Safer and Sustainable-by-Design Hydroxyapatite Nanobiomaterials for Biomedical Applications: Assessment of Environmental Hazards

Developments in the nanotechnology area occur ensuring compliance with regulatory requirements, not only in terms of safety requirements, but also to meet sustainability goals. Hence, safer and sustainable-by-design (SSbD) materials are also aimed for during developmental process. Similar to with any new materials their safety must be assessed. Nanobiomaterials can offer large advantages in the biomedical field, in areas such as tissue repair and regeneration, cancer therapy, etc. For example, although hydroxyapatite-based nanomaterials (nHA) are among the most studied biomaterials, its ecotoxicological effects are mostly unknown. In the present study we investigated the toxicity of seven nHA-based materials, covering both different biomedical applications, e.g., iron-doped hydroxyapatite designed for theragnostic applications), hybrid collagen/hydroxyapatite composites, designed for bone tissue regeneration, and SSbD alternative materials such as titanium-doped hydroxyapatite/alginate composite, designed as sunscreen. The effects were assessed using the soil model Enchytraeus crypticus (Oligochaeta) in the natural standard LUFA 2.2 soil. The assessed endpoints included the 2, 3 and 4 days avoidance behavior (short-term), 28 days survival, size and reproduction (long term based on the OECD standard reproduction test), and 56 days survival and reproduction (longer-term OECD extension). Although overall results showed little to no toxicity among the tested nHA, there was a significant decrease in animals’ size for Ti-containing nHA. Moreover, there was a tendency for higher toxicity at the lowest concentrations (i.e., 100 mg/kg). This requires further investigation to ensure safety.


Introduction
Nanotechnology is a multi-million-dollar industry, estimated to reach a market size of approx. 34 billion dollars by 2030. The diversity of applications ranges from energy and electronics to healthcare, cosmetics, aerospace, etc. [1]. As outlined in the United Nations Sustainable Development Goals 2030, any new material or product should not only be functional and cost-effective but also safe and sustainable. This is to ensure compliance with regulation and promote a clean environment, aligned with a zero-pollution ambition [2]). The implementation of safe and sustainable by design (SSbD) materials [3,4] offers nanotechnology a large advantage, due to the cost of developing and testing materials. Safe and sustainable by design refers to the process of anticipating potential toxic impacts of a product or material on human and/or environmental health, and to address these concerns during the innovation process, i.e., altering the product design if needed before marketing [5]. Nanobiomaterials-natural or synthetic nanoscale materials, that interact with biological systems to perform a specific function, e.g., treat, support, or replace damaged tissue-have a tremendous potential in biomedical applications, e.g., tissue repair and regeneration, drug and gene delivery, cancer therapy, medical imaging, and theragnostics [6]. Hydroxyapatite (HA) is a calcium phosphate-based compound which is among the most studied biomaterials in the medical field due to its biocompatibility and for being the main constituent of the mineral part of bone and teeth [7,8]. The low mechanical properties of HA ceramic are surpassed by its combination with a large number of polymers (e.g., collagen, polylactic acid, alginates, polyethylene) producing nanocomposites with broader applications [8], e.g., photothermal therapy [9]. Despite the great promise of HA based nanomaterials (nHA) in the biomedical field, its ecotoxicity must still be thoroughly assessed to ensure that it is safe and sustainable.
The available data on nHA are often inconsistent (please note that in the vast majority of the studies, the tested nHA were different nHA forms which hampers the comparison). Some in vitro studies report induction of inflammation, but no cytotoxicity to various cell lines [10][11][12]. However, the induction of cyto-toxicity (cell death) following nHA exposure has also been reported [13][14][15]. Cytotoxicity correlated with nHA uptake and cell load, suggesting that the increase in cytoplasmic calcium levels, as result of intracellular dissolution of nHA over time, is likely the cause of cell death [13]. Nevertheless, Zhao et al. [14] found that shape has a more important role on nHA cytotoxicity than cell uptake, as needle-and plate-shaped nHA were more toxic to BEAS-2B cells than rod-shaped nHA, the latter for which higher particle-cell association and increased cellular uptake were observed [14]. Size has been shown to influence the cytotoxicity of sphere-shaped nHA to HepG2 cell [15], with higher toxicity of the 45 nm nHA (in comparison to 26, 78 and 175 nm nHA). The higher toxicity was associated with higher levels of apoptosis markers and higher concentration of nHA around the nucleus of the cells [15]. Among the in vivo studies available, rod-shaped nHA caused no effects to wistar rats exposed through food for 1 year (evaluated parameters were body weight, food consumption, clinical observations, survival, biochemical and hematological analysis, and histopathological evaluation) [16]. A study on Drosophila melanogaster showed no evidence of toxicity in terms of biochemical parameters (lipid peroxidation and nitro blue tetrazolium assay), behavioral changes in larvae and adults, nor chronic survival toxicity [17]; the nHA used was 20 nm with the presence of needle-like structures. On the other hand, rod-shaped nHA with 70-80 by 40-50 nm induced oxidative stress and cell damage to gut cell of D. melanogaster larvae, caused developmental delay in the late third instar larvae, decreased hatching, and behavior and phenotypical abnormalities [18]. The effects reported were attributed to interference of nHA with the calcium and phosphorus absorption pathway [18]. Few studies have analyzed in vitro and in vivo effects simultaneously. Needle-and rod-shaped nHA caused no cytotoxicity (in vitro) to catfish B-cells (3B11) and catfish T-cells (28s.3) up to 300 µg/mL; however, hatching inhibition was reported in zebrafish (in vivo) and, although all the larvae hatched at 120 h post fertilization, needle-shaped nHA caused spine malformation in 75% of the embryos [19]. In another study, nHA caused no cytotoxicity to murine fibroblast cells (L929), but induced an increase in lipid peroxidation levels in rat liver cells (in vitro) [20]. The same study reported no effects to guinea pigs, in vivo, when evaluating skin sensitization potential, blood parameters, oxidative stress in liver and brain and DNA damage [20]. It is noticeable that ecotoxicity information on nHA is missing, representing a major current gap for safe and sustainable materials.
The aim of the present study was to investigate the ecotoxicological effects of seven HA or HA-hybrid nanomaterials. These HA materials [part of the repository of test materials within the EU H2020-NMBP-2017 BIORIMA project (GA 760928)] covered various biomedical applications and SSbD alternative materials, i.e., (1) commercial hydroxyapatite was purchased from Sigma-Aldrich (Merck Life Science S.L.U., Portugal) (Sigma-HA)-used as a standard for comparison, (2) calcium hydroxyapatite (CaP-HA), (3) superparamagnetic iron doped hydroxyapatite (Fe-HA) [21]-a bioresorbable superparamagnetic nanoparticle proposed for theragnostic applications, (4) titanium-doped hydroxyapatite (Ti-HA), and (5) hybrid alginate/titanium-doped hydroxyapatite (Ti-HA-Alg) [22,23]-proposed an alternative to TiO 2 NPs to boost SPF in sunscreens, as they display high biocompatibility, bands of absorption/reflection of the UV radiation in the range suitable for application as sunscreen, and no photocatalytic effect thanks to the absence of titanium oxide crystals, and (6) collagen/hydroxyapatite hybrid scaffold (HA-Coll) [24], and (7) collagen/iron-doped hydroxyapatite hybrid scaffold (Fe-HA-Coll) [25]-proposed as bio-hybrid materials for regenerative medicine, as scaffolds are biocompatible and display good ability to stimulate and support cell adhesion and proliferation, holding promise, for instance, for therapies targeting bone tissue regeneration [26]. For the ecotoxicology study, the globally important soil invertebrate model species Enchytraeus crypticus (Oligochaeta) [27] was used. The effects were investigated based on the OECD standard 28 days enchytraeid reproduction test (ERT) [27] and on a longer-term exposure of 56 days, a standard extension. Further endpoints were included, i.e., animal size and intermediate sample times (7, 14, 21 days). In addition, for 2 of the nHA materials (6 and 7) it was not possible to homogeneously mix these with the soil, hence a short-term avoidance behavior test was used as screening tool to see if the organisms could detect the materials. In addition, the standard 2 days test was enhanced with additional sample times, i.e., 2, 3 and 4 days.

Test Organism
The test species Enchytraeus crypticus, (Westheide and Graefe, 1992) was used. The cultures were kept in agar, consisting of Bacti-Agar medium (Oxoid, Agar No. 1) and a sterilized mixture of four different salt solutions at final concentrations of 2 mM CaCl 2 ·2H 2 O, 1 mM MgSO 4 , 0.08 mM KCl, and 0.75 mM NaHCO 3 , under controlled conditions of temperature (19 ± 1 • C) and photoperiod (16:8 h light:dark). The animals were fed on ground autoclaved oats twice per week.
All of the synthesized materials were made available within the BIORIMA project materials repository. Briefly, CaP-HA, Fe-HA, Ti-HA were prepared by aqueous precipitation reaction. For CaP-HA, ortho-phosphoric acid solution (H 3  CaP-HA, Fe-HA and Ti-HA, after stirring the reactions were let without stirring or heat overnight. After that, the particles were resuspended by stirring for 15 min, the pellet was collected by centrifugation, and washed with water 3 times. The final powders were dried at 40 • C and sieved under 150 µm. The hybrid materials Ti-HA-Alg, HA-Coll and Fe-HA-Coll were produced by a nature-inspired biomineralization process, consisting of the heterogeneous nucleation of titanium/iron-doped hydroxyapatite NPs (Ti/Fe-HA) on biodegradable natural polymers: alginate (alginic acid sodium salt from brown algae; Sigma Aldrich, Saint Louis, MO, USA) or collagen (Opocrin SPA, Modena, Italy). In both cases, ortho-phosphoric acid solution added with the polymer (Alg or Coll), was dropped onto calcium hydroxide containing Ti or Fe 2+ / 3+ solution, at room temperature under gentle stirring; maturation at room temperature was kept for 1 h after the addition finished. Subsequently, the hybrid product was washed with water 3 times by sieving the gel and finally it was dried by freeze-drying process to obtain porous structures/scaffolds. For Ti-HA-Alg, to achieve microparticles, the dried product was sieved under 150 µm, then micronized in dry and cold conditions. All the materials were added to soil as dry powders as recommended for non-soluble materials [28]. Each replicate was spiked individually to ensure total raw amounts of the tested material. The soil was homogeneously mixed, deionized water was added until 50% of soil's maxWHC, and the soil was mixed again. The soil was left to equilibrate for 1 day prior the test start. The concentration ranges and tests performed for each material is presented on Table 1. HA-Coll and Fe-HA-Coll have a fibrous structure and when mixed onto soil a homogeneous distribution was not possible. The exposure to spiked media that is not homogeneously distributed will not allow for the assessment of effects as a consequence of exposure, and high variability is expected. Hence, avoidance tests were performed (instead of the reproduction tests) to assess enchytraeids behavior.

Test Procedures 2.4.1. Reproduction Tests
The tests followed the standard guideline for the Enchytraeid Reproduction Test (ERT, 28 days) [27], plus the OECD extension, as described in, e.g., Ribeiro et al. [29]. In short, the test was extended for 28 additional days (56 days in total) and extra monitoring sampling times at days 7, 14, 21, (28) and 56. Endpoints included survival for all sampling periods, reproduction at days 28 and 56, i.e., number of juveniles and population, respectively, and size at day 28. Four replicates per treatment were carried out, except at days 7, 14 and 21 (1 replicate). At test start, ten synchronized age organisms (young, 18-20 days old after cocoon laying) were introduced in each test vessel with moist soil (7, 14, 21, and 28-days exposure: Ø4 cm vessel, 20 g of soil, and 56 days exposure: Ø5.5 cm vessel, 40 g of soil) and food supply (22 ± 2 mg, autoclaved rolled oats). The test ran up to 56 days at 20 ± 1 • C and 16:8 h photoperiod. Food (11 ± 1 mg: until day 28, and 33 ± 3 mg: from 28 to 56 days) and water (based on weight loss) was replenished weekly. On sampling days 7, 14, 21, and 28, adults were carefully removed from the soil and counted (survival). The juveniles were counted at day 28 and 56 using a stereo microscope, to assess reproduction. After being fixed for 24 h with ethanol and Bengal rose (Merck Life Science S.L.U., Portugal) (1% in ethanol), soil samples were sieved through meshes with decreasing pore size (1.6, 0.5, and 0.3 mm) in order to separate the enchytraeids from most of the soil and facilitate counting. For the replicates that continued until day 56, adults were carefully removed from the soil at day 28. The adult organisms collected at day 28 were photographed after staining with Bengal rose, and size (area, mm2) was assessed using the software ImageJ (v.1.52a, Wayne Rasband, National Institutes of Health, USA) [30].

Avoidance Tests
Avoidance tests were performed following the earthworm avoidance test guideline [31] using E. crypticus with adaptations as described in Bicho et al. [32]. In short, plastic containers (2.5 × 6.5ø cm) with one removable plastic divider were used; each replicate contained 50 g of soil, 25 g in each side, i.e., the control and spiked soil. After this, the wall was gently removed and ten adult organisms with well-developed clitellum were placed on the contact line of the soils. The boxes were covered with a lid (containing small holes) and kept, for 48, 72 and 96 h, at 20 ± 1 • C and a photoperiod of 16:8 h (light-dark). Five replicates per treatment were used (at 48 h and 72 h monitoring, the replicates were checked but kept until the 96 h sample). At the end of the test period, the divider was again inserted in the separation line between the two soils and each side of the box was independently searched for worms.

Data Analysis
For all the tests, to assess differences between treatments and controls, one-way analysis of variance (ANOVA) was performed, followed by the post hoc Dunnett's method for multiple comparisons at a significance level of 0.05 (SigmaPlot v.14.0, Systat Software, Inc., San Jose, CA, USA). Avoidance was calculated as the percentage of worms that avoided the treated soil in the test container from the total number of worms in that container. The mean percentages of net responses (NR) were calculated as follows (Equation (1)): where C is the number of organisms observed in the control soil, T is the number of organisms observed in test soil and N is the total number of organisms per replicate. A positive (+) NR indicates avoidance, and a negative (−) NR indicates a non-response (or attraction) to the test substance

Materials Characterization
A summary of the main characteristics of the nHAs tested is provided in Table 2. For further details please see SI. In summary, among the "powder", Sigma-HA forms the least stable suspension (larger aggregates-larger hydrodynamic diameter and surface charge closer to zero); CaP-HA, Fe-HA and Ti-HA forms smaller aggregates and are negatively charged. The substitution of HA atoms by Fe or Ti changes the shape of the nHA from round to needle-or rod-like, for Fe-HA and Ti-HA, respectively, also affected by the temperature condition during synthesis and maturation step. The agglomerated state of Ti-HA-Alg hybrid particles and the fibrous structure of HA-Coll and Fe-HA-Coll are clearly shown in SEM images (Figures S1-S7).

Ecotoxicological Tests
The validity criteria were fulfilled as within the standard OECD test, i.e., in controls, adult mortality < 20% and the number of juveniles > 50 per replicate, with a coefficient of variation < 50%.
The results showed that in the standard exposure test (28 days) no dose-response negative impact pattern occurred in terms of survival or reproduction ( Figure 1A). For Ti-HA and TiHA-Alg, while there was an increase in reproduction, the size of the animals was reduced (Figure 2). For the standard extension exposure (56 days), although there was no dose-response pattern, there was a decrease in population numbers observed at the lowest tested concentration (100 mg NM/kg soil) for all materials, except Ti-HA-Alg ( Figure 1B,C).
For the avoidance test, the validity criteria were fulfilled, i.e., less than 20% mortality and homogeneous distribution (no avoidance) in controls. The results showed that enchytraeids did not avoid the soil containing HA-Coll or Fe-HA-Coll, instead there was a dose-dependent preference for the spiked soil ( Figure 3). The same pattern was observed at all exposure periods (48, 72, 96 h). Table 2. A summary of the tested nHA (commercial hydroxyapatite (Sigma-HA), calcium hydroxyapatite (CaP-HA), superparamagnetic iron doped hydroxyapatite (Fe-HA), and titanium-doped hydroxyapatite nanoparticles (Ti-HA), alginate/titanium-doped hydroxyapatite hybrid particles (Ti-HA-Alg), collagen/hydroxyapatite hybrid scaffold (HA-Coll) and collagen/iron-doped hydroxyapatite hybrid scaffold (Fe-HA-Coll)) main characteristics, including chemical composition, hydrodynamic size, surface charge and shape based on TEM/SEM pictures. DLS: dynamic light scattering, TEM: transmission electron microscopy, SEM: scanning electron microscopy.        negative impact pattern occurred in terms of survival or reproduction ( Figure 1A). For Ti-HA and TiHA-Alg, while there was an increase in reproduction, the size of the animals was reduced (Figure 2). For the standard extension exposure (56 days), although there was no dose-response pattern, there was a decrease in population numbers observed at the lowest tested concentration (100 mg NM/kg soil) for all materials, except Ti-HA-Alg ( Figure 1B,C).  For the avoidance test, the validity criteria were fulfilled, i.e., less than 20% mortality and homogeneous distribution (no avoidance) in controls. The results showed that enchytraeids did not avoid the soil containing HA-Coll or Fe-HA-Coll, instead there was a dose-dependent preference for the spiked soil ( Figure 3). The same pattern was observed at all exposure periods (48, 72, 96 h). Values represent average ± standard error (AV ± SE). *: p < 0.05 (Dunnett's method).

Discussion
No dose-response impact of the nHA-based materials was observed in terms of survival and reproduction in the standard test (28 days), but in the extended exposure (56 days) the lowest concentration caused an impact to E. crypticus. Higher effects at lower or intermediate concentrations of NMs have been previously reported to soil invertebrates exposed to, e.g., Ag [33,34], Ni [35], and Au [36]. The source of this increased toxicity is probably related to higher dispersibility/lower agglomeration of the particles at lower concentrations, which results in a higher availability. This was also reported for nHA in aqueous media, by Zhao et al. [19] whose results showed highest effect in terms of zebrafish hatching and embryo development at the lowest tested concentration (3 μg/mL), also associated with the lowest particles agglomeration.
Furthermore, it was of interest to note that there was a combination between a significant reduction in the size of animals exposed to Ti-HA and Ti-HA-Alg for 28 days, and    For the avoidance test, the validity criteria were fulfilled, i.e., less than 20% mortality and homogeneous distribution (no avoidance) in controls. The results showed that enchytraeids did not avoid the soil containing HA-Coll or Fe-HA-Coll, instead there was a dose-dependent preference for the spiked soil ( Figure 3). The same pattern was observed at all exposure periods (48, 72, 96 h). Values represent average ± standard error (AV ± SE). *: p < 0.05 (Dunnett's method).

Discussion
No dose-response impact of the nHA-based materials was observed in terms of survival and reproduction in the standard test (28 days), but in the extended exposure (56 days) the lowest concentration caused an impact to E. crypticus. Higher effects at lower or intermediate concentrations of NMs have been previously reported to soil invertebrates exposed to, e.g., Ag [33,34], Ni [35], and Au [36]. The source of this increased toxicity is probably related to higher dispersibility/lower agglomeration of the particles at lower concentrations, which results in a higher availability. This was also reported for nHA in aqueous media, by Zhao et al. [19] whose results showed highest effect in terms of zebrafish hatching and embryo development at the lowest tested concentration (3 μg/mL), also associated with the lowest particles agglomeration.
Furthermore, it was of interest to note that there was a combination between a significant reduction in the size of animals exposed to Ti-HA and Ti-HA-Alg for 28 days, and

Discussion
No dose-response impact of the nHA-based materials was observed in terms of survival and reproduction in the standard test (28 days), but in the extended exposure (56 days) the lowest concentration caused an impact to E. crypticus. Higher effects at lower or intermediate concentrations of NMs have been previously reported to soil invertebrates exposed to, e.g., Ag [33,34], Ni [35], and Au [36]. The source of this increased toxicity is probably related to higher dispersibility/lower agglomeration of the particles at lower concentrations, which results in a higher availability. This was also reported for nHA in aqueous media, by Zhao et al. [19] whose results showed highest effect in terms of zebrafish hatching and embryo development at the lowest tested concentration (3 µg/mL), also associated with the lowest particles agglomeration.
Furthermore, it was of interest to note that there was a combination between a significant reduction in the size of animals exposed to Ti-HA and Ti-HA-Alg for 28 days, and their increased number of juveniles. This is not an uncommon situation, where stressed animals respond by increasing one on trait compromising another, i.e., increase the reproduction compromising growth, hence prioritizing energy allocation to the next generation numbers. The growth inhibition can result from impairment in feeding activity/damage to digestive track [37], or could be related to developmental effects. In D. melanogaster, rod-shaped nHA exposure, caused developmental delay in the late third instar larvae, among other effects including phenotypical abnormalities in the hatched flies [18]. Considering that size was only reduced for Ti-doped NMs, Ti could be the source of the observed effects. TiO 2 NPs inhibit growth in zebrafish [38]. It is known that Ti can be released from Ti-based bone implants applied in mini pig maxillae [39], thus Ti release from the NMs and consequent toxicity cannot be ruled out. The extended exposure of 56 days showed that Ti-HA (100 mg/kg) caused a reduction in the population numbers, but this was not the case for Ti-HA-Alg.
Enchytraeids showed a clear attraction behavior or preference for HA-Coll and Fe-HA-Coll spiked soil, independently of the exposure period. Previous studies have shown that the attraction for spiked soil can be caused by neurotoxic effects resulting in the animals' incapacity to avoid the spiked soil [32,40]. However, it was observed that animals feed on Fe-HA-Coll; because of its orange color, Fe-HA-Coll was easily visualized inside the animals' gut ( Figure S8). HA-Coll must also be ingested, although being whitish it could not be easily identified. This could be somehow associated to the preference for the spiked soil, a palatable material. Neurotoxicity seems unlikely to be the reason considering the little to no effects observed in terms of survival and reproduction, for the majority of the nHA, and including the Fe-HA material.
Although the role of shape or size on nHA toxicity has been highlighted [14,15,19], our results do not allow for the confirmation of any association. For example, Ti-HA, which caused a decrease in animals' growth, has a rod-shape, whereas Sigma-HA and CaP-HA are spherical and Fe-HA look more needle-like ( Figure S3). Since the tested NMs have other aspects varying besides shape or size, any assumption would be more of a speculation.

Conclusions
We provided ecotoxicological data for seven nHA materials, which was a current gap in the literature. Considering the overall low ecotoxicity to E. crypticus, the nHA tested materials do not seem to cause high risk to soil inhabiting invertebrates. Special attention should be given to lower concentrations of NMs, for nHA materials in the 100-320 mg/kg range, where increased impacts were observed. From all nHA SSbD material options, since Ti-HA and Ti-HA-Alg caused growth inhibition, likely due to the presence of Ti-these nHA seemed to be the least safe.