Use of Caenorhabditis Elegans and Zebrash Embryo as an Alternative Model for Biocompatibility Assessment of Natural Latex Biocomposite

The number of animals killed used for research purposes has increased signicantly in recent years. For this reason, several alternatives of animal models are widely used. In this study, two phases of calcium phosphates synthesized by the sol-gel method were added by precipitation to the surface of the natural rubber latex membranes for bone applications. Membranes cytotoxicity and toxicity was evaluated by in vitro and in vivo methods. Hemolysis levels were less than 3%, which indicates that there was no signicant hemolytic activity. It was demonstrated that latex presents enzymes that can dissociate osteoblasts MC3T3, so it was necessary to eliminate these enzymes to obtain a good adhesion and cellular proliferation without morphological alterations. The in vivo toxicity test with Caenorhabditis elegans worms did not show evidence of toxicity with a sample of hydroxyapatite incubated for 5 days, exhibiting the lowest survival rate of the worms. In the Zebrash embryotoxicity assay, 60% of embryos survived and there was no evidence of embryos with malformation or developmental delay. In summary, alternative animal models, which are not commonly used to assess biocompatibility, have provided reliable in vivo results that allow suggesting the use of these membranes in the bone biomedical.


Introduction
Bones and teeth are natural composites formed by collagen molecules linked in linear chains arranged in ber shapes [1]. Among these molecules, there are small regularly spaced interstitial compartments where nanocrystals of a known inorganic solid, such as calcium phosphate, are present, accounting for 65% of the total bone mass [2]. Calcium phosphates (CaP´s) have been widely used in bone regeneration applications because they show osteoconductive and, in some cases, osteoinductive features. The release of calcium and phosphorus ions regulates the activation of osteoblasts and osteoclasts, facilitating bone regeneration [3]. However, like any ceramic material, these compounds have high brittleness and low elasticity, which can impair their mechanical performance after implantation [4].
In this sense, incorporating CaP´s into a polymeric matrix can be a possible solution. Combining CaP´s with polymers has become a popular method for fabricating composite scaffolds, because CaP´s play an active role in increasing cell adhesion and are able to induce differentiation of osteoprogenitor cells [5].
As the number of available polymers is almost unlimited, there is a wide potential for such combinations.
Natural rubber latex (NRL) is extracted from the Hevea brasiliensis tree, belonging to the Euphorbiaceae family, native to the Amazon basin in Brazil. This material consists of 40 to 45% by weight of polymeric molecules that form the rubber poly(cis-1,4-isoprene), 4 to 5% by weight of other organic constituents, such as proteins, lipids and carbohydrates, in addition to 50% of water [6]. The NRL membranes have been widely used in the clinical area, due to the excellent angiogenic properties presents in NRL [7]. Mrué [8] used the NRL membranes for the rst time as prosthesis for a segment of the cervical esophagus of dogs, which demonstrated inducing property of tissue regeneration (neo-esophagus).
Balabanian et al. [9] implanted latex granules in the alveolar bone cavities after dental extraction in rats and observed that there was an acceleration of the new bone formation and a decrease in the presence of connective tissue around the implant. These results were further elucidated by Ereno et al. [10] where they used the NRL membranes for bone repair in rabbit calvaria. In this experiment, the NRL membrane was applied on the fracture, which prevented the migration of epithelial and connective tissue and facilitated the action of regenerative cells.
The increase in the number of researches in the biotechnology area has stimulated the search for alternative methodologies for assessing toxicity. Rat and mice are the conventional animal models used for toxicity and pharmacological tests. However, alternative methods seek to reduce the number of animals used in the experiment, as well as the cost of the experiments, since the animals used in the research need to be conditioned, fed and kept in the best possible health and hygiene conditions, otherwise they will not may be used for scienti c purposes [11].
Caenorhabditis elegans is a small nematode that can be used as an animal model in biotechnological research. Unlike toxicity tests using cell cultures, C. elegans toxicity assays provide data for a compete organism, including systems such as digestive, reproductive, endocrine and neuromuscular, besides presenting 60-80% genome homology with humans [12]. These animals are small (1 mm long), proliferative, with a short lifecycle and have few maintenance requirements, which facilitates the study in a multicellular organism [13].
Another alternative model for the in vivo toxicity tests is the embryo test using the Danio rerio sh, better known as Zebra sh. Zebra sh genome shares approximately 70% homology with humans, so the bases of embryonic development are fundamentally similar between Zebra sh and mammalian embryos [14]. Due to the transparent nature of the eggs and embryos, Zebra sh allows the visualization of morphological and structural anomalies throughout the animal's body [15]. Despite their apparent simplicity, these models have aided biomedical research, particularly in the functional characterization of novel drug targets that has been identi ed using genomics technologies. However, their use in the eld of biomaterials research has been limited.
In this work, two types of CaP´s were synthesized via the sol-gel method (calcium pyrophosphate and hydroxyapatite) and incorporated separately into the NRL membranes. NRL used in this work, are composed of several proteins that can cause allergic reactions, and due to that clari cation step was performed to reduce them. Membrane toxicity was assessed in vitro by hemolytic and cell culture (MC3T3 osteoblast) tests and in vivo using the nematode C. elegans and Zebra sh embryos.  LGCBio, Brazil) was used instead of the PBS solution. After the incubation period, the PBS solutions (or α-MEM) were collected for later use in in vitro and in vivo toxicity tests. These solutions were called eluates and identi ed as: NRL-e24, NRL-e120, CPP-e24, CPP-e120, HA-e24 and HA-e120.

Hemolysis assay
The protocol was carried out according to Onuma et al. [18] where sheep blood cells (Newprov®, Brazil) were used in this assay. 50 µL of the NRL-e24, NRL-e120, CPP-e24, CPP-e120, HA-e24 and HA-e120 were incubated together with 50 µL of a 5% (w/v) erythrocyte solution (0.9% (w/v) saline solution) at 37°C for 1 h. After this period, the eluates were centrifuged at 3000 rpm for 3 min and 50 µL of the supernatant were pipetted into 96-well microplates. The absorbance measurements at 540 nm were determined using a microplate reader (PowerWave Epoch2, BioTek Intruments®, USA). The 100% hemolysis value was de ned using the positive control (50 µL of PBS solution with 100 µL of 1% (v/v) Triton X-100 (Sigma-Aldrich®, USA), while the 0% hemolysis value was obtained using the negative control (10 µL of PBS solution) [19]. The hemolysis rate was calculated using following Eq. (1): 1 ABS sample is absorbance of the samples, ABS negative control is absorbance of negative control, and ABS positive control is absorbance of positive control.

Adhesion assay on the surface of membranes
The NRL-m was incubated with culture medium for 48 h to extract and neutralize the enzymes that make the adhesion assay impossible and after that time, the medium was discarded [20].
medium enriched with the CPP and HA (1 mg/mL) was added to the pure NRL-m. After 48 h, the medium was carefully removed from the corner of the wells so as not to extract the phosphates that precipitated on the membrane surface. The membranes were sterilized by ultraviolet light for 20 min. The adhesion assay was realized in 24-well plates, seeding 2x10 4 cells/cm of MC3T3 osteoblast directly on the membranes surface and cultivating with the α-MEM medium for 48 h. The membranes were washed with PBS solution (pH = 7.2), xed in 2.5% (v/v in PBS) glutaraldehyde (Sigma-Aldrich ® , USA) for 1 h at room temperature. Subsequently, the samples were dehydrated in increasing series of ethanol (50%, 60%, 70%, 80%, 90% and 100%) and dried at room temperature. The analyzes were evaluated by FEG-SEM (model 7500F, JEOL ® , Japan) at 20 kV.
In vivo toxicity assays Toxicity assay in Caenorhabditis elegans C. elegans worms of wild-type strain N2 were synchronized by treatment with sodium hypochlorite (Sigma-Aldrich ® , USA) and the 4th larval stage (L4) was used for the experiment. The worms were washed with NaCl 50 mM (Sigma-Aldrich ® , USA) and about 20-25 worms were added to the wells of 96well plates containing 54 µL NaCl 50 mM, 36 µL Brain Heart Infusion medium, 10 µg/mL cholesterol, 90 µg/mL kanamycin and 200 µg/mL ampicillin. 100 µL of samples of NRL-e24, NRL-e120, CPP-e24, CPP-e120, HA-e24 and HA-e120 were added to the wells and the plates were kept at 25 ºC for 24 and 48 h. Toxicological results were analyzed by worm shape and mobility (rod-shaped worms were considered dead and sinusoidal worms were considered alive) and by nematode reproduction. The visualization was carried out using a Zeiss ® Discovery V12 (Germany) microscope attached to a camera. The statistical analysis of the results was performed using the GraphPad Prism 5 ® software (San Diego, CA, USA).
Zebra sh embryos used for chemical exposure were obtained from spawning adults in appropriate tanks overnight with a male/female ratio of 2:1. Embryos were collected within 1 h post fertilization (hpf) and were transferred to 96-well plates. 100 µL of NRL-e120, CPP-e120 and HA-e120 samples were tested in the static mode and embryonic medium (NaCl; KCl;

Results And Discussion
Characterization of calcium phosphate phases In this study, the choice of the sol-gel method for the CaP´s synthesis was based on the simplicity of the method and the ability to obtain particles with nanometric structure, aiming at the in vivo biomineralization process. Using the same chemical precursors and different temperatures, it is possible to obtain two different CaP´s phases, namely: CPP and HA, with properties needed as biomaterial, such as osteoinductivity, bioactivity and nontoxic. Figure 1A shows bond [21]. The XRD patterns (Fig. 1B) con rmed the existence of the CaP´s phases obtained in each heat treatment performed in the sol-gel precipitate [17]. Both samples revealed broad peaks at 2θ = 20-40°, characteristic of nanocrystalline phases. The intense peaks at 2θ = 25°, 32° and 35° indicated that product obtained at 400°C was Ca 2 P 2 O 7 base on CSM card #96-100-1557. The thermal treatment at 600°C con rmed the formation the Ca 5 (PO 4 ) 3 OH, identi ed by the peaks 26°, 31°and 33° (card #96-900-1234). In addition, morphological characterization was carried out (Fig. 1C). The CPP micrographs showed liform particles grouped in the form of agglomerates with a size range of 170-650 nm, while the particles of HA presented sizes between 110-620 nm and hexagonal shape.

Calcium phosphate incorporation into the NRL membranes
The NRL membranes were prepared according to the Sect. 2.3. The proteins responsible for the allergic reactions of NRL have a molecular weight of about 14 kDa, and due to that, they can be easily separated by clari cation unit (centrifugation), [22] where reduction of 27% of the proteins were achieved.
The incorporation process of CaP´s phases was ful lled by precipitation, unlike the casting method commonly used in the literature [23][24][25]. Figure 2 (A-C) shows SEM micrographs of the surfaces of the NRL membranes without and with the inclusion of the phases containing CPP and HA. The pure NRL membrane (NRL-m) exhibited a smooth surface, with evidence of some irregularities ( Fig. 2A), which can be explained by the fact that, in solution, some of the NRL proteins are released and, after drying, they are precipitated on the surface [26]. With a magni cation of 2000x, it was possible to observe that the CPP particles are homogeneously dispersed in the NRL matrix (CPP-m) (Fig. 2B). In the micrograph of the membrane incorporated with HA (HA-m), a heterogeneous layer of HA was denoted on the material surface (Fig. 2C). The lower uniformity of the HA particles in the HA-m membrane in relation to the CPP-m membranes may be related to the different solubility of each CaP´s phase, because the higher the Ca/P ratio, the lower the solubility in water [27]. HA has the highest Ca/P ratio with a value of 1.67, while CPP has a ratio of 1.0. Thus, this large difference in solubility makes CPP easier to be adsorbed by the polymeric membrane. To con rm the presence of CPP and HA particles in the membranes, we used the EDX technique coupled to the SEM. In Fig. 2 (D-E), the EDX spectra are demonstrated and the presence of both chemical elements Ca and P, constituents of CPP and HA, can be observed on the membranes surface.
The obtaining of CaP´s-NRL composites with different CaP´s has been the object of study by different researchers [28,29]. However, there are few studies that have precisely explained the molecular interactions between the non-rubber constituents of NRL and inorganic particles. However, it is known that, on the CaP´s particles surface, there are Ca 2+ ions that are attracted by the negative charges found in the NRL proteins and lipids.
In vitro toxicity assays Evaluation of hemolytic activity Hemolytic activity assay was accomplished to assess whether the components released by the samples NRL-e24, NRL-e120, CPP-e24, CPP-e120, HA-e24 and HA-e120 are able to rupture the red blood cell membrane, causing hemoglobin release [16,30]. Figure 3 shows the positive control with red color, indicating that hemolysis occurred due to the release of hemoglobin, and the transparent samples (negative control and other samples), where there was no hemolysis. According to Fischer et al. [31] rates below 10% of hemolysis suggest that the samples are considered non-toxic. The released hemoglobin percentages were calculated and it was proved that the eluates at 24 and 120 h exhibited hemolysis levels less than 3%, con rming that there was no damage to the red blood cell membrane and the type of CaP´s did not in uence hemolytic activity. Our results are complemented by those obtained by De Barros et al. [32] who tested the hemolytic activity of the NRL membranes with oxytocin, and values below 5% were achieved. Floriano et al. [33] also veri ed that the NRL membranes coated with CaP´s and with ketoprofen respectively, did not present hemolytic effect.

Cytotoxicity assay in osteoblasts
The in vitro test was performed with MC3T3 osteoblasts. The samples NRL-e24, NRL-e120, CPP-e24, CPP-e120, HA-e24 and HA-e120 were placed in indirect contact with the cells for 24 and 48 h. After 24 h of plating, it was found that the cells were adhered and the culture medium was replaced. The MTT assay was carried out at 48 h and it was observed that all samples, except the positive control (100%), demonstrated percentage values of cell viability less than 50%. This result indicates that the eluates tested showed cytotoxic behavior, which is in contradiction with the countless studies in the literature that demonstrate the excellent biocompatibility and bioactivity in vivo of NRL [9,10,34,35].
However, in this study, biocompatibility assays were evaluated in the serum (eluates) not in the NRL membranes. It is known that NRL contains proteolytic enzymes, such as serine and cysteine proteases, [16,[36][37][38] which can in uence the cell viability assay. To prove this hypothesis, a cell culture test was accomplished using only NRL serum previously centrifuged (at 19000 rpm), and after 5 min of incubation, it was microscopically veri ed that there was no evidence of cell adhesion and the MC3T3 osteoblast cell body was dissociated. This result indicated that the proteolytic enzymes present in NRL may be able to dissociate the MC3T3 cells, causing the low values of cell viability obtained in the MTT assays. Borges et al. [20] obtained similar results when studying the eluate of the incubation of the NRL membrane without dilution, exhibiting a cell viability of approximately 48%.
From these results, a new cell adhesion test was realized. At this time, to eliminate the enzymes that prevent adhesion, the membranes were prepared according to Sect. 2.7. The micrographs of the membranes surface after 48 h of culture (Fig. 4(A-C)) revealed that the cell line established points of adhesion and the cells did not undergo morphological changes. The existence of lamellipodia and philopodia of osteoblasts (red arrows), as well as intercellular connections (blue arrows), indicated that there was a good adhesion of osteoblasts to the membranes surface [20]. In tissue engineering, in addition to cell adhesion, cell proliferation and differentiation are necessary to guarantee greater bioactivity of the material [39]. In this study, it was possible to observe the cell division in the NRL-m, CPPm and HA-m membranes ( Fig. 4 (D-F)) through the FEG-SEM analysis, suggesting the cell proliferation qualitatively.
According to Deligianni et al. [40] osteoblasts present better proliferation, adhesion and mineralization on rough and sandy surfaces. Thus, the CaP´s phases are excellent candidates for polymer coating, because, in addition to these characteristics, CaP´s are formed by the main bone constituents, bonding chemically and directly to the bone, resulting in a better bone repair. Figure 4 (D-F) exhibits a close connection between clusters of CPP and HA and osteoblasts on the NRL membranes surface (green arrows). These results suggest good cell viability on the membranes. Finally, the in vitro biocompatibility of the prepared membranes did not show appreciable differences between them.

In vivo toxicity assays
Toxicity assay in Caenorhabditis elegans The toxicity assay using C. elegans followed the methodology described in Sect. 2.4.1 where all membranes were incubated separately in PBS solution for 24 and 120 h. The eluates resulting from these incubations were placed in contact with the worms and analyzed after 24 and 48 h. Figure 5 (A-C) reveals the percentages of survival, where in the rst 24 h of the experiment, all samples did not signi cantly reduce the survival of C. elegans worms. After 48 h of assay, the survival rates of the animals exposed to the samples were analyzed again and the eluates NRL-e24, NRL-e120, CPP-e24, CPP-e120 and HA-e24 did not show mortality. However, the worms exposed to the HA-e120 sample presented toxic effects, with only 42% survival.
In this study, the toxicity also was veri ed by the worm development and by monitoring the body shape. Live worms (sinusoidal shape) were observed for all samples and a predominant amount of dead worms were identi ed by their rod-shape in the HA-e120 sample (Fig. 5 (D-F)). The toxicity of this sample may be related to the high presence of Ca 2+ ions in the HA-e120 eluate (5 days). As previously mentioned, the difference in the Ca/P ratio between the two CaP´s resulted in a lower solubility for HA, which caused the HA particles to not be fully adsorbed on the NRL membranes surface (Fig. 2C), being easily released in the HA-e-120 eluate. Ca 2+ plays an essential role in physiological processes and in cell proliferation or differentiation. Alvarez et al. [41] summarized evidence from various experiments on C. elegans worms that helped to understand the role of Ca 2+ in living systems. However, their results showed that the system is quite complex and there is only a ne line that separates the levels of calcium ows that can be considered normal from those that can lead to pathology. Therefore, an excess of Ca 2+ ions in the HA-e120 eluates (after 5 days of incubation) may have induced the death of a greater number of worms. Nematode reproduction was also evaluated (Fig. 5E) and in all samples, except HA-e120, the presence of the 1st larval stage (L1) from the 4th larval stage (L4) reproduction indicates the non-toxic character of these samples.

Embryotoxicity assay in Danio rerio (Zebra sh)
A second in vivo model was used to investigate the biocompatibility of the eluates incubated for 5 days (NRL-e120, CPP-e120 and HA-e120). Recently, Zebra sh proved to be an ideal model for studying biocompatibility and bone disease [42]. Zebra sh embryos were analyzed by determining different physiological parameters, such as percentage of viability and phenotypic morphological alteration. Figure 6A shows the survival rate of the three samples and the control in relation to the contact time with the embryos. The samples were obtained employing the same methodology used in the C. elegans assay and the eggs were exposed for 24, 48, 72 and 96 hpf. Quantitatively, all samples, including the control, exhibited dead embryos from 24 hpf. The survival rate in the solvent control was approximately 76%, which was reduced to nearly 60% when the embryos were incubated in the studied eluates. It was noted that the reduction was not dependent on the exposure time or the type of eluate. These results suggest low embryotoxicity, as the values obtained by the eluates were near to those of the control sample (containing only embryonic medium, without embryos), which is widely used as control in this type of assay [43].
In view of these results and knowing that the embryonic stages are extremely sensitive for the determination of toxicity, [15] the number of live, dead, delay in development and deformed embryos was counted. This study showed only the existence of live or dead animals (more than 50% of Zebra sh embryos were found alive) and none of the tested samples induced deformation or delay in embryonic development (Fig. 6B). This result was supported by the analysis of the images obtained from the development of live embryos. Through the transparency of the embryos, it was possible to observe different stages of normal development, according to the phases of gastrula, pharynx, early larvae and larva ( Fig. 7(A-D)), which were found in all samples [44]. The certi cation that the embryos were alive was carried out through microscopic observation of the sequence of development and movement of the embryos. The existence of worms with curvature of the tail and pericardial edema were considered dead.
These results are in line with those reported by Makkar et al. [45] who used the Zebra sh model to study the biocompatibility of two dental bioceramics, one of which is the mineral trioxide aggregate (MTA) composed mainly of inorganic elements containing calcium. The eluates of MTA contain silicates, aluminates and calcium oxides, which facilitated the increase in abnormal metabolic activity in embryonic cells, causing a percentage of embryonic deaths from oxidative stress and cellular apoptosis.

Conclusion
In this work, the sol-gel method was used to obtain two different CaP´s that were incorporated into the NRL membranes. The cytotoxicity of the samples was tested by conventional in vitro methods, in addition to two alternative in vivo models that have so far been little explored to evaluate the biocompatibility of biomaterials. The membranes were obtained by the precipitation method of the CaP´s phases in the NRL matrix and, by means of FEG-SEM micrographs, it was observed that the CaP´s phases were incorporated into the membrane surface. The hemolytic activity assay did not con rm cell damage, with hemolysis level less than 3%. The cytotoxicity assay by MTT in MC3T3 osteoblasts was not e cient, because enzymes contained in NRL dissociated the cells, making the test impossible. Therefore, it was necessary to wash the pure NRL membranes with complete α-MEM medium and then incorporate the CaṔ s phases. After these procedures, it was observed that the cells adhered to the surface of all membranes with intimate contact through lamellipodia and philopodia. The in vivo biocompatibility study using the C. elegans nematode did not demonstrate signi cant toxicity in the samples incubated for 24 and 48 h. During the larvae counting, the worm reproduction capacity was observed, due to the presence of L1 larvae from the reproduction of L4 larvae. The embryotoxicity of all samples was tested using embryos of the D. rerio sh (Zebra sh) and it was found that all samples contained live and dead embryos, without malformation or developmental delay, suggesting low embryotoxicity. The alternative animal models used in this study allowed assessing the in vivo biocompatibility of the studied membranes, indicating their potential to be used in biomedical applications.     (A) Survivability of embryonic Zebra sh exposed to control sample, NRL-e120, CPP-e120 and HA-e120 and (B) average the embryos number of live, dead, delayed in development and deformed counted in each sample after 96 hpf.