelerated by a serum constituent and not by inhibition of insoluble apatite formation as is the case in purely orthophosphate-based brushite cements [5]. To our knowledge, this is the first report of biologically mediated degradation of an inorganic biomaterial, rather than simple dissolution as is the case with gypsum and a-TCP-based bioceramics.
5. Conclusion We have demonstrated that by using pyrophosphoric acid in a cement paste, it is possible to form a matrix consisting of crystalline and amorphous components. The pyrophosphate ions may have been associated with the amorphous components or have been absorbed to the surface of the brushite crystals. The presence of the amorphous phase and/or the pyrophosphate ions prevented the formation of hydroxyapatite in the cement matrix during in vitro ageing, as has previously been shown to occur in purely orthophosphate-based brushite cements, resulting in a linear mass loss for the duration of the experiment. Having demonstrated the potential of biologically mediated control of an inorganic biomaterial’s degradation for the first time, further work will seek to precisely identify key parameters for predictable manipulation of this phenomenon.
Acknowledgments The authors acknowledge the financial support of the EPSRC (LMG) and the provision of a CASE studentship by Smith and Nephew Group Research Centre, York, UK. MJT is grateful to the Royal Society for the award of a University Research Fellowship.
Appendix A. Supplementary materials Supplementary data associated with this article can be found in the on-line version at doi:10.1016/j.biomaterials.2005.11.012.
References [1] Ohura K, Bohner M, Hardouin P, Lemaitre J, Pasquier G, Flautre B. Resorption of, and bone formation from, new b-tricalcium phosphate–monocalcium phosphate cements: an in vivo study. J Biomed Mater Res 1996;30:193–200. [2] Fraysinnet P, Gineste L, Conte P, Fages J, Rouquet N. Short-term implantation effects of a DCPD based calcium phosphate cement. Biomaterials 1998;19:971–7. [3] Flautre B, Delecourt C, Blary MC, Van Landuyt P, Lemaitre J, Hardouin P. Volume effect on biological properties of a calcium phosphate hydraulic cement: experimental study in sheep. Bone 1999;2:35S–9S. [4] Penel G, Leroy N, Van Landuyt P, Flautre B, Hardouin P, Lemaitre J, et al. Raman microspectrometry studies of brushite cement: in vivo evaluation in a sheep model. Bone 1999;25:81S–4S. [5] Grover LM, Knowles JC, Fleming GJP, Barralet JE. In-vitro ageing of a brushite calcium phosphate cement. Biomaterials 2003;24: 4133–41. [6] Bohner M, Matter S. Brushite hydraulic cement stabilized with a magnesium salt. US Patent number 6,733,582, 2004. [7] Amjad Z, Koutsoukos PG, Nancollas GH. The crystallisation of hydroxyapatite and fluorapatite in the presence of magnesium ions. J Colloid Interf Sci 1984;101:250–6. [8] Rosseau S, Lemaitre J, Bohner M, Frei C. Long term ageing of brushite cements in physiological conditions: an in-vitro study. In: Proceedings of the GRIBOI, Shanghai, 2002. [9] Bohner M, Theiss F, Apelt D, Hirsiger W, Houriet R, Rizzoli G, et al. Compositional changes of a dicalcium phosphate dihydrate cement after implantation in sheep. Biomaterials 2003;24:3463–74. [10] Eidelman N, Brown WE, Meyer JL. Selective inhibition of crystal growth on octacalcium phosphate and nonstoichiometric hydroxyapatite by pyrophosphate at physiological concentrations. J Cryst Growth 1991;113:643–52. [11] Grover LM, Gbureck U, Wright AJ, Barralet JE. Cement formulations in the calcium phosphate H2O–H3PO4–H4P2O7 system. J Am Ceram Soc 2005;88:3096–103. [12] Curry NA, Jones DW. Crystal structure of brushite, calcium hydrogen orthophosphate dihydrate: a neutron diffraction investigation. J Chem Soc A 1971:3725–9. [13] Dickens B, Schroeder LW, Brown WE. Crystallographic studies of the role of Mg as a stabilising impurity in b-Ca3(PO4)2 I. The crystal structure of pure b-Ca3(PO4)2. J Solid State Chem 1974;10:232–48. [14] Mandel NS. The crystal structure of calcium pyrophosphate dihydrate. Acta Crystallog B 1975;31:1730–4. [15] Ishikawa K, Asaoka K. Estimation of ideal mechanical strength and critical porosity of calcium phosphate cement. J Biomed Mater Res 1995;29:1537–43. [16] Ishikawa K, Takagi S, Chow LC, Suzuki K. Reaction of calcium phosphates with different amounts of tetracalcium phosphate and dicalcium phosphate anhydrous. J Biomed Mater Res 1999;46: 504–10. [17] Mirtchi AA, Lemaitre J. Calcium phosphate cements: study of the btricalcium phosphate–monocalcium phosphate system. Biomaterials 1989;10:475–80. [18] Brown EH, Lehr JR, Smith JP, Frazier AW. Preparation and characterisation of some calcium pyrophosphates. Agri Food Chem 1963;11:214–22.
ARTICLE IN PRESS L.M. Grover et al. / Biomaterials 27 (2006) 2178–2185 [19] Elliott JC. Structure and chemistry of the apatites and other calcium orthophosphates. Amsterdam: Elsevier; 1994. p. 55–58. [20] Fleisch H, Maerki J, Russell RGG. Effect of pyrophosphates on dissolution of hydroxyapatite and its possible importance in calcium homeostasis. Proc Soc Exp Biol Med 1966;122:317–20. [21] Constantz BR, Barr BM, Ison IC, Fulmer MT, Baker J, McKinney L, et al. Histological, chemical and crystallographic analysis of four calcium phosphate cements in different rabbit osseus sites. J Biomed Mater Res 1998;43:451–61. [22] Terkeltaub RA. Inorganic pyrophosphate generation and disposition in pathophysiology. Am J Physiol Cell Physiol 2001;281:C1–C11. [23] Johnson KA, Hessle L, Vaingankar S, Wennberg C, Mauro S, Narisawa S, et al. Osteoblast tissue-nonspecific alkaline phosphatase antagonises and regulates PC-1. Am J Physiol Regul Integr Comp Physiol 2000;279:R1365–77. [24] Murshed M, Harmey D, Millan JL, McKee MD, Karsenty G. Unique coexpression in osteoblasts of broadly expressed genes
[25] [26]
[27]
[28]
[29]
2185
accounts for the spatial restriction of ECM mineralization to bone. Genes Dev 2005;19:1093–104. Xie J, Riley C, Chittur K. Effect of albumin on brushite transformation to hydroxyapatite. J Biomed Mater Res 2001;57:357–65. Xu Y, Cruz TF, Pritzker KPH. Alkaline-phosphatase dissolves calcium pyrophosphate dihydrate crystals. J Rheumatol 1991;18: 1606–10. Xu Y, Pritzker KPH, Cruz TF. Characterization of chondrocyte alkaline-phosphatase as a potential mediator in the dissolution of calcium pyrophosphate dihydrate crystals. J Rheumatol 1994;21:912–9. Shinozaki T, Xu Y, Cruz TF, Pritzker KPH. Calcium pyrophosphate dihydrate (CPPD) crystal dissolution by alkaline-phosphatase— interaction of alkaline-phosphatase on CPPD crystals. J Rheumatol 1995;22:117–23. Grynpas MD, Pillia RM, Kandel RA, Renlund R, Filiaggi M, Dumitrui M. Porous calcium polyphosphate scaffolds for bone substitute applications in vivo studies. Biomaterials 2002;23:2063–70.