Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone

Non-collagenous proteins (NCPs) are specialized biomacromolecules within the extracellular matrix (ECM) that regulate the mineraliza- tion of calcified tissues, such as bone and dentin. Numerous in vitro studies have demonstrated that natural polyanionic NCPs and their analogues can mediate intrafibrillar mineralization, characterized by the infiltration of apatite minerals into collagen fibr ils. However, these studies primarily utilize self-assembled collagen fibrils or demineralized mature tissues, leaving it unclear whether pristine embry- onic bone ECM at a developmental stage permissive to mineral deposition can regulate intrafibrillar mineralization independen tly or requires polyanionic NCP substitutes to promote the process artificially. To address this, we employed an ex vivo model of endochondral ossification using metatarsals isolated from 15-day-old embryonic mice (E15). In addition to a supersaturated calcium (Ca) and inorganic phosphate (Pi) medium, we introduced fetuin -A, a native polyanionic NCP or poly-DL-aspartic acid (pAsp), commonly used as an NCP substitute. The incorporation of either additive was essential for the effective mineralization of embryonic metatarsals. Both fetuin-A and pAsp played a direct role in facilitating the infiltration of Ca-Pi precursors into the avascular cartilaginous matrix. Raman spectroscopy and electron microscopy confirmed the formation of hydroxyapatite (HAp) exhibiting diverse levels of crystallinity, with fetu in-A sup- plementation resulting in the greatest HAp ac cumulation within the rudiments. HAp was localized in the perichondrium, a region con- ducive to initial mineralization and enriched with a fibrillar network of collagen types I and II. Three -dimensional reconstructions im- plementing Dijkstra’s algorithm revealed the association between HAp and collagen fibrils either organized in an intrafibrillar, extrafi- brillar, or combined arrangement.

Bone, Embryonic, Organ Explant, Mineralization, Non-Collagenous Proteins

Bone is a calcified connective tissue that stores minerals, cre- ates the skeletal system, and actively contributes to metabolic functions.1 In mammals, all long bones are formed via endo- chondral ossification, which is a replacement of hyaline carti- lage anlagen with bone tissue. 2 Collagen fibrils —the primary structural units of cartilage and bone —are internally strength- ened by minerals at the nano-structural level via a process called mineralization to improve their structural integrity and re- sistance to internal and external load s.3 Bone minerals consist of platelet-shaped hydroxyapatite (HAp) crystals with a length, width, and thickness of 12–100 nm, 10–40 nm, and 0.61–5 nm, respectively.4–7 However, biogenic HAp typically exhibits a non-stoichiometric phase, characterized by a high carbonate (CO32–) ion content, ranging from approximately 5.8% to 7.4% by weight, along with minor substitutions of other anions and cations, such as Na +, Mg 2+, K+, Cl−, and F −.8,9 This imbalance makes biological HAp highly reactive and effective at exchang- ing calcium (Ca), inorganic phosphate (Pi), and other ions, as bones are intricately connected with blood vessels and con- stantly exposed to blood flow.10 The complex “intrafibrillar mineralization” mechanism by which nano-sized HAp crystals nucleate and grow within colla- gen fibrils during bone tissue formation has been studied for decades.3,11–13 Recent in vitro experiments have successfully achieved intrafibrillar mineralization by treating reconstituted collagen fibrils with supersaturated Ca and Pi solution with poly-DL-aspartic acid (pAsp) as a negatively charged “polyan- ionic” non -collagenous proteins (NCPs) substitute.14,15 This polyanionic macromolecule was demonstrated in vitro to inter- act with Ca and Pi ions in solution and prevent spontaneous nu- cleation of HAp crystals by stabilizing amorphous calcium phosphate (ACP) precursor phase via a polymer-induced liquid precursor (PILP) process.14,16,17 This negatively charged precur- sor interacts with a positively charged region at the C -terminus end of the gap zones of collagen fibrils, which facilitates the infiltration of ACP within the fibril and subsequent ACP to HAp transformation. 18 The PILP process was further demon- strated to be used as a biomimetic system to study different pa- rameters controlling the mineralization process, such as the role of the polymer molecular weight 19 and extended beyond single fibril to promote the mineralization of dense collagen scaffolds and the remineralization of demineralized bone. 20,21 .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 2 Advanced remineralization studies using decalcified and chem- ically fixed mature dentin tissue have shown that the native ex- tracellular matrix (ECM), which includes collagen fibrils and other associated macromolecules contains the information needed to guide mineralization when exposed to metastable Ca- Pi solution, resulting in the formation of ACP within the tis- sue.22 These experiments highlight the need to determine whether the ECM of developing bone possesses inherent func- tional properties that regulate mineralization or whether the presence of polyanionic NCPs is necessary to trigger mineral deposition within collagen fibrils. Fetuin-A is an embryonic NCP with polyanionic characteristics, classified as a phosphorylated glycoprotein (42 –68 kDa) and a member of the cystatin superfamily.23–25 It consists of two struc- turally linked cystatin D1 and D2 domains, along with a third domain rich in proline and glycine.26 At the onset of endochon- dral ossification, hypertrophic chondrocytes and cartilage ECM exhibit high expression of fetuin-A, suggesting its potential in- volvement in tissue mineralization. 25,27 Additionally, this poly- anionic NCP normally circulates in the blood and extracellular fluid with a half-life of several days, to regulate the Ca -Pi pre- cursors and prevent spontaneous vascular calcification.28,29 Previous comparative in vitro studies demonstrated that fetuin- A promotes intrafibrillar formation of HAp crystals in both the reconstituted collagen fibrils and decalcified bone tissue via mineralization by an inhibitor size exclusion (ISE) mecha- nism.12 During the ISE process, fetuin -A binds selectively to Ca-Pi precursors30 and inhibits the apatite precipitation outside the collagen fibrils. 12 This allows the precursors to exclusively infiltrate the fibril and subsequently nucleate to form HAp crys- tals within the intrafibrillar compartments.12 The ISE route con- trasts with the PILP process in which a liquid ACP precursor phase forms extrafibrillarly and enters the collagen fibrils via capillary forces. 14 Despite the differences, both mechanisms highlight the intricate interplay between collagen fibrils, poly- anionic NCP substitutes, and Ca -Pi minerals to control and fa- cilitate mineralization.12,14 Nevertheless, all those studies were conducted either using self- assembly collagen fibrils 13–15 or decalcified mature tissue. 12,22 Hence, it remains unclear whether the pristine and unmineral- ized ECM of embryonic bone contains sufficient capability to regulate the mineralization directly or whether native polyan- ionic NCPs and similarly charged macromolecule analogues are necessary to promote artificial mineralization within embryonic bone. Accordingly, our present study aimed to investigate these issues using an ex vivo model of endochondral ossification in embryonic mouse metatarsals, an organ explant culture pio- neered by Burger and colleagues 31 with a novel upgrade on the incorporation of natural polyanionic NCP or its polymer ana- logue. We hypothesized that the polyanionic -type macromole- cules: fetuin-A and pAsp are crucial to initiate mineral precur- sors infiltration and promote artificial minera lization within embryonic bone tissue. It is important to understand the role of these negatively charged macromolecules and identify the best promoter of collagen mineralization during embryonic bone formation to aid the development of novel bio -inspired materi- als that are greatly beneficial in promoting fracture repair and treating bone-related diseases.

Development of Embryonic Bone Mineralization Model The 15th gestational days (E15) metatarsals were isolated as pre- viously described. 32 At this developmental stage, no matrix mineralization was observed at mid -diaphysis ( Figure 1 A). Moreover, the rudiments are still avascular and translucent, with proliferating/hypertrophic chondrocytes, collagen type -II fibrils, and aggrecan-type proteoglycans as their primary com- position.33,34 While metatarsals at E14 consist almost entirely of small-sized chondrocytes, the widening of lacunae or the begin- ning of cartilage hypertrophy is seen in the mid -diaphysis of E15, which is the preliminary sign that the tissue is ready to accept mineral deposition.31 The cartilage matrix is pristine and unmineralized, and the bone collar is still absent compared to later development stages.31,32 Small amounts of soft tissue, such as tendons and muscles remain attached to the outside of the rudiments, which are very difficult to remove during dissection. However, these soft tissues remove themselves after a few days in culture. From this point onward, artificial mineralization of E15 meta- tarsals was conducted using four different incubation mediums at two observation times: 7 and 9 days -of-culture. The culture mediums were classified as cell culture medium in the absence of CaPi minerals (Control), medium with CaPi minerals only (CaPi), and medium with CaPi minerals plus polyanionic addi- tives (CaPi + pAsp and CaPi + fetuin-A). Physical Appearance Change and Mineralization Features After 7 and 9 days in culture, the E15 metatarsals treated with CaPi + pAsp and CaPi + fetuin-A became stiff and appeared white under the microscope compared to the Control and CaPi samples ( Figure 1 B). Since all samples were washed thor- oughly with water before observation, this physical transfor- mation may indicate mineral deposition within the metatarsals. It should be noted that the metatarsals cultured in the CaPi me- dium (i.e., supersaturated Ca and Pi solution with respect to HAp stoichiometry 35,36) did not appear white under dark field (DF) optical imaging, which reflect the absence of mineral within the tissue. Comparatively, there were fewer solid precip- itates observed in the bottom of the culture well of both samples supplemented with pAsp and f etuin-A than in the CaPi -only sample. Altogether, this white color transformation, an increase in rudiment stiffness, and less mineral sedimentation in the cul- ture well are the first indication of a mineralization within E15 metatarsals treated with NCP substitutes. Organic and Inorganic Phases Identification To investigate both the organic and inorganic phases of E15 metatarsals, Raman observation on a selected region of the mid- diaphysis (Figure 1B) was conducted after 7 and 9 days-of-cul- ture. The resulting 100 spectra from each sample mapping area were averaged and normalized with respect to amide III band (~1250 cm –1) in agreement with previous studies. 37 The com- bined Raman spectra in the range of 250–3000 cm−1 region was displayed for both culture days ( Figure 1 C). The vibrational band assignments were identified and described (Supporting In- formation: Table S1). The peaks corresponding to the collagen, namely amides, along with the ring of proline, hydroxyproline, and phenylalanine were present in all samples both at 7 and 9 treatment days (Fig- ure 1C and Supporting Information: Table S1). Amide I and III bands are indicative of the collagen phase signatures and .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 3 were observed across all treatment conditions. In contrast, am- ide II band, which is associated with other proteins and lipids was exclusively detected in the Control sample on both incuba- tion days. The presence of glycosaminoglycans (GAGs) could also be identified by the peaks corresponding to the pyranose ring at 1045 cm−1 and the symmetry O–SO3− bond at 1062 cm−1. Importantly, the peaks correspond to the v1, v2, v3, and v4 modes of phosphate (PO 43−) were present only in the samples incu- bated in CaPi + fetuin-A with little change from 7 to 9 days-of- culture. The appearance of these peaks indicates the presence of HAp crystals within the metatarsals. The spectrum of samples incubated in CaPi + pAsp after 7 days only displayed a low - intensity peak at 961 cm −1, corresponding to the v1 mode of PO43−, whose intensity increased slightly after 9 days. In addi- tion, the v2 and v4 peaks were absent at 7 days and at low inten- sity at 9 days of incubation in CaPi + pAsp, indicating that min- eralization was slower than with CaPi + fetuin-A, with less HAp formed. The absence of the v1 and v4 modes of PO43− in the Con- trol and CaPi samples showed that without additives no miner- alization of the embryonic bone tissue takes place. Taken together, these results imply that fetuin -A and pAsp can both promote artificial mineralization of E15 metatarsals, with fetuin-A being more effective. It should be highlighted that Ra- man spectroscopy is a surface technique, with a beam penetration depth of approximately up to 12 µm. Therefore, electron microscopy (EM) examination is necessary to evaluate mineral localization deep inside the metatarsals (discussed be- low). Coils Arrangement and Intrafibrillar Mineralization A more ordered coils structural arrangement of collagen fibril is correlated to intrafibrillar mineralization, or it can also be in- terpreted as fibril preparation to receive minerals deposit.38 The quantification of coil arrangement level (CAL) values ( Equa- tion 1) on Raman spectra resulted in the highest overall value on CaPi + fetuin -A treatment on both 7 and 9 days of c ulture. The CAL mean values (Supporting Information: Table S2 ) were increased as follows: Control (0.511 ± 0.08) < CaPi + pAsp (0.667 ± 0.07) < CaPi (0.669 ± 0.13) < CaPi + fetuin -A (1.078 ± 0.06) for day 7, and Control (0.534 ± 0.09) < CaPi (0.641 ± 0.13) < CaPi + pAsp (0.705 ± 0.13) < CaPi + fetuin-A (1.411 ± 0.16) for day 9. With prolonged incubation time, the CAL variability of the additive -supplemented samples in- creased, signifying that the fibrils became more ordered with time in culture as a result of mineral deposition within the fi- brillar compartment. Statistical analysis revealed that samples with CaPi only and CaPi plus additives have a significant increase in CAL ( p < Figure 1. Optical images of E15 metatarsals on (A) day 0 in which the mineralized core was absent at the mid-diaphysis (black asterisk) and (B) after 7 and 9 days-of-culture. The red boxes indicate the Raman mapping area in each sample. (C) Raman spectra of E15 metatarsals cultured for 7 and 9 days with each line representing a different treatment: Control (black), CaPi (red), CaPi + pAsp (blue), and CaPi + fetuin-A (green) (Inset: amide bands of collagen in 1200 –1700 cm−1 range). Amide II band (purple asterisk) was only visible in the Control samples at both incubation days. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 4 0.001) compared to the Control sample (Figure 2A), and a sig- nificant increase ( p < 0.001) in the value for CaPi + fetuin-A from day 7 to day 9 of culture ( Figure 2 B). The absence of PO43− peaks in the CaPi sample suggests that, even though no mineral was formed, enough Ca and Pi infiltrated into the tissue to induce coil arrangement changes in the collagen fibrils. These observations also demonstrate the direct role of fetuin-A and pAsp in facilitating the infiltration of Ca-Pi precursors into the avascular cartilage rudiments to affect coil arrangements and promote mineralization. Mineral Development in Relation to Tissue Mineralization To investigate the correlation between the formation of HAp crystals and the collagen matrix, Raman quantitative analyses on several aspects (Figure 2C–H) were determined and com- pared. Tissue mineralization was evaluated by both the mineral to GAGs ratio (MGR) and the mineral to matrix ratio (MMR) values. The MGR distribution within the mineralized matrix of the GAGs-rich ECM (Equation 2) depicted that while all sam- ples with NCP substitutes have mineral deposition in their ECM, there were differences in the mineralization level be- tween them; with fetuin-A supplemented samples producing the highest overall MGR value on both culture days. The MGR mean values (Supporting Information: Table S2 ) were in- creased as follows: Control = CaPi (null) < CaPi + pAsp (0.171 ± 0.05) < CaPi + fetuin-A (2.785 ± 0.18) for day 7, and Control = CaPi (null) < CaPi + pAsp (1.091 ± 0.13) < CaPi + fetuin -A (3.912 ± 0.29) for day 9. All additive-treated samples were sig- nificantly different (p < 0.001) from Control and CaPi samples (Figure 2 C). Moreover, the MGR of the pAsp and fetuin -A supplemented samples increased with time in culture ( p < 0.001) suggesting a greater mineral deposition into the tissue with prolonged culture time (Figure 2D). Furthermore, when the overall quantity of minerals was com- pared to the total collagen matrix (MMR, Equation 3), the CaPi + fetuin-A sample on both culture days also yielded the highest overall value when compared with the other treatment condi- tions. The MMR mean values (Supporting Information: Table S2) were increased as follows: Control = CaPi (null) < CaPi + pAsp (0.827 ± 0.24) < CaPi + fetuin -A (3.071 ± 0.36) for day 7, and Control = CaPi (null) < CaPi + pAsp (1.104 ± 0.28) < CaPi + fetuin -A (4.151 ± 0.64) for day 9. The MMR for the CaPi + fetuin-A sample was higher (p < 0.001) when compared to the CaPi + pAsp sample ( Figure 2E). This also implies that the rate of mineralization is faster in the presence of fetuin -A than pAsp. Accordingly, when the culture time was prolonged to 9 days, the ratio increased (p < 0.001) further for both fetuin- A and pAsp treated samples ( Figure 2 F), indicating the in- crease of tissue mineralization level during this period. Mature minerals have higher crystallinity values due to the crystal apatite having a more perfect lattice organization with fewer ionic substitutions or stoichiometric arrangement. 39 The crystallinity index (CI ) ( Equation 4 ) of the CaPi + fetuin -A samples was greater than that of the other samples at both cul- ture time points. The CI mean values (Supporting Information: Table S2) were increased as follows: Control = CaPi (null) < CaPi + pAsp (0.271 ± 0.04) < CaPi + fetuin -A (0.341 ± 0.01) for day 7, and Control = CaPi (null) < CaPi + pAsp (0.313 ± 0.01) < CaPi + fetuin-A (0.352 ± 0.01) for day 9. The lower CI in the pAsp compared to the fetuin-A supplemented samples (p < 0.001) is reflective of a well-defined and organized HAp crys- tal structures with fetuin -A addition (Figure 2 G). The respective CI of both NCP supplemented samples increased ( p < 0.001) from day 7 to day 9 of the culture suggesting that crys- tal growth and maturation was a time dependent process (Fig- ure 2H). In general, tissue mineralization (MGR and MMR) has a paral- lel correlation to mineral crystallinity (CI). All quantitative analyses yielded the same results, in which the CaPi + fetuin-A sample at both culture days resulted in the highest overall value in contrast to the other treatments. Concisely, artificial HAp crystals formation inside E15 metatarsals was successfully ini- tiated and promoted by adding natural polyanionic NCP, such as fetuin-A (1 mg/mL) or its polymer substitute, such as pAsp (25 µg/mL) in a supersaturated Ca (2.5 mM) and Pi (1 mM) solution. Collagen-associated mineralization occurred in fetuin- Figure 2. Statistical analysis of E15 metatarsals culture with (A –B) CAL, (C–D) MGR, (E–F) MMR, and (G–H) CI as compared to each specific medium (left side) and culture time (right side). Zero (null) values for MGR, MMR, and CI in the Control and CaPi samples due to the absence of v1PO43− peaks. *p < 0.05; **p < 0.01; ***p < 0.001. . .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 5 A supplemented cartilage as indicated by the highest coil struc- tural change and tissue mineralization is correlated with mineral crystallinity, which reveals the interplay between col lagen fi- brils, polyanionic NCP substitutes, and Ca-Pi minerals. Mineral Localization in Cartilaginous Tissue Transmission electron microscopy (TEM), scanning -transmis- sion electron microscopy (STEM), and energy dispersive X-ray spectroscopy (EDX) observations were used to investigate min- eral localization in the tissue sections. In accordance with the Raman measurement, mineral presence was absent in the Con- trol and CaPi samples after 7 and 9 days of culture (Figure 3– TEM). Mineral was sparsely detected in the metatarsals in CaPi + pAsp sample at 7 and 9 days -of-culture. In contrast, CaPi + fetuin-A revealed the presence of minerals throughout the inner region (i.e., near cartilage core) of the metatarsals. These TEM findings have also confirmed that mineralization takes place in- side the metatarsals rather than just on their surface. Minerals in the CaPi + pAsp and CaPi + fetuin -A samples have similar plate-shaped crystals with a width/thickness of about 5 –20 nm and a length of 30–100 nm. The minerals of the CaPi + fetuin - A samples were more neatly distributed along the collagen fi- brillar network, a pattern not so obvious in the CaPi + pAsp samples. In agreement with the previous data, this indicates that fetuin-A is a robust promoter of crystal apatite deposition. Elemental mapping after 9 days -of-culture of the samples with NCP additives showed the presence of both Ca and P, confirm- ing that the tissue is deposited with calcium -phosphate mineral (Figure 3–EDX). Combined with bright field (BF) and high an- gle annular dark field (HAADF) imaging, the analysis indicated that in the CaPi + fetuin -A samples, the mineral was arranged adjacent to collagen fibrils ( Figure 3 –STEM). Overall, the mineral within the CaPi + fetuin-A sample was shown to local- ize to the perichondrium, forming an electron-dense layer about 30–40 µm thick (Figure 4A). As the perichondrium in E15 met- atarsals contains both collagen types -I and -II fibrils, the min- eralized collagen type -I fibrils within this layer represents the initiation of the bone collar, which fully forms at later stages of bone development. 31,34,40,41 Within CaPi + fetuin -A supple- mented metatarsals, some regions of the collagen fibril ( d~20 nm) were mineralized, and a cloud -like mineral complex was attached to the fibril surface (Figure 4B). These findings are in agreement with the role of fetuin-A in promoting collagen min- eralization.12,18 Interestingly, not only the collagen matrix but also the cells within the cartilage core of fetuin -A treated samples were Figure 3. Comprehensive TEM (day 7 and day 9) and STEM -EDX (day 9) observations on the inner region (i.e., near cartilage core) of E15 metatarsals cultures. Irregular-shaped cells (blue asterisks) were adjacent to the unmineralized collagen fibrillar (green arrowhea ds) network in Control and CaPi samples. The presence of minerals (orange arrowheads) not associated with collagen fibrils was detected in t he CaPi + pAsp samples. A distinguishable pattern between unmineralized (green arrowheads) and mineralized (red arrowheads) collagen fibrils was observed in CaPi + fetuin-A supplemented samples, which matched with their Ca and P distribution. BF: bright field; HAADF: high angle annular dark field; STEM: scanning-transmission electron microscopy; EDX: energy dispersive x-ray. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 6 heavily calcified. The cells were relatively small ( d~4–8 µm) with irregularly shaped morphologies ( Figure 4 C–D). These cells possibly undergo shrinkage, necrosis, and apoptosis be- cause the culture conditions were insufficient to maintain cell survival, which was followed by rapid mineral deposition and HAp crystallization within their apoptotic cytoplasm. Three-Dimensional (3D) Structure of Mineralized Embry- onic Metatarsal To reconstruct and visualize the calcified collagen matrix, fo- cused-ion beam scanning electron microscopy (FIB -SEM) to- mography was conducted on CaPi + fetuin -A treated samples since these were the only ones that contained overt areas of min- eralization. Large area mapping of a cross section through the metatarsal using the backscattered electron (BSE) revealed the distribution of HAp (Figure 5A), and two areas of the perichon- drium at the mid-diaphysis were selected as the region of inter- ests (ROIs) (Figure 5A inset) for tomography. In contrast to the initial E15 metatarsal ( Figure 1 A) before culture ( L~0.85–1 mm), the cartilage rudiments become relatively longer ( L~1.2– 1.5 mm) after being cultured with CaPi + fetuin-A (Figure 5A), indicating tissue growth. However, it is important to note that this growth is minimal (Δ L~0.2–0.5 mm) in comparison to the typical elongation of cultured embryonic cartilage rudiments during endochondral ossification (ΔL~1.5–2 mm), likely due to culture conditions that were insufficient to support prolonged cell survival.31,32 A 3D reconstruction of the transverse volumetric stacks showed that the perichondrium layer was heavily mineralized in the lat- eral direction towards the cartilage core at the mid -diaphysis (Figure 5B). A cytoplasm portion of the irregular -shaped cell adjacent to the perichondrium layer was fully embedded in min- eralized ECM, which is portrayed as a 3D mineralized peri- chondrium model. Parallel to the perichondrium layer a fila- mentous structure was observed from longitudinal image stacks and displayed as a 3D filamentous model (Figure 5C). This fil- amentous structure was likely constructed from numerous col- lagen fibrils that aggregated into a collagen fiber together with HAp crystal deposited intra - or extrafibrillarly. It was difficult to differentiate between collagen fibrils/fibers and HAp crystals as they were closely packed together with similar image conflu- ency and contrast. However, implementing a Dijkstra’s—short- est path method—algorithm42 (Supporting Information: Figure S2) to follow the trace or common path of the collagen fibrils and combining it with the previous model, it was revealed that numerous smaller filamentous-like structures and rough apatite morphology were responsible for the 3D collagen + HAp model at the perichondrium (Figure 5C and Supporting Information: Figure 4. Selected STEM (HAADF) images of E15 metatarsals culture with CaPi + fetuin-A addition. (A) The highly mineralized perichondrium layer extends to the inner region of the cartilage tissue, where numerous calcified cells (yellow asterisks) reside. (B) The partially mineralized collagen fibril (red arrows) with the mineral-fetuin-A complex adhered to the surface in a cloud-like structure (yellow arrowheads). At the cartilage core, minerals were also heavily localized in (C) the cells and (D) some regions of the collagen matrix either still unmineralized (green arrowheads) or mineralized (red arrowheads). .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 7 Video S1). Since collagen fibrils often interweave, branch, and overlap, this algorithm purposely computes optimal paths within two or more adjacent voxels that minimize intensity dif- ferences or follow geometric constraints through noisy BSE data. When the structural paths are discontinuous and the volu- metric morphology is spherical at specific site, the algorithm is unable to reconstruct fibrils, resulting in HAp classification. Therefore, each filament structure was identified as a collagen fibril that had a length of approximately 1–6 µm. All these col- lagen fibrils were interconnected and embedded within dense HAp crystals. According to this 3D model, these HAp crystals were portrayed to be associated with collagen fibrils either organized in an in- trafibrillar, extrafibrillar, or combined arrangement. This model projection also reinforces the TEM and STEM results (Figure 3 and Figure 4), indicating that HAp crystals in CaPi + fetuin - A samples were organized in a regular pattern along collagen fibrillar network. It should be highlighted that extrafibrillar crystal arrangement is predominant within the perichondrium at this time studied. Nevertheless, further evaluations are neces- sary to determine the exact location of HAp crystal deposition.

Our ex vivo embryonic metatarsal culture provides a highly suit- able and physiological ECM for investigating early bone min- eralization. This model more accurately replicates the initial bone tissue mineralization process compared to earlier in vitro approaches, such as the mineralization of self-assembled colla- gen fibrils or the remineralization of demineralized mature bone tissue. Uncalcified embryonic metatarsal tissue at E15 provides an authentic and functional cartilaginous ECM template for mineralization.31,32 Moreover, the absence of blood vessels in the tissue at this gestational age infers that the mineral infiltra- tion solely depends on a simple diffusion or convection mecha- nisms.33,34,40,43 These characteristics create an excellent model for investigating the functional role of embryonic bone ECM and polyanionic NCP substitutes in regulating Ca-Pi precursors to promote tissue mineralization during endochondral ossifica- tion. To the best of ou r knowledge, this is the first ex vivo at- tempt to utilize natural polyanionic NCP or its analogue to pro- mote artificial mineralization of the embryonic long bone. During embryonic bone development, the blood carries the Ca- Pi ions within the vascular network in its supersaturated and Figure 5. (A) Overlay area of E15 metatarsals cultured with CaPi + fetuin -A. Within the perichondrium of the mid -diaphysis (inset: close up area), the ROIs were selected for serial surface imaging in (B) transverse (blue dashed box) and (C) longitudinal (green dashed box) orientations (x, y, and z axes match between the direction in the inset and each 3D model). The representative 3D models from volumetric stacks were displayed as (B) a mineralized perichondrium model composed of the irregular cell (red) embedded within mineralized ECM (green) and (C) a filamentous model composed of calcified collagen fibrils orientated parallel to the perichondrium. The shortest path method or Dijkstra’s algorithm imple- mentation on the filamentous model exposes a structure of collagen fibrils (yellow) embedded with in minerals (green semi-transparent volume). Intrafibrillar HAp crystals deposited within the projected collagen fibrillar network are indicated by blue arrows. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 8 balanced state with respect to HAp stoichiometry, [Ca] = 2.05– 2.58 mM and [Pi] = 0.78–1.49 mM.35,36 The avascular cartilage anlagen takes up the mineral precursors from the surrounding extravascular fluid after they are released from adjacent capil- laries via paracellular pathways 44,45 and larger negatively charged macromolecules (42–68 kDa), such as albumin and fe- tuin-A enter the interstitial fluid through the opening of the tran- sendothelial channel to continuously maintain the Ca -Pi ionic equilibrium and prevent spontaneous apatite nu cleation.46–48 Accordingly, we chose two polyanionic macromolecules known to regulate the mineral precursors in vitro : pAsp (25 µg/mL with approximate size of 11 kDa or d~2.9 nm) and fe- tuin-A (1 mg/mL with approximate size of 64 kDa or d~5.3 nm) in addition to the serum-equivalent Ca and Pi concentrations of 2.5 mM and 1 mM, respectively.12,14,35 The first intriguing point of discussion is the inability of E15 cartilage rudiments to mineralize spontaneously upon exposure to a supersaturated CaPi solution, despite the rudiments pos- sessing a functional ECM and being at a developmental stage conducive to mineral deposition. It must be noted that these conditions are different from those of normal culture conditions (e.g. tissue grown in 95% air and 5% CO 2 supply31,32) where matrix mineralization continues for several days. In this latter case, the cells play a critical role in controlling matrix homeo- stasis and mineral deposition. In the case of our system, miner- alization only occurred within the tissue in the presenc e of ei- ther fetuin-A or pAsp along a supersaturated CaPi medium. The hydrated state of cartilage maintains its structural integrity by intermolecular electrostatic -steric forces predominantly be- tween negatively charged chondroitin sulfate networks within the GAGs chain.49 Together with small cartilage pores (d~6–14 nm), these properties may provide high resistance to fluid flow and water redistribution on the tissue because of low permea- bility.50–52 Moreover, in unloaded cultures, the mineral infiltra- tion rate into the tissue is also impaired in compliance with pre- vious studies.53 Therefore, it is unlikely that supersaturated Ca- Pi medium alone can promote spontaneous cartilage minerali- zation as observed in our work. Without a mineral regulator, like fetuin -A or pAsp, Ca -Pi precursors nucleate faster and spontaneously precipitate within the solution rather than diffuse into the cartilage anlagen. There is a possibility that when the Ca and Pi concentrations fall below those selected in our study (i.e., [Ca] < 2.5 mM and [Pi] < 1 mM), the precursors may not undergo spontaneous nu- cleation in solution, allowing them to infiltrate the tissue with- out the necessity of either fetuin-A or pAsp as a mineralization regulator. However, a cartilage ECM has been shown to tolerate substantially higher Ca concentrations in the presence of normal Pi levels without inducing apatite precipitation, a phenomenon not observed in aqueous solutions. 54,55 Under these conditions, precursors would merely diffuse in and out of the tissue, making mineralization unfeasible. Native polyanionic NCPs and their analogues, such as fetuin-A and pAsp inhibit crystal nucleation and precipitation in solu- tion, subsequently promoting intrafibrillar mineralization of collagen in vitro.12,14 Our results show that the incorporation of fetuin-A or pAsp within Ca and Pi-rich fluid is necessary to in- itiate mineral precursor infiltration and promote embryonic bone mineralization. One possible explanation is that these pol- yanionic macromolecules m ay form a semi -liquid or amor- phous complex with the Ca-Pi precursors in solution as widely postulated.12,14 This amorphous "ACP -NCPs" complex may adhere to the cartilage rudiment surface and progressively dif- fuse through the pores of a dense perichondrium layer to pro- mote mineralization. Our Raman measurements at the mid -diaphysis region dis- closed that the embryonic cartilage supplemented with addi- tives beside CaPi solution, especially with fetuin -A treatment resulted in similar spectra (Figure 1C) to the bone, even when compared to calcified cartilage. 56,57 Additionally, significant changes to the coil structural arrangement (Figure 2A–B) indi- cated that the collagen fibrils become more ordered due to the deposition of crystals within the fibrils thus aligning with pre- vious studies. 38 It is noteworthy that CaPi -only treatment also yielded a change in the coil structure without mineralization and the value is comparable to that observed in the CaPi + pAsp treated rudiments. In contrast, the CaPi + fetuin-A gave a higher value that may a ccount for the association of the mineral with collagen fibrils, implying this association was also absent in the CaPi + pAsp samples. Although electron microscopy in combi- nation with 3D reconstruction showed the mineral following a filamentous structure, indicative of its association with collagen fibrils (Figure 5C), it could not allow for a clear distinction of whether it is deposited within the gap zones or on the fibril sur- face. Thus, we concur that the HAp crystals either organized in an intrafibrillar, extrafibrillar, or combined arrangement. It is possible that in the presence of CaPi only, small quantities of mineral precursors can enter the cartilage rudiments and alter the coil arrangement in the ECM. However, native NCPs, such as chondrocalcin, osteopontin, osteocalcin, and bone sialopro- tein are possibly absent in the cartilage tissue at this stage of development (i.e., primary ossification) and only synthesized at a later ossification stage in accordance to previous in vivo stud- ies.58–61 Therefore, it is likely that the unexpected mineral pre- cursors that bypass the tissue hydraulic permeability barrier and infiltrate the cartilage anlagen without the support of polyan- ionic additives cannot crystallize properly within collagen fi- brils or they are only sufficient to alter the coil structural order moderately. We reveal that the perichondrium of the developing "cartilage to bone" anlagen is the specific region that is heavily mineral- ized (Figure 4A). Our finding is similar to in vivo observations that endochondral ossification of long bone begins with the for- mation of a mineralized "bone collar" layer around this site, which contains overlapping collagen type-I and type-II compo- sition at this gestational period.31,34,40 This finding is not surpris- ing because it is established that type -II collagen plays an im- portant role during endochondral ossification of bone tissues. 62 Recent in vivo observations on the endochondral ossification of mice auditory ossicles and the bony labyrinth tissues demon- strated that mineralization with high mineral density take place within the conjoint type-I and -II collagen network, suggesting a potential link b etween the presence of collagen type -II and hypermineralization of bone during endochondral ossifica- tion.63 Proteoglycans associated with collagen type -II exhibit a higher molecular mass compared to those bound to collagen type-I fibrils. 64 Given the propensity of Ca ions to localize at proteoglycan sites, 65,66 the presence of collagen type -II within the osteoid matrix—predominantly consisting of collagen type- I fibrils—may facilitate enhanced mineral deposition. This no- tion is supported by our observation of a highly mineralized re- gion within the perichondrium, consistent with the elevated mineralization typically associated with type-I and -II collagen- containing tissues. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 9 An important aspect to consider is which type of collagen in the cartilage anlagen contains the intrafibrillar deposition of crystal apatite. Although collagen type-II has a similar 67 nm periodic spacing as collagen type -I in accordance to previous studie s,67 collagen type -II cannot be mineralized intrafibrillarly in vitro by using pAsp. 68 Moreover, in vivo cartilage mineralization studies have also demonstrated that early crystal nucleation is localized within extrafibrillar sites, where proteoglycans are heavily distributed and not within collagen type -II fibrils. 65,69 Therefore, we propose that if intrafibrillar mineral deposition occurs inside E15 cartilage rudiments as the collagen becomes more ordered following the increase of CAL values and as il- lustrated by the 3D model within perichondrium, it likely occurs exclusively within collagen type-I fibrils. Our work reveals that fetuin-A promotes embryonic bone min- eralization more effectively than pAsp. Despite the similar pol- yanionic characteristics of pAsp to the acidic domain of NCP, this macromolecule is not an NCP like fetuin -A. We assume that fetuin -A o r other natural NCPs have specialized mecha- nisms for regulating the mineral precursors in the tissue, which are beyond the electrostatic attraction between mineral ions and the charged surface of macromolecules.70 Furthermore, we con- sidered the possibility that both pAsp and fetuin -A facilitate collagen mineralization of embryonic bone through the PILP process.14 However, the Raman spectra obtained are distinctly different between each additive with the PO43− intensity of pAsp significantly lower compared to fetuin -A treatment. Smaller pAsp molecules complexed with Ca-Pi precursors would be ex- pected to penetrate faster and deeper into the dense-packed col- lagen fibrils. If the PILP mechanism is the optimal way for em- bryonic bone mineralization, pAsp should have a higher pro- moting influence than fetuin -A and not otherwise. Hence, we highlight that the ISE process, which emphasizes the regulation of mineralization dynamics as the most favourable route during embryonic bone mineralization, and this aligns with previous studies.12 Overall, we demonstrate that artificial mineralization of embry- onic cartilage anlagen can be initiated and promoted with an external mineral supply and polyanionic NCP substitutes. Tis- sue mineralization is concentrated within the perichondrium, and the HAp crystals are associated with collagen fibrils either organized in an intrafibrillar, extrafibrillar, or combined ar- rangement. Nevertheless, both intra- and extrafibrillar mineral- ization are fundamental to the development of embryonic bones and the determination of their final mechanical properties.

This study represents the first demonstration of embryonic bone artificial mineralization by utilizing natural polyanionic non - collagenous protein or its analogue within an ex vivo model of endochondral ossification. While embryonic bone tissue pos- sesses an inherent molecular composition and functional ad- vantages, its extracellular matrix alone is insufficient to directly promote mineralization. However, the presence of a polyan- ionic macromolecule, such as fetuin-A or poly-DL-aspartic acid within a supersaturated calcium -phosphate medium exerts a significant and direct influence on promoting embryonic bone mineralization, potentially by stabilizing the amorphous min- eral phase. The mi neralized embryonic metatarsals have or- ganic and inorganic phases that are comparable to the bone tis- sue with hydroxyapatite as the sole mineral phase according to Raman spectroscopy analysis. Electron microscopy observation combined with a three-dimensional reconstruction depicted that hydroxyapatite crystals are heavily localized within the peri- chondrium in association with collagen fibrils either organized in an i ntrafibrillar, extrafibrillar, or combined arrangement. Overall, fetuin-A is an effective pr omoter of early mineraliza- tion during embryonic endochondral ossification. Experimental Section Reagents and Solutions All chemicals were purchased from Sigma -Aldrich (Dorset, UK), unless otherwise stated. Animal Welfare Pregnant female C57BL/6 mice were purchased from Charles Rivers Lab (Kent, UK) and maintained under conventional housing conditions with 12 hours light/dark cycle. All animal experiments were approved by The Roslin Institute’s Animal Users Committee and th e animals were maintained in accord- ance with UK Home Office guidelines for the care and use of laboratory animals. Isolation and Culture of Embryonic Mouse Metatarsals Pregnant female mice were sacrificed by cervical dislocation and their E15 embryos were collected following decapitation in accordance with home office guidelines in the UK. The middle three metatarsals were isolated and dissected under a dissecting microscope in accordance with previous protocols.32 Through- out the dissection procedure, the metatarsals were kept under preparation medium, which composed of 0.8 mL alpha mini- mum essential medium ( α-MEM) (without ribonucleosides), 10.45 mL sterile phosphate-buffered saline (PBS), and 22.5 mg bovine serum albumin (BSA) (Fraction V). For the mineralization process, metatarsals were cultured in 24- well plates with each well containing one metatarsal in 600 µL of specific culture medium with the final concentration as de- scribed (Supporting Information: Table S3). Each specific cul- ture medium was buffered to normal blood pH 7.4 with 0.25 mM hydroxyethylpiperazine ethane sulfonic acid (HEPES) so- lution and incubated in closed environment under physiological body temperature (37 °C) without a carbon dioxide (CO 2) sup- ply. The culture medium was changed every 3 days throughout the culture period and the rudiments were collected after 7 and 9 days of culture. Metatarsals were individually washed several times with ultrapure water to remove excess mi nerals on the surface and fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer solution pH 7.4 at 37 °C for 1 hour. Metatar- sals were then stored in 0.1 M sodium cacodylate buffer solu- tion pH 7.4 at 4 °C until required for processing. Raman Spectroscopy Metatarsals were characterized using a Thermo Scientific DXR3 Raman microscope equipped with 785 nm laser using 10X objective with NA = 0.25 and 3 µm laser spot size. Each metatarsal was analyzed following optimal measurement pa- rameters of 10 mW laser power and 12 s exposure time, with each spectrum as the average of 3 accumulations. Raman map- ping of each metatarsal within an area of mid -diaphysis region was carried out using 5 µm step between measurement points with focus tracking under dark field (DF) imaging condition to produce a consisted minimum 100 spectra. Spectra were .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 10 fluorescence corrected (5 th order) by the instrument OMNIC software to exclude background variations. A total of resultant 800 spectra (4 culture medium treatments for 2 time points) were then processed by OriginPro 2024b software for subse- quent asymmetric least squares (ALS) backgr ound subtraction and 5th order fast Fourier transform (FFT) filter smoothing. For quantitative analysis, integrated peak areas (ʃA) of the spec- tra were carefully selected (Supporting Information: Figure S1 and Table S4) based on the Raman assessment studies of carti- lage and bone tissue.37,56,57,71 The ʃA values under the normalized curve within each band region were determined by OriginPro 2024b software with interpolation to rectangle edges. Raman visualization was made by averaging and normalizing all repre- sentative spectra relative to amide III band in tensity in accord- ance with previous studies.37 Coil arrangement level (CAL). A relative quantification to measure structural changes in collagen fibril as a consequence of intrafibrillar mineralization. 38 This value was calculated by dividing the α-helical (ordered) proteins band (Amide III) to the random-coil (disordered) proteins band (Amide I). CAL = ∫𝐴Amide III ∫𝐴Amide I (1) Mineral to GAGs ratio (MGR). A relative estimation of the total mineralized matrix in comparison to ECM rich GAGs or rela- tive amount of bone to cartilage tissue .72,73 This value was de- termined by dividing the pyranose ring-symmetric stretch of O- SO3− band (GAGs) to the primary phosphate band (v1PO43−). MGR = ∫𝐴𝑣1PO4 3− ∫𝐴GAGs (2) Since the GAGs band highly overlaps with the tertiary phos- phate band ( v3PO43−) in agreement with previous studies37, all ʃAGAGs values for each specific treatment were approximated from the subtraction with the ʃAv3PO43− value of pure HAp min- eral spectrum ( rruff-database, https://rruff.info/hydroxylap- atite/R050512). Mineral to matrix ratio (MMR). A relative determination of the overall ratio between mineral and collagen in the tissue, which is correlated to the bone ash fraction. 74 This value was calcu- lated by dividing the primary phosphate band ( v1PO43−) to the α-helical (ordered) proteins band (Amide III). MMR = ∫𝐴𝑣1PO4 3− ∫𝐴Amide III (3) Crystallinity index (CI). A relative quantification of apatite phase, which represents the average size of crystallites, perfec- tion, and ordering in a crystal.75 This value was determined by using the width of the primary phosphate band, which is math- ematically described as the reciprocal of the full width at half maximum (FWHM) of the v1PO43− band. CI = 4.9 FWHM of 𝑣1PO4 3− (4) The value of 4.9 cm−1 corresponds to the average FWHM of the standard magmatic apatite (Conodont or phosphatic microfos- sil) and biological apatite have CI value in range of ≤ 0.5.76 Statistical analysis. A two-way analysis of variance (ANOVA) was performed with Bonferroni test to make a comparison of both culture treatments and culture times for each specific quan- titative value. Differences were considered significant at p < 0.05, highly significant at p < 0.01, and very significant at p < 0.001. Electron Microscopy For electron microscopy, metatarsals were post-fixed in 1% os- mium tetroxide in 0.1 M sodium cacodylate buffer solution pH 7.4 for 45 minutes, then washed in three (10 minutes) changes of 0.1 M sodium cacodylate buffer solution pH 7.4. Next, spec- imens were dehydrated in 50%, 70%, 90%, and 100% ethanol three times for 15 minutes each, followed by two (10 minutes) changes in propylene oxid e. Samples were then embedded in TAAB 812 epoxy resin. Transmission electron microscopy (TEM). The resin-embedded samples were cut (1 µm thickness) on a Leica Ultracut ultrami- crotome, stained with toluidine blue, and viewed in a light mi- croscope to select suitable areas for investigation. The ultrathin sections (60 nm thickness) were cut from selected areas and transferred to quantifoil TEM grids, stained in uranyl acet ate and lead citrate solutions, then examined with JEOL JEM-1400 Plus TEM. Representative images were collected on a GATAN OneView camera. Scanning-transmission electron microscopy (STEM) and en- ergy dispersive X -ray spectroscopy (EDX). Ultrathin sections (60 nm thickness) on quantifoil TEM grids were fitted on STEM grid holder, then viewed using Zeiss Crossbeam 550 equipped with a field -emission gun operation on dual channel STEM detector: bright field (BF) and high angle annular dark field (HAADF), operating at 20 kV accelerating voltage with 300 pA probes current and 3 mm working distance. Selected areas were further investigated for its elemental composition by EDX. Focused-ion beam scanning electron microscopy (FIB-SEM) and serial surface imaging . The chosen sample in resin block was attached to the carbon-covered stub, mounted to the holder, then observed with Zeiss Crossbeam 550 field -emission gun (FEG) FIB-SEM using two detectors: secondary electron (SE) and backscattered electron (BSE), operatin g at 1 kV accelerat- ing voltage with 200 pA probe current and 5 mm working dis- tance. Large area mapping of the sample using the BSE was carried out first via ATLAS 5 softwar e to produce an overlay area of the metatarsal and specific region of interest (ROI) at the perichondrium was selected. The stage was then tilted to 54o to set the sample to be perpendicular to the FIB column. A plat- inum layer was deposited on top of ROI to protect the sample surface, then a trench was milled using FIB beam (30 kV; 15 nA) to create a cross -sectional of the ROI. After polishing of the ROI cross-section, automated serial milling and slice imag- ing took place to create both SE and BSE image sta cks. Each slice (20 nm thickness) was generated with 300 pA probe cur- rent with scan speed = 4, N = 40, dwell time = 300 µs, and pixel size (X, Y) = 6.5 nm. Image analysis and three-dimensional (3D) reconstruction. The obtained BSE image stacks were carefully aligned and seg- mented using Avizo 9.0 and Dragonfly software, respectively. The raw 3D model was then processed in Blender 3.6 software by a self -made geometry nodes (Supporting Information: Fig- ure S2 ) based on Dijkstra’s shortest path algorithm. 42,77 This algorithm was used to find the globally optimal fibril pathways between two or more selection (seed) points inside a given vol- ume. The representative pathways were combined with the raw 3D model to create a new volumetric 3D model to illustrate and differentiate between inorganic minerals and collagen fibrils. The video of collagen fibrils pathway tracking and mineral lo- calization (Supporting Information: Video S1) was generated and edited in Blender 3.6 software using the same image stacks .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 11 segmentation of mineralized perichondrium layer in longitudi- nal direction which was transferred from Dragonfly software. The video was rendered using cycles render engine and OptiX denoising optimization with a high resolution of 1080p, 30 frames per second (fps), and very high contrast adjustment. Associated Content Data Availability Statement The data that support the findings of this study are available from the authors upon reasonable request. Supporting Information The Supporting Information is available. Integrated peak areas (ʃA) selection, Dijkstra’s shortest path geometry nodes, Raman spectra vibrational band assign- ments and quantitative assessment values, and final concen- tration of mineralization mediums (PDF). Video S1 (MP4). Author Information Corresponding Author *Fabio Nudelman – School of Chemistry, The University of Edinburgh, Joseph Black Building, Edinburgh, UK; : https://orcid.org/0000-0001-7464-4309; Email: [email protected] Authors †Muhammad Wisnugroho – School of Chemistry, The University of Edinburgh, Joseph Black Building, Edin- burgh, UK; : https://orcid.org/0009-0003-1761-6629 Present Address: †Department of Physics, The University of Indonesia, Biophysics Laboratory, Depok, Indonesia Fraser H. J. Laidlaw – School of Physics and Astronomy, The University of Edinburgh, James Clerk Maxwell Build- ing, Edinburgh, UK; : https://orcid.org/0000-0002-5907- 0447 Andrei V. Gromov – School of Chemistry, The University of Edinburgh, Joseph Black Building, Edinburgh, UK; : https://orcid.org/0000-0001-9254-4192 Colin Farquharson – The Roslin Institute and Royal (Dick) School of Veterinary Studies, The University of Ed- inburgh, Easter Bush, Midlothian, UK; : https://or- cid.org/0000-0002-4970-4039 Author Contributions The manuscript was written through contributions of all au- thors. All authors have given approval to the final version of the manuscript. Conflict of Interest The authors declare no conflict of interest.

The authors acknowledge financial support from the Biotech- nology and Biological Sciences Research Council (BBSRC) in the form of an Institute Strategic Programme Grant (BBS/E/RL/230001C). The authors thank the Indonesia En- dowment Fund for Education (LPDP) from the Ministry of Fi- nance, Republic of Indonesia for granting the scholarship and supporting this research. Electron microscopy data acquisitions were performed at the TEM (the Wellcome Trust Multi -User Equipment Grant WT104915MA) and Cryo FIB-SEM (EPSRC grant No. EP/P030564/1) facilities at the University of Edin- burgh. The authors also thank Martin Singleton and Stephen Mitchell from Wellcome Trust Centre for Cell Biology, Univer- sity of Edinburgh for their assistance regarding the TEM usage. For the purpose of open access, the authors have applied a CC- BY public copyright licence to any Author Accepted Manu- script version arising from this submission. Abbreviations α-MEM, alpha minimum essential medium; ΔL, length differ- ence; ʃA, integrated peak areas; 3D, three -dimensional, ACP, amorphous calcium phosphate; ALS, asymmetric least squares; ANOVA, analysis of variance; BF, bright field; BSA, bovine serum albumin; BSE, backscattered electron; Ca, calcium; CAL, coil arrangement level; CI, crystallinity index; CO2, car- bon dioxide; CO32–, carbonate; d, diameter; DF, dark field; E15, embryonic phase at 15th days; ECM, extracellular matrix; EDX, energy dispersive X -ray spectroscopy; EM, electron micros- copy; FEG, field-emission gun; FIB-SEM, focused -ion beam scanning electron microscopy; FFT, fast Fourier transform; FWHM, full width at half maximum; GAGs, glycosaminogly- cans; HAADF, high angle annular dark field; HAp, hydroxyap- atite; HEPES, hydroxyethylpiperazine ethane sulfonic acid; ISE, inhibitor size exclusion; L, length; MGR, mineral to GAGs ratio; MMR, mineral to matrix ratio; NCPs, non -collagenous proteins; pAsp, poly-DL-aspartic acid; PBS, phosphate -buff- ered saline; Pi, inorganic phosphate; PILP, polymer -induced liquid precursor; PO43−, phosphate; ROIs, regi on of interest; TEM, transmission electron microscopy; SE, secondary elec- tron; STEM, scanning-transmission electron microscopy;

(1) Roohani, I.; Cheong, S.; Wang, A. How to build a bone? - Hydroxyap- atite or Posner’s clusters as bone minerals. Open Ceramics 2021, 6, 100092. (2) Kronenberg, H. M. Developmental regulation of the growth plate. Na- ture 2003, 423, 332–336. (3) Glimcher, M. J. Recent studies of the mineral phase in bone and its possible linkage to the organic matrix by protein -bound phosphate bonds. Philosophical Transactions of the Royal Society of London. B, Biological Sciences 1984, 304, 479–508. (4) Robinson, R. A.; Watson, M. L. Collagen‐crystal relationships in bone as seen in the electron microscope. The Anatomical Record 1952, 114, 383–409. (5) Posner, A. S.; Betts, F. Synthetic amorphous calcium phosphate and its relation to bone mineral structure. Accounts of Chemical Research 1975, 8, 273–281. (6) Eppell, S. J.; Tong, W.; Katz, J. L.; Kuhn, L.; Glimcher, M. J. Shape and size of isolated bone mineralites measured using atomic force mi- croscopy. Journal of Orthopaedic Research 2001, 19, 1027–1034. (7) Reznikov, N.; Bilton, M.; Lari, L.; Stevens, M. M.; Kröger, R. Fractal- like hierarchical organization of bone begins at the nanoscale. Science 2018, 360, eaao2189. (8) Daculsi, G.; Bouler, J. M.; LeGeros, R. Z. Adaptive crystal formation in normal and pathological calcifications in synthetic calcium phos- phate and related biomaterials. International review of cytology 1997, 172, 129–191. (9) Dorozhkin, S. V.; Epple, M. Biological and medical significance of calcium phosphates. Angewandte Chemie International Edition 2002, 41, 3130–3146. (10) Copp, D. H.; Shim, S. S. The homeostatic function of bone as a mineral reservoir. Oral Surgery, Oral Medicine, Oral Pathology 1963, 16, 738–744. (11) Landis, W. J.; Song, M. J.; Leith, A.; McEwen, L.; McEwen, B. F. Mineral and organic matrix interaction in normally calcifying tendon visualized in three dimensions by high -voltage electron microscopic tomography and graphic image reconstruction. Journal of Structural Biology 1993, 110, 39–54. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 12 (12) Price, P. A.; Toroian, D.; Lim, J. E. Mineralization by inhibitor exclu- sion. Journal of Biological Chemistry 2009, 284, 17092–17101. (13) Xu, Y.; Nudelman, F.; Eren, E. D.; Wirix, M. J. M.; Cantaert , B.; Nijhuis, W. H.; Hermida -Merino, D.; Portale, G.; Bomans, P. H. H.; Ottmann, C.; Friedrich, H., Bras, W.; Akiva, A.; Orgel, J. P. R. O.; Meldrum, F. C.; Sommerdijk, N. Intermolecular channels direct crystal orientation in mineralized collagen. Nature Communications 2020, 11, 5068. (14) Olszta, M. J.; Cheng, X.; Jee, S. S.; Kumar, R.; Kim, Y.-Y.; Kaufman, M. J.; Douglas, E. P.; Gower, L. B. Bone structure and formation: A new perspective. Materials Science and Engineering: R: Reports 2007, 58, 77–116. (15) Deshpande, A. S.; Beniash, E. Bioinspired synthesis of mineralized collagen fibrils. Crystal Growth & Design 2008, 8, 3084–3090. (16) Dey, A.; Bomans, P. H. H.; Müller, F. A.; Will, J.; Frederik, P. M.; de With, G.; Sommerdijk, N. A. J. M. The role of prenucleation clusters in surface-induced calcium phosphate crystallization. Nature Materials 2010, 9, 1010–1014. (17) Wang, L.; Li, S.; Ruiz -Agudo, E.; Putnis, C. V.; Putnis, A. Posner’s cluster revisited: direct imaging of nucleation and growth of nanoscale calcium phosphate clusters at the calcite -water interface. Cryst. Eng. Comm. 2012, 14, 6252. (18) Nudelman, F.; Pieterse, K.; George, A.; Bomans, P. H. H.; Friedrich, H.; Brylka, L. J.; Hilbers, P. A. J.; de With, G.; Sommerdijk, N. A. J. M. The role of collagen in bone apatite formation in the presence of hydroxyapatite nucleation inhibitors. Nature Materials 2010, 9, 1004– 1009. (19) Jee, S.-S.; Thula, T. T.; Gower, L. B. Development of bone-like com- posites via the polymer-induced liquid-precursor (PILP) process. Part 1: Influence of polymer molecular weight. Acta Biomaterialia 2010, 6, 3676–3686. (20) Jee, S. S.; Kasinath, R.; DiMasi, E.; Kim, Y.; Gower, L. B. Oriented hydroxyapatite in turkey tendon mineralized via the polymer -induced liquid-precursor (PILP) process. Cryst. Eng. Comm. 2011, 13, 2077. (21) Li, Y.; Thula, T. T.; Jee, S.; Perkins, S. L.; Aparicio, C.; Douglas, E. P.; Gower, L. B. Biomimetic mineralization of woven bone-like nano- composites: Role of collagen cross -links. Biomacromolecules 2012, 13, 49–59. (22) Lausch, A. J.; Quan, B. D.; Miklas, J. W.; Sone, E. D. Extracellular matrix control of collagen mineralization in vitro. Advanced Func- tional Materials 2013, 23, 4906–4912. (23) Pedersen, K. O. Fetuin, a new globulin isolated from serum. Nature 1944, 154, 575–575. (24) Brown, W. M.; Dziegielewska, K. M.; Saunders, N. R.; Christie, D. L.; Nawratil, P.; Müller-Esterl, W. The nucleotide and deduced amino acid structures of sheep and pig fetuin. European Journal of Biochemistry 1992, 205, 321–331. (25) Terkelsen, O. B. F.; Jahnen-Dechent, W.; Nielsen, H.; Moos, T.; Fink, E.; Nawratil, P.; Müller-Esterl, W.; Møllgård, K. Rat fetuin: Distribu- tion of protein and mRNA in embryonic and neonatal rat tissues. Anat- omy and Embryology 1998, 197, 125–133. (26) Jahnen-Dechent, W.; Schinke, T.; Trindl, A.; Müller -Esterl, W.; Sa- blitzky, F.; Kaiser, S.; Blessing, M. Cloning and targeted deletion of the mouse fetuin gene. Journal of Biological Chemistry 1997, 272, 31496–31503. (27) Saunders, N. R.; Deal, A.; Dziegielewska, K. M.; Reader, M.; Sheardown, S. A.; Møllgård, K. Expression and distribution of fetuin in the developing sheep fetus. Histochemistry 1994, 102, 457–475. (28) Schinke, T.; Amendt, C.; Trindl, A.; Pöschke, O.; Müller -Esterl, W.; Jahnen-Dechent, W. The serum protein α2-HS glycoprotein/fetuin in- hibits apatite formation in vitro and in mineralizing calvaria cells. Jour- nal of Biological Chemistry 1996, 271, 20789–20796. (29) Ketteler, M.; Bongartz, P.; Westenfeld, R.; Wildberger, J. E.; Mahnken, A. H.; Böhm, R.; Metzger, T.; Wanner, C.; Jahnen-Dechent, W.; Floege, J. Association of low fetuin -A (AHSG) concentrations in serum with cardiovascular mortality in patients on dialysi s: a cross - sectional study. The Lancet 2003, 361, 827–833. (30) Heiss, A.; Pipich, V.; Jahnen-Dechent, W.; Schwahn, D. Fetuin-A is a mineral carrier protein: Small angle neutron scattering provides new insight on fetuin -A controlled calcification inhibition. Biophysical Journal 2010, 99, 3986–3995. (31) Thesingh, C. W.; Burger, E. H. The role of mesenchyme in embryonic long bones as early deposition site for osteoclast progenitor cells. De- velopmental Biology 1983, 95, 429–438. (32) Houston, D. A.; Staines, K. A.; MacRae, V. E.; Farquharson, C. Cul- ture of murine embryonic metatarsals: A physiological model of endo- chondral ossification. Journal of Visualized Experiments 2016, 118, 54978. (33) Osdoby, P.; Caplan, A. I. First bone formation in the developing chick limb. Developmental Biology 1981, 86, 147–156. (34) Haaijman, A.; D’Souza, R. N.; Bronckers, A. L. J. J.; Goei, S. W.; Burger, E. H. OP-1 (BMP-7) affects mRNA expression of type I, II, X collagen, and matrix Gla protein in ossifying long bones in vitro. Jour- nal of Bone and Mineral Research 1997, 12, 1815–1823. (35) Kokubo, T.; Ito, S.; Huang, Z. T.; Hayashi, T.; Sakka, S.; Kitsugi, T.; Yamamuro, T. Ca, P‐rich layer formed on high‐strength bioactive glass‐ceramic A‐W. Journal of Biomedical Materials Research 1990, 24, 331–343. (36) Lorenzo, C.; Hanley, A. J.; Rewers, M. J.; Haffner, S. M. Calcium and phosphate concentrations and future development of type 2 diabetes: The insulin resistance atherosclerosis study. Diabetologia 2014, 57, 1366–1374. (37) van der Meijden, R. H. M.; Daviran, D.; Rutten, L.; Walboomers, X. F.; Macías‐Sánchez, E.; Sommerdijk, N.; Akiva, A. A 3D cell‐free bone model shows collagen mineralization is driven and controlled by the matrix. Advanced Functional Materials 2023, 33, 2212339. (38) Kerns, J. G.; Buckley, K.; Churchwell, J.; Parker, A. W.; Matousek, P.; Goodship, A. E. Is the collagen primed for mineralization in specific regions of the turkey tendon? An investigation of the protein –mineral interface using Raman spectroscopy. Analytical Chemistry 2016, 88, 1559–1563. (39) Farlay, D.; Panczer, G.; Rey, C.; Delmas, P. D.; Boivin, G. Mineral maturity and crystallinity index are distinct characteristics of bone min- eral. Journal of Bone and Mineral Metabolism 2010, 28, 433–445. (40) Pechak, D. G.; Kujawa, M. J.; Caplan, A. I. Morphological and histo- chemical events during first bone formation in embryonic chick limbs. Bone 1986, 7, 441–458. (41) Bruder, S. P.; Caplan, A. I. First bone formation and the dissection of an osteogenic lineage in the embryonic chick tibia is revealed by mon- oclonal antibodies against osteoblasts. Bone 1989, 10, 359–375. (42) Dijkstra, E. W. A note on two problems in connexion with graphs. Numerische Mathematik 1959, 1, 269–271. (43) Maroudas, A. Biophysical chemistry of cartilaginous tissues with spe- cial reference to solute and fluid transport1. Biorheology 1975, 12, 233–248. (44) Karnovsky, M. J. The ultrastructural basis of capillary permeability studied with peroxidase as a tracer. The Journal of Cell Biology 1967, 35, 213–236. (45) Bundgaard, M. The three -dimensional organization of tight junctions in a capillary endothelium revealed by serial -section electron micros- copy. Journal of Ultrastructure Research 1984, 88, 1–17. (46) Simionescu, N.; Siminoescu, M.; Palade, G. E. Permeability of muscle capillaries to small heme-peptides. Evidence for the existence of patent transendothelial channels. The Journal of Cell Biology 1975, 64, 586– 607. (47) Palade, G. E. The microvascular endothelium revisited. In Endothelial Cell Biology in Health and Disease; Simionescu, N., Simionescu, M., Eds.; Springer: Massachusetts, 1988; pp 3–22. (48) Garnett, J.; Dieppe, P. The effects of serum and human albumin on calcium hydroxyapatite crystal growth. The Biochemical Journal 1990, 266, 863–868. (49) Seog, J.; Dean, D.; Plaas, A. H. K.; Wong -Palms, S.; Grodzinsky, A. J.; Ortiz, C. Direct measurement of glycosaminoglycan intermolecular interactions via high -resolution force spectroscopy. Macromolecules 2002, 35, 5601–5615. (50) McCutchen, C. W. The frictional properties of animal joints. Wear 1962, 5, 1–17. (51) Hardingham, T. Chondroitin sulfate and joint disease. Osteoarthritis and Cartilage 1998, 6, 3–5. (52) Majda, D.; Bhattarai, A.; Riikonen, J.; Napruszewska, B. D.; Zimow- ska, M.; Michalik-Zym, A.; Tӧyrӓs, J.; Lehto, V.-P. New approach for determining cartilage pore size distribution: NaCl -thermoporometry. Microporous and Mesoporous Materials 2017, 241, 238–245. (53) Klein‐Nulend, J.; Veldhuijzen, J. P.; Burger, E. H. Increased calcifica- tion of growth plate cartilage as a result of compressive force in vitro. Arthritis & Rheumatism 1986, 29, 1002–1009. (54) Benderly, H.; Maroudas, A. Equilibria of calcium and phosphate in hu- man articular cartilage. Annals of the Rheumatic Diseases 1975, 34, 46-47. (55) Hunter, G. K. An ion -exchange mechanism of cartilage calcification. Connective Tissue Research 1987, 16, 111–120. (56) Lakshmi, R. J.; Alexander, M.; Kurien, J.; Mahato, K. K.; Kartha, V. B. Osteoradionecrosis (ORN) of the mandible: A laser Raman spectro- scopic study. Applied Spectroscopy 2003, 57, 1100–1116. (57) Crawford-Manning, F.; Vardaki, M. Z.; Green, E.; Meakin, J. R.; Ver- gari, C.; Stone, N.; Winlove, C. P. Multiphoton imaging and Raman .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint Polyanionic Non-Collagenous Proteins and Their Analogues Promote Artificial Mineralization of Embryonic Mouse Bone Wisnugroho et al. 2025 13 spectroscopy of the bovine vertebral endplate. The Analyst 2021, 146, 4242–4253. (58) Poole, A. R.; Pidoux, I.; Reiner, A.; Choi, H.; Rosenberg, L. C. Asso- ciation of an extracellular protein (chondrocalcin) with the calcification of cartilage in endochondral bone formation. The Journal of Cell Biol- ogy 1984, 98, 54–65. (59) Strauss, P. G.; Closs, E. I.; Schmidt, J.; Erfle, V. Gene expression dur- ing osteogenic differentiation in mandibular condyles in vitro. The Journal of Cell Biology 1990, 110, 1369–1378. (60) Bianco, P.; Fisher, L. W.; Young, M. F.; Termine, J. D.; Robey, P. G. Expression of bone sialoprotein (BSP) in developing human tissues. Calcified Tissue International 1991, 49, 421–426. (61) McKee, M. D.; Glimcher, M. J.; Nanci, A. High -resolution immuno- localization of osteopontin and osteocalcin in bone and cartilage during endochondral ossification in the chicken tibia. The Anatomical Record 1992, 234, 479–492. (62) Savontaus, M.; Rintala‐Jämsä, M.; Morko, J.; Rönning, O.; Metsäranta, M.; Vuorio, E. Abnormal craniofacial development and expression patterns of extracellular matrix components in transgenic Del1 mice harboring a deletion mutation in the type II collagen gene. Orthodontics & Craniofacial Research 2004, 7, 216–226. (63) Kuroda, Y.; Kawaai, K.; Hatano, N.; Wu, Y.; Takano, H.; Momose, A.; Ishimoto, T.; Nakano, T.; Roschger, P.; Blouin, S.; Matsuo, K. Hy- permineralization of hearing-related bones by a specific osteoblast sub- type. Journal of Bone and Mineral Research 2021, 36, 1535–1547. (64) Douglas, T.; Heinemann, S.; Bierbaum, S.; Scharnweber, D.; Worch, H. Fibrillogenesis of collagen types I, II, and III with small leucine - rich proteoglycans decorin and biglycan. Biomacromolecules 2006, 7, 2388–2393. (65) Poole, A. R.; Pidoux, I. Immunoelectron microscopic studies of type X collagen in endochondral ossification. The Journal of Cell Biology 1989, 109, 2547–2554. (66) Hoshi, K.; Ejiri, S.; Ozawa, H. Localizational alterations of calcium, phosphorus, and calcification -related organics such as proteoglycans and alkaline phosphatase during bone calcification. Journal of Bone and Mineral Research 2001, 16, 289–298. (67) Studer, D.; Chiquet, M.; Hunziker, E. B. Evidence for a distinct water- rich layer surrounding collagen fibrils in articular cartilage extracellu- lar matrix. Journal of Structural Biology 1996, 117, 81–85. (68) Habelitz, S.; Hsu, T.; Hsiao, P.; Saeki, K.; Chien, Y.; Marshall, S. J.; Marshall, G. W. The natural process of biomineralization and in‐vitro remineralization of dentin lesions. In Advances in Bioceramics and Bi- otechnologies II ; McKittrick, J. M., Narayan, R., Lin, H. -T., Eds.; Wiley: Ohio, 2014; pp 13–24. (69) Arsenault, A. L.; Ottensmeyer, F. P. Quantitative spatial distributions of calcium, phosphorus, and sulfur in calcifying epiphysis by high res- olution electron spectroscopic imaging. Proceedings of the National Academy of Sciences 1983, 80, 1322–1326. (70) Manning, G. S. Limiting laws and counterion condensation in poly- electrolyte solutions I. Colligative properties. The Journal of Chemical Physics 1969, 51, 924–933. (71) Lim, N. S. J.; Hamed, Z.; Yeow, C. H.; Chan, C.; Huang, Z. Early de- tection of biomolecular changes in disrupted porcine cartilage using polarized Raman spectroscopy. Journal of Biomedical Optics 2011, 16, 017003. (72) Esmonde-White, K. A.; Esmonde -White, F. W. L.; Morris, M. D.; Roessler, B. J. Fiber -optic Raman spectroscopy of joint tissues. The Analyst 2011, 136, 1675–1685. (73) Casal-Beiroa, P.; Balboa -Barreiro, V.; Oreiro, N.; Pértega -Díaz, S.; Blanco, F. J.; Magalhães, J. Optical biomarkers for the diagnosis of osteoarthritis through Raman spectroscopy: Radiological and bio- chemical validation using ex vivo human cartilage sampl es. Diagnos- tics 2021, 11, 546. (74) Taylor, E. A.; Lloyd, A. A.; Salazar-Lara, C.; Donnelly, E. Raman and Fourier transform infrared (FT -IR) mineral to matrix ratios correlate with physical chemical properties of model compounds and native bone tissue. Applied Spectroscopy 2017, 71, 2404–2410. (75) Sa, Y.; Guo, Y.; Feng, X.; Wang, M.; Li, P.; Gao, Y.; Yang, X.; Jiang, T. Are different crystallinity-index-calculating methods of hydroxyap- atite efficient and consistent? New Journal of Chemistry 2017, 41, 5723–5731. (76) Pucéat, E.; Reynard, B.; Lécuyer, C. Can crystallinity be used to deter- mine the degree of chemical alteration of biogenic apatites? Chemical Geology 2004, 205, 83–97. (77) Everts, M. H.; Bekker, H.; Roerdink, J. B. T. M. Visualizing white matter structure of the brain using Dijkstra’s algorithm. In 2009 Pro- ceedings of 6th International Symposium on Image and Signal Pro- cessing and Analysis; IEEE: Salzburg, 2009; pp 569–574. .CC-BY 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted June 12, 2025. ; https://doi.org/10.1101/2025.06.10.658745doi: bioRxiv preprint