A Gallium-doped cement for the treatment of bone cancers. The effect of ZnO ↔ Ga2O3 substitution of an ionomeric glass series on the rheological, mechanical, pH and ion-eluting properties of their corresponding glass polyalkenoate cements

The primary treatment for patients suffering from bone cancers is resection of the tumour followed by reconstruction of the damaged bone. Despite the administration of post-operative chemotherapy, tumour recurrence continues to present itself as a severe complication leading to re-operation. Attempts to incorporate chemotherapeutic drugs into bone cements elicits local toxic effects on healthy bone, which could compromise implant fixation. Alternatively, the local administration of gallium (Ga) may prove to be more effective. This report considers the development of a Ga ionomeric glass series (0.48SiO2-0.355ZnO-0.06CaO-0.08SrO-0.02P2O5-0.005Ta2O5, with 0.01–0.05 mol% substitution for ZnO). X-ray Diffraction (XRD) confirmed the amorphous glass structure and Energy Dispersive x-ray Fluorescence (EDXRF) verified the successful addition of Ga into the glass series at the expense of Zinc (Zn). A Ga-GPC series was then formulated by mixing the glass particles with aqueous poly(acrylic) acid (PAA) and trisodium citrate (TSC). Fourier transform infrared (FTIR) spectroscopy demonstrated no structural changes to the GPC matrix with the incorporation of Ga. Measurements of the rheological properties demonstrated an exponential increase in setting time with increasing Ga content. Furthermore, the addition of ≥ 3 mol% Ga demonstrated deleterious effects on the GPC’s mechanical properties and an analysis of pH confirmed that it decreased with increasing Ga content, suggesting a reduction in glass reactivity and PAA cross-linking. Finally, inductively coupled plasma—optical emission spectrometry (ICP-OES) demonstrated the controlled release of Ga across the GPC series, which will prove beneficial to future in vitro studies.

Inspired by the incorporation of antibiotics into bone cement, researchers have investigated the employment of PMMA bone cement as a local antineoplastic drug (ANPD) carrier; however, these drugs have low therapeutic indices and demonstrable risks around bone necrosis and possible loss of implant fixation [20][21][22][23][24]. Furthermore, although PMMA bone cement is the gold standard for orthopedic fixation, it possesses a native arsenal of cytotoxic effects (thermal necrosis, monomer toxicity and periprosthetic osteolysis) and is bioinert, thus presenting the opportunity to investigate new materials to repair bones damaged by cancer [20]. Alternatively, inorganic bone cements, including calcium phosphate cements (CPCs) and glass polyalkenoate cements (GPCs), have also been investigated as local ANPD carriers as it was hypothesized that their biological properties might oppose the toxic nature of these drugs and support osteogenesis [25][26][27][28][29][30][31][32]. Unfortunately, studies on ANPD-loaded CPCs demonstrated similar cytotoxic effects on healthy bone [27,31]; moreover, the use of resorbable calcium phosphate poses a contradiction, as its introduction to a lytic area, such as metastatic bone lesions, poses the risk of initiating or aggravating hypercalcemia. Furthermore, an ANPD-loaded GPC showed cytotoxic effects on NIH-3T3 mouse fibroblasts [32]. The current knowledge of chemotherapeutic bone cements demonstrates that, regardless of cement type, ANPD toxicity is inevitable and alternative chemotherapeutic agents should be investigated [20].
Hart et al [33] studied the antitumour activity of the salts of inorganic group IIIa metals, concluding that Ga nitrate had the potential for clinical usefulness in treating solid tumours. Furthermore, several studies have noted the synergistic interactions of Ga nitrate with chemotherapeutic drugs [34][35][36][37]. In-depth studies into its anti-cancer role have demonstrated the effect of Ga on the cellular iron metabolism, resulting in the inhibition of ribonucleotide reductase formation, an enzyme responsible for the formation of deoxyribonucleotides [38][39][40]; therefore, halting DNA synthesis and the proliferation of rapidly dividing (malignant) cells. Ga has also been found to inhibit calcium resorption [41,42]. Studies by Warrell et al [43] and Cvitkovic et al [44] demonstrated that Ga nitrate is superior to bisphosphates for inhibiting bone resorption. Furthermore, Ga nitrate is FDAapproved for the treatment of cancer-related hypercalcemia [45]. Therefore, the local administration of Ga by a GPC may prove beneficial to the treatment of such patients.
GPCs have been used in both restorative and orthodontic dentistry for over 40 years and in ear, nose and throat surgery for over 20 years [46][47][48]; however, the presence of aluminum (Al) in the glass phase of such GPCs have inhibited their translation to orthopedic applications [49,50], resulting in the recent development of Alfree GPCs [51][52][53][54][55][56][57][58][59][60]. GPCs are set by an isothermic neutralization reaction between a silicate-based ionomeric glass and aqueous poly (acrylic acid) (PAA). When mixed, the free PAA protons attack the glass particles and liberate cations, which then cross-link the PAA chains, resulting in a polysalt matrix reinforced by reacted and unreacted glass particles [61]. GPCs are considered bioactive due to their release of therapeutic ions loaded into the cement's glass phase [62,63]. Furthermore, two studies have investigated the incorporation of gallium (Ga) into GPC compositions [59,60], demonstrating the effects on the structural properties of their base ionomer silicate glass, as well as the rheologic, mechanical and ion eluting properties of the corresponding GPCs. While considerable amounts of Ga were incorporated (dopant levels of 8 and 16 mol%), these GPCs were only able to elute up to 3 parts per million (ppm) of Ga 3+ , and no attempts were made to investigate the chemotherapeutic potential or cytotoxic effects of these cements in vitro [59].
In previous studies on a related glass series, Alhalawani and colleagues developed a novel SiO 2 -ZnO-CaO-SrO-P 2 O 5 -Ta 2 O 5 based GPC for wrist fracture fixation [53,54,64]. The purpose of this current body of work is to expand on this and incorporate variable amounts of Ga (0-5 mol%), at the expense of zinc (Zn), into the glass phase of these GPCs to tailor them for the treatment of bone cancers. Specifically, this study aims to understand further the role of Ga incorporation on the GPC setting chemistry, pH and their rheological, mechanical and ion release properties.

Materials and methods
2.1. Glass synthesis and cement preparation Six Ga containing ionomeric glass compositions were formulated (table 1). C-TA0 represents the Ga-free (control) glass, while the remaining glasses (C-TA1 through C-TA5) varied from 1-5 mol% Ga, substituted at the expense of Zn. Calculated amounts of the glass precursor, analytical grade reagents (Fisher Scientific, Ottawa, and Sigma-Aldrich, Oakville, both Canada) were weighed out and thoroughly mixed in a glass beaker. The mix was then placed in a platinum crucible and fired in a furnace (Zircar Hot Spot 110, Florida, New York, USA) at 1650°C for 1.5 h. Molten glass was shock quenched into water at room temperature (∼23°C), and the obtained glass frit was dried in an incubator (37°C) for 24 h. The dried glass frit was ball milled and sieved to a maximum particle size of 45 μm. Glass powder was annealed by reaching the temperature of 640°C over 3 h, remaining at the temperature for 12 h, and returning to room temperature over a further 3 h (total, 16 h). Annealing was performed to reduce glass reactivity (relieve internal stresses) and achieve longer GPC working times GPC samples were prepared by thoroughly mixing a ratio of 1 g glass powders, 0.4 g PAA (PAA35, Mw=55 000, Advanced Healthcare Ltd), 0.6 ml deionized (DI) water and 0.1 g tri-sodium citrate (TSC) (Sigma-Aldrich, Oakville, Canada). Briefly, PAA and TSC were mixed with DI water until they had completely dissolved. This aqueous gel was mixed in room temperature (23±1°C) with the glass powder to obtain the final GPC.

X-ray diffraction (XRD)
XRD diffraction patterns of the glass powders were collected using a multipurpose x-ray diffractometer (Malvern Panalytical, Malvern, Worcestershire, UK) with a Cu source, from 10 to 90°, at a scan step size 0.1°, using a generator voltage of 45 kV and a tube current of 40 mA.

Energy dispersive x-ray fluorescence (EDXRF)
EDXRF spectra of the glass powers were obtained using an Epsilon 4 Food and environment EDXRF Spectrometer (Malvern Panalytical, Malvern, Worcestershire, UK) under four different excitation conditions including: using a silver filter with tube voltage of 50 kV, a copper filter with a tube voltage of 50 kV, an aluminum filter with a tube voltage of 12 kV and no filter with a tube voltage of 10 kV. Qualitative analysis was performed to identify the elements present in the sample, and quantitative analysis was performed to obtain the %mass of each identified element. Both quantitative and qualitative analyses were carried out using the Omnian XRF analysis program (Malvern Panalytical, Malvern, Worcestershire, UK).

Rheological properties
The rheological properties (n=3) of the GPCs were recorded at room temperature (23±1°C) using a stopwatch. Working times (Tw) were defined as the time from the start of mixing to when the materials began to exhibit elastic properties. The net setting times (Ts) were determined in accordance with ISO 9917:2007 [65]. Briefly, a mold (7 mm length, 8 mm width and 5 mm height) was filled to the surface with liquid GPC. An indenter (mass=400 g, diameter=1.06 mm) was lowered on the GPCs and was allowed to rest for 5 s. If an indentation was formed on the GPC, then the needle was repositioned to a clean (non-indented) area, and the process was repeated. Ts was determined once no indentation was visible to the naked eye.

Fourier transform infrared (FTIR) spectroscopy
Cement cylinders (diameter, 4 mm & height, 6 mm) were prepared by filling a stainless-steel mold with freshly mixed cement. The mold was then sandwiched between two stainless-steel plates lined with acetate sheets and incubated at 37°C for 100 min. Samples were then immersed in 10 ml of DI water (37°C) and aged for either 1, 7 or 30-days. At the time of testing, samples were ground using a mortar and pestle and FTIR spectra were collected in ambient air (23±1°C) using a Cary 630 FTIR Spectrometer (Agilent Technologies, Santa Clara, CA, USA) via the total reflectance technique with a ZnSe crystal. The analysis was performed in the wavenumbers ranging from 4000 to 650 cm −1 with a spectral resolution of 4 cm −1 .
2.6. Determination of mechanical properties 2.6.1. Determination of compressive strength and Young's modulus The compressive strengths of the six GPCs were determined in accordance with ISO 9917:2007 [65]. Briefly, cement cylinders (diameter, 4 mm & height, 6 mm) were prepared and aged in 10 ml of DI water (37°C) and aged for either 1, 7 or 30-days (n=5). The samples were tested in ambient air (23±1°C) using an Instron 5966 universal Testing Machine (Instron Corp., Norwood, MA, USA) fitted with a ±10 kN load cell with a crosshead  (1)): where, p is the sample failure load (N) sample and d is the sample diameter (mm). The Young's modulus of each GPC was obtained from the elastic region of the stress-strain curves that were collected during the determination of compressive strength.

Determination of biaxial flexural strength
The biaxial flexural strengths of the six GPCs were determined in accordance with the method described by Williams et al [66]. Cement discs (16 mm diameter and 2 mm thick) were prepared by filling silicon rubber molds with freshly mixed cement. The mold was then sandwiched between two stainless-steel plates lined with acetate sheets and incubated at 37°C for 100 min. Samples were then placed in 10 ml of DI water (37°C) and aged for either 1, 7 or 30-days (n=5).
The cement discs were tested in ambient air (23±1°C) using the same universal Testing Machine described in section 2.6.2. with a crosshead speed of 1 mm min −1 . The discs were centered on a three-ball bearing support, and the load was applied as a point to the center of the disc. The biaxial flexural strength (σ b ) was calculated by the following relation: where, F is the sample failure load (N), t is the sample thickness (mm) and a is the radius of the support diameter (3.825 mm).

pH analysis
The effect of Ga incorporation on the pH of the GPC series was measured in ambient conditions (23±1°C) using an Orion VersaStar Pro pH Benchtop Meter (Thermo Fisher Scientific, Waltham, MA) equipped with a flat membrane microelectrode (MI-406, Microelectrodes, Bedford, NH) and a rate reference electrode (MI-041, Microelectrodes). Cement cylinders (diameter, 4 mm; height, 6 mm) were prepared and aged in 10 ml of DI water at 37°C for 1, 7 and 30-days (n=3). Prior to testing, the pH meter was calibrated using pH buffer solution 4.00±0.02 and 7.00±0.02 (Thermo Fisher Scientific, Waltham, MA).

Ion release
The ion release of the GPC series was measured in ambient air (23±1°C) using inductively coupled plasmaoptical emission spectrometry (ICP-OES 5110, Agilent, Santa Clarita, CA, USA). Calibration standards (0, 1, 5, 10 and 30 ppm) were prepared from 1000 ppm stock solutions (Sigma-Aldrich, Oakville, Canada) for the glass composition elements. Cement cylinders were aged in 10 ml of DI water at 37°C and tested after 1, 7 and 30days (n=5).

Statistical analysis
Statistical analysis was performed using Prism GraphPad 9.0 (GraphPad Software. Inc., San Diego, CA, USA). The Shapiro-Wilk test was used to test normality, and parametric one-way ANOVA was used to analyze all data, with the Tukey's test for multiple comparisons. Results were considered statistically significant when p values were <0.05.

Results and discussion
3.1. X-ray diffraction (XRD) The XRD diffraction patterns across the glass series are displayed in figure 1. It is crucial that the structure of these glasses be amorphous, as a crystalline structure, or a glass with crystalline phases, would be resistant to acid attack by the PAA. The diffuse wave in the XRD diffractogram (∼20-40°) represents the amorphous silica glass structure [67]. The peak detected at ∼15°may indicate a carbon-oxide species formed by the carbonate precursors used to form the glasses as Choi et al [68] and Huh et al [69] have previously assigned comparable peaks to microcrystalline cellulose and graphene oxide, respectively. Furthermore, the diffraction patterns remain consistent across the glass series, confirming that any changes observed in the GPC properties may be solely attributed to the incorporation of Ga rather than phase changes in the glass structures.

Energy dispersive x-ray fluorescence (EDXRF)
EDXRF was performed to study the elemental composition of the glass series. Table 2 compares the theoretical (calculated) %mass values with the quantitative data obtained from EDXRF. The results demonstrate the successful incorporation of Ga at the expense of Zn across the glass series (Ga increase from 0%-12.73%mass and Zn decrease from 42.86%-33.84%mass). Furthermore, slight discrepancies were observed between the theoretical and experimental concentrations of each element, the largest of which was observed for P 2 O 5 (mean difference of −4.93% between the E.C. and the T.C.). Other major discrepancies (>1%) between the E.C. and T. C. include ZnO (mean difference of −4.26% mass), SrO (1.48%mass) and SiO 2 (−2.62%mass). The majority of the glass precursor reagents employed the use of the oxide or carbonate of each element (i.e., CaCO 3 , SrCO 3 , SiO 2 etc). However, the reactive nature of P 2 O 5 has led to the use of ammonium dihydrogen phosphate ((NH 4 ) H 2 PO 4 ) as an alternative reagent for the incorporation of phosphorus in this glass series, which may explain the observed discrepancies. While calculating each reagent amount to add to the glass mix, a correction is required when the exact compound is not being used. For instance, no correction was required for the amount of SiO 2 since the exact element was added to the mix, while SrCO 3 required correction due to a discrepancy between its molecular weight (Mw) and the Mw of the compound obtained in the glass (SrO). This correction occurs by multiplying the calculated mass by the ratio of the Mw of the glass compound and the glass precursor. In the majority of cases, this ratio is above 1; however, glass precursor for P 2 O 5 , (NH 4 ) H 2 PO 4 , has a smaller Mw; therefore, the correction ratio is less than 1 and is likely to be the cause of the large discrepancy observed for P 2 O 5 .

Rheological properties
The rheological properties of the GPCs are of significant importance as these properties are assumed to indicate the effect of Ga incorporation on the cement's setting chemistry. These results are in agreement with previous studies conducted on Ga-incorporated GPCs [59,60]. For instance, Wren et al [60] noted that Tw increased from 125-200 s and Ts increased from 150 s to >550 s with the addition of 16 mol% Ga. The substitution of Ga in place of Zn is expected to result in a more stable and electroneutral glass structure since Ga 3+ provides a larger number of positive charges compared to Zn 2+ . This increase in the glass's network stability would ultimately reduce its reactivity, causing a delay in cation liberation and their consequent gelation process with the PAA carboxyl groups (COOH) and an increased Tw and Ts [53]. Furthermore, Wren et al [60] demonstrated that both Tw and Ts decreased by using a higher molecular weight PAA, explained by entangled polymer dynamics and reptation theory, which states that the relaxation time of long polymer chains is proportional to the cube of their molecular mass. However, the incorporation of Ga in place of Zn may also affect this relaxation time.
The Edwards tube concept describes the movement of a polymer chain to be confined to a tube, with an average diameter a, which is defined by the position of other polymer chains [70]. Moreover, the number of monomers in a single entanglement strand is defined as the number of Kuhn monomers, Ne, present in chain size length a, which is calculated using the following relation [70]: where b is the Kuhn length. Furthermore, the reptation time (t rep ) is described by [70]: where z is the friction coefficient corresponding to each monomer unit, K B is the Boltzmann's constant, T is the temperature, and N is the total number of monomer units in the polymer chain. By rearranging equation (3) and substituting into b in equation (4), the following expression is obtained [70]: This relation demonstrates that, t rep is proportional to the square of the average tube diameter (a). Ga has a larger atomic radius than Zn (187 pm and 139 pm, respectively) and a smaller difference in electronegativity with oxygen compared to Zn (1.63 and 1.79, respectively); therefore, it may be assumed that the cross-linking of Ga 3+ and COOin the GPC's polysalt matrix results in a larger bond length compared to Zn cross-linking [61]. Since the tube diameter of a single polymer chain is determined by the presence of other polymer chains, this dispersion of the surrounding strands may cause an increase in its average tube diameter (a). Furthermore, increasing Ga content may increase Ga-COO bonding, further increasing polymer chain dispersion and the average tube diameter, resulting in a larger Ts. While this is a crude analysis of entangled polymer dynamics, it provides a possible explanation for the increase in Ts across the GPC series.

Fourier transform infrared spectroscopic (FTIR)
FTIR spectroscopy was performed to investigate the setting kinetics between the glass powder and the aqueous PAA. The FITR spectra collected for the 1, 7 and 30-day samples are displayed in figures 3(a)-(c), respectively. The broad peak centered at ∼3250 cm −1 is assigned to the O-H stretching of water [71]. The peaks centered at ∼1550 cm −1 , ∼1455 cm −1 , ∼1406 cm −1 and ∼1322 cm −1 are assigned to asymmetric/symmetrically stretching vibration of a carboxyl group bonded to cations that were liberated from the glass structure [53,72,73]. Furthermore, the peak centered at ∼1050 is assigned to Si-O-Si bridges [53]. Alhalawani et al [59] reported the formation of a shoulder peak at ∼1700 cm −1 in the 1-day GPCs mixed with glasses containing 8 mol% and 16 mol% Ga, which was assigned to unreacted COOH functional group of the PAA [73], and did not appear in their spectra at seven and 30-days. The presence of this shoulder peak may be related to the alterations of glass reactivity caused by the substitution of Ga for Zn, as discussed in section 3.3. The relatively high Ga doping of these glasses may have reduced their reactivity enough to result in a high presence of unreacted COOH groups that was detected by FTIR analysis of their corresponding GPCs at early maturation (1-day). In contrast, the lower Ga dopant levels in the present GPC series did not result in the detection of this peak at any time point. Moreover, the FTIR spectra do not demonstrate the evidence of any structural changes to this GPC series with the addition of 1-5 mol% Ga into the glass phase.

Mechanical properties
The compressive strength (σ c ), biaxial flexural strength (σ b ), and Young's modulus (E) of each cement tested over 1, 7 and 30 days are presented in figure 4. These mechanical properties demonstrated two trends: (1) A global increase in mechanical properties as the cements were aged, attributed to continuous ion liberation from unreacted glass particles during GPC maturation that subsequently increases the crosslinking between PAA chains [73,74].
(2) A general decrease in the mechanical properties with increasing Ga over the GPC series. Figure 4(a) displays σ c as a function of Ga content and time. The greatest σ c was observed for C-TA1 after 30days of maturation (∼19.73  2.27 MPa), and the lowest σ c was observed for C-TA5 after 1-day of maturation (∼7.23  1.17 MPa). No significant differences were observed between σ c of C-TA0, C-TA1 and C-TA2 at any of the time points (P>0.05) while C-TA3, C-TA4 and C-TA5 all demonstrated significantly lower σ c when compared to C-TA0 (P<0.05). Moreover, at 1-day, C-TA4 and C-TA5 showed significantly lower σ c compared to C-TA3 (P<0.05), however at 7-and 30-days, no differences were observed between these three GPCs (P>0.05).  (∼4.07  0.14 MPa). Overall, no significant differences were observed between the σ b of C-TA0 and C-TA1 at any time point (P>0.05). Furthermore, C-TA2, C-TA3, C-TA4 and C-TA5 all demonstrated significantly lower σ b than C-TA0 and C-TA1 at both 1-and 7-day time points (P<0.05). However, at 30-days, no differences were observed between C-TA1, C-TA2, C-TA3 and C-TA4 (P>0.05), and the σ b of C-TA5 was observed to be significantly lower than all other GPCs (P<0.05).
The modulus (E) of each GPC was calculated from the stress-strain curve obtained during the determination of compressive strength. Figure 4(c) displays E as a function of Ga concentration and time. C-TA0 demonstrated the greatest E after 7-days of maturation (∼144  16 MPa), while C-TA5 demonstrated the lowest E after 1-day of maturation (∼41  3.6 MPa). No significant differences were observed between C-TA0, C-TA1 and C-TA2 at 1-day (P>0.05), while C-TA3, C-TA4 and C-TA5 were significantly lower than C-TA0 and C-TA1 (P<0.05). At 7-days, the E of all the GPCs containing Ga were significantly lower than C-TA0 (P<0.05). However, by 30days, no significant differences were observed between C-TA0, C-TA1, C-TA2 and C-TA3 (P>0.05), while C-TA4 and C-TA5 demonstrated a significantly lower E compared to the other four GPCs (P<0.05).
The results of the σ c and σ b are in agreement with the results reported in previous studies [59,60], which reported a decrease in these mechanical properties with the incorporation of Ga in their respective GPC series; although, both of the previous studies noticed an increase in the mechanical properties from their 8-to-16 mol% Ga GPC. The current study demonstrates that 1-2 mol% Ga incorporation does not affect this GPC's mechanical properties, while  3 mol% Ga causes deleterious effects. Moreover, a comparison of E values suggests that C-TA0 matures faster, demonstrating no significant increase from 7-to 30-days (P=0.0823), while the GPCs with Ga demonstrated a gradual increase in E through each maturation point (aside from C-TA1, which demonstrated a slight decrease from 1-day to 7-days), all showing significant differences between 7-and 30-day samples (P<0.05). This trend would follow the principle that a reduction in glass reactivity would decrease cross-linking between the glass cations and the PAA chains [73].
Furthermore, this trend may also be explained by the theory of entangled polymer dynamics discussed in section 3.3. [70]. If the difference in molecular diameter and electronegativity between Zn and Ga results in an increased bond length between the polymer chains, then the consequent increase in average tube diameter (a) would also increase the number of Kuhn monomers in a single entanglement stand (N E ) (equation (3)). Furthermore, if the total number of Kuhn monomers in a single polymer chain is represented by N, then the total number of entanglements is taken as the ratio of N E /N. As N E increases, the number of total entanglements experienced by a single polymer strand would decrease and result in deleterious effects on the cement's mechanical properties.

pH analysis
The pH of the GPC series is presented in figure 5. No significant differences were observed between the pH of C-TA0 and C-TA1 after 1-day of maturation, and no significant differences were observed between C-TA0, C-TA1 and C-TA2 after 7 and 30-days of maturation (P>0.05). In contrast, C-TA3, C-TA4 and C-TA5 all demonstrated a significantly lower pH compared to C-TA0, C-TA1 and C-TA2 at all time points (P<0.05). Moreover, no differences were observed between C-TA3 and C-TA4 at any time point. At 7-days and 30-days, C-TA4 demonstrated a significantly lower pH compared to C-TA3 and C-TA5 (P<0.05). Overall, two trends were observed: (1) an increase in pH with time, (2) a general decrease in pH with increasing Ga content. The first trend follows suit with the continuous maturation of the GPCs. When the glass is exposed to aqueous PAA, the hydrogen ions (H + ) attack the glass structure and liberate cations that bond to the PAA chains. Therefore, as this process continues, the total number of free H + decreases, resulting in increased pH [61,73]. Moreover, the general decrease in pH with increasing Ga content demonstrates an increase in free H + , implying a decrease in glass reactivity and PAA cross-linking as discussed in section 3.3.

Ion release
The ion release properties were investigated to understand the effect of Ga incorporation on the solubility of the GPC series. Furthermore, the ion release properties of GPCs are important since the concentration of eluted ions play an essential role in their therapeutic properties. The release profiles of Ga, Zn, Sr and Si were studied only due to their content in the precursor glasses and their therapeutic potential in clinical applications. Figure 6 displays their ion release profiles over 1, 7 and 30 days. In general, the release profiles for each GPC increased over time.
The Ga release profiles (figure 6(a)) demonstrate a controlled release of Ga across the GPC series, ranging from 0-60 ppm. Moreover, the Ga release increased significantly between each GPC at seven and 30-days (P<0.05). Alhalawani et al [59] noted a similar trend in Ga release with increasing Ga content. However, this trend could only achieve a maximum release of ∼3 ppm after 30-days of incubation, which is attributed to their relatively lower Ts (∼250 s) [59]. In contrast, the Ts of this GPC series is much longer (33-96 min) and therefore allows for greater levels of Ga mobility/release before complete crosslinking occurs to the set cement. C-TA1 demonstrated the lowest Ga release (other than the control) after 30-days of incubation (∼3 ppm), and C-TA5 demonstrated the largest release after 1-day of incubation (∼60 ppm). The GPC series in the present study presents a wide range of Ga release, which may be used to evaluate chemotherapeutic potential and alterations to bone metabolism.
The Zn release profiles across the GPC series ( figure 6(b)) are of particular importance since Zn represents the majority of the precursor glass compositions (table 1). Furthermore, Zn possesses anti-bacterial properties and its elution may regulate the occurrence of postoperative infection [53,[75][76][77]. The Zn release of C-TA0 was significantly greater than all other GPCs at each of the time points (P<0.05), suggesting that Zn release decreases with the immediate incorporation of Ga ( 1 mol% Ga) and could be explained by a decrease of glass reactivity with Ga incorporation. Since Zn is the primary cation present in the glass series, it may be responsible for the majority of cross-linking between the PAA chains. Therefore, a decrease in Zn ion release could represent a decrease in their cross-linking between the PAA chains with the inclusion of Ga. However, the Zn release across the five Ga-containing GPCs remains relatively unchanged across the series. Similar results were observed with the Sr elution profile (figure 6(c)), where C-TA0 was found to be significantly higher than all other GPCs after 1 and 7-days (P<0.05); furthermore, the Sr elution from C-TA0 remained significantly higher than C-TA1, C-TA2, C-TA3 and C-TA4 after 30-days maturation. In-depth IR studies during the initial setting period would be required to investigate further the role of Ga, Zn and Sr in the cements setting chemistry.
The Si release profiles are displayed in figure 6(d). The leaching of ions from the glass leaves an ion-depleted silica gel layer around unreacted glass particles [74], and therefore, Si 4+ elution may provide evidence of glass reactivity. No significant differences were observed at 1 and 30-days, however the elution of Si of C-TA0 was significantly higher than all other glasses after 7-days of elution (P<0.05), suggesting that Ga containing GPCs mature at a slower rate.

Conclusion
This study investigated the substitution of Ga at the expense of Zn (dopant levels ranging from 0%-5%) on an ionomer silicate glass. These studies expand on the current knowledge of Ga-incorporated GPCs. Previous studies [59,60] demonstrated the synthesis and analysis of two Ga-loaded GPCs (8 and 16 mol%, at the expense of Zn). GPCs loaded with 3 mol% of Ga demonstrated similar results to the previous studies [59,60], showing increased rheological properties and decreased mechanical properties. These phenomena have been discussed by the difference in elemental properties between Ga and Zn and entangled polymer dynamics using Edward's Tube concept [70]. Furthermore, an analysis of pH confirmed that it decreased with increasing Ga content, suggesting increasing glass network stability leading to decreased reactivity. Finally, the ion release profiles demonstrated a controlled release of Ga across the GPC series, which should prove an advantage in future works when analyzing their dose dependent cytotoxic effects on healthy bone and cancer cells.