Incorporating zinc into HA was reported to exhibit a significant effect on the cytotoxicity of scaffolds against bone osteosarcoma cells, where increasing zinc content increased the cytotoxicity% [3]. Hence, Zn-HA was prepared and evaluated for its physicochemical properties. XRD patterns of the pure HA and doped samples with different content of zinc (2 and 6%) are shown in Fig. 1. For all samples, there are similar diffraction peaks that are associated with the monoclinic structure of HA (JCPDS no. 76–0694) where the crystal structure of HA was maintained despite the addition of zinc. The effect of doping on the crystallite size and the lattice parameters (a and c) of the samples were calculated by Rietveld refinement and illustrated in Table 2. As presented in Table 2, doping HA with increasing concentrations of zinc ions resulted in a noticeable decrease in crystallite size. This observation is consistent with literature findings, where zinc ions—due to their smaller ionic radius compared to calcium—substitute calcium in the HA lattice, creating lattice strain and defect sites that hinder crystal growth [44,45,46]. This leads to smaller crystallites and increased surface energy, which could positively influence scaffold resorption and cellular responses in vivo."

XRD patterns (a) and FTIR spectra (b) of HA and Zn-doped samples (2 and 6%)

The FTIR spectra of all prepared samples are illustrated in Fig. 1. The spectra of pure HA and Zn-HA showed the characteristic bands corresponding mainly to the HA structure. All the spectra showed double bands at 605 and 563 cm−1 which were assigned to the ν4 vibrational mode of the phosphate group [47]. The weak band that appeared at 473 cm−1 is attributed to the ν2 vibrational mode of the phosphate group. The stronger bands observed at 1020–1100 cm−1 as a shoulder in all spectra are related to ν3 vibrational mode corresponding to the phosphate group [48, 49]. The bending vibrational mode of the OH group appeared at 630 cm−1 [44].

FT-IR studies are primarily sensitive to vibrations of covalent bonds and functional groups; they can also provide indirect information about ionic substitutions through lattice vibrations or changes in associated functional group bands. However, in this study, no significant shifts or new peaks were observed in the phosphate or hydroxyl regions, suggesting that there is no clear FT-IR evidence of Zn2⁺ incorporation into the crystal structure of HA [45]. The addition of zinc into HA does not result in observable new FTIR peaks, as Zn2⁺ ions integrate into the lattice by occupying ionic positions rather than forming covalent bonds. While zinc doping might induce subtle shifts or distortions in pre-existing spectral bands (such as phosphate-related vibrations), these changes are often minimal and insufficient to definitively confirm zinc incorporation using FTIR alone [50]. Direct verification of zinc presence necessitates alternative analytical methods, such as XRD, X-ray photoelectron spectroscopy (XPS), or energy-dispersive X-ray spectroscopy (EDX). As a result, FTIR is not a reliable tool for detecting zinc within Zn-doped HA.

The TEM micrographs of HA and Zn-doped HA are shown in Fig. 2. It can be seen that the doping process of HA with zinc ions has a significant effect on the particle size of obtained samples where the size reduces with increasing zinc content. TEM images also showed that the pure HA formed in a round particulate morphology while the formation of rod-like particle crystallites increases with the higher content of Zn ions [46].

TEM micrographs of HA and Zn-doped samples (2 and 6%)

In a previous study [3], Elsayed et al. investigated the impact of three key variables on alginate-gelatin hydrogel scaffolds: Zn-HA morphology, crosslinking methods, and the effect of 5-FU loading. Zn incorporation reduced HA crystallite size and altered its morphology, shifting from rounded particles to rod-like crystallites. Higher Zn content also led to increased cytotoxicity against bone cancer cells. The scaffolds were crosslinked using calcium chloride (CaCl₂) alone or with glutaraldehyde, with the latter slightly modifying scaffold appearance but not significantly affecting drug incorporation or thickness. 5-FU was successfully incorporated into the hydrogel without chemical interaction, and drug-loaded scaffolds exhibited an initial burst release within 10–15 min, followed by complete release in four hours. The cytotoxicity of scaffolds increased with Zn doping and 5-FU loading, demonstrating promising potential for targeted bone cancer therapy.

Different Zn-HA hydrogel scaffolds were reported to show promising cytotoxicity against bone osteosarcoma cells with almost no effect on the normal cells [3]. In the current study, the prepared blank and drug-loaded scaffolds were evaluated in terms of surface area, pore size, morphology as well as cell attachment capability. The scaffolds were also stored for one year and evaluated for their physical and chemical stability.

The gas adsorption technique was utilized for surface area and pore volume analyses due to its ability to provide accurate and quantitative assessment of micro- and meso-porosity in porous biomaterials. These parameters are critical in scaffold-based drug delivery systems as they directly influence drug loading capacity, release kinetics, and cell–material interactions [36, 37]. The surface area can be determined using the Brunauer–Emmett–Teller (BET) model based on the relative pressure of gas [36]. The pore volume and the surface area could also be estimated using the Barrett-Joyner-Halenda (BJH) method or the density functional theory (DFT) [37]. In the BET method, the surface area is calculated from the adsorption isotherm by fitting to the BET equation without accounting for the pore size [38]. This method has its assumptions and limitations; however, it is appropriate for samples with uniform gas adsorption. On the other hand, the BJH method calculates the surface area and pore size distribution from the gas desorption and relates the pore size to the volume of gas adsorbed at a specified relative pressure [37]. This method is sensitive to the pore geometry as it assumes a cylindrical pore shape. The DFT method is a more advanced approach and is recommended by “The International Standard Organization” (ISO) to calculate pore size distribution [37].

In this study, the surface area, and pore volume of the blank and drug-loaded scaffolds were evaluated using BET, BJH, and DFT methods. Despite the previously mentioned differences, the results calculated using either of the three methods showed higher surface area and pore volume values of the blank scaffolds compared to the drug-loaded ones (Scaf-1 & Scaf-2, Table 3). Different studies have suggested that the large surface area of scaffolds is correlated with smaller pore sizes [13, 15]. For example, very small pores might hinder cell migration toward the center of the construct and hence, limit the diffusion of nutrients and removal of waste. Conversely, very large pores reduce the available specific surface area, affecting cell attachment. In this regards, a conflicting data were reported in the literature [35] [13, 15]. It is still unclear how scaffold pore size and cell activity are related. Previous bone tissue engineering studies have suggested an optimal range of mean pore sizes (96–150 μm) for facilitating attachment. Other research highlights the need for larger pores (300–800 μm) to support successful bone growth within scaffolds [35].

SEM imaging was carried out to better evaluate the pore morphology and size, especially in the large size range [37, 51]. Figure 3 shows the morphology of the blank scaffolds (BScaf-1 and BScaf-2) and drug-loaded ones (Scaf-1 & Scaf-2). The blank scaffolds showed a less porous morphology compared to the drug-loaded ones that showed interconnected pores and needle-shaped drug particles. scaffolds’ pore sizes were measured by ImageJ® software. This software is frequently used to analyze SEM images to measure pore sizes and porosity [3, 16, 52, 53]. The average pore size of Scaf-1 was smaller than Scaf-2 (mean = 9.79 μm and 16.19 μm, SD = 20.5 and 47.8, respectively). These results are in agreement with the studied surface area values and the reported correlation between the smaller surface area of the scaffolds and the larger pore size [13, 15]. It is worth mentioning that the high standard deviation values indicate the non-uniformity of the pore size of the hydrogel scaffolds, which is considered a preferred characteristic of this type of scaffold. The complexity of their non-uniform porous structure allows better mechanical properties and maintains their elastic state, which is critical for the effective use of implanted biomaterials [3, 54].

SEM images of the blank scaffolds (BScaf-1 and BScaf-2) and the drug-loaded scaffolds (Scaf-1 & Scaf-2)

SEM imaging was also carried out after soaking the drug-loaded scaffolds in sterile media for 3 days to evaluate the morphology and pore size after 5-FU release (Fig. 3). It was found that the needle-shaped crystals of 5-FU disappeared, and the scaffolds retained the porous structure with average larger pore sizes (mean = 18.37 μm, and 27.14 μm, SD = 102 and 115 for Scaf-1 and Scaf-2, respectively) compared to the non-soaked scaffolds. The increase in pore size after soaking the scaffolds might be attributed to the biodegradability of the hydrogel-forming materials used as well as the drug release [55].

While the visual observation via SEM might initially seem contradictory to the higher total surface area and pore volume obtained by gas adsorption for the blank scaffolds, this can be understood by considering the nature of the information each technique provides. Gas adsorption methods quantify the total accessible surface area and pore volume across a broad range of pore sizes, encompassing contributions from both larger and smaller pores. In contrast, SEM provides direct visual information primarily highlighting the morphology and presence of larger pores. Therefore, the higher total surface area and pore volume values for the blank scaffolds determined by gas adsorption strongly suggest a greater contribution from smaller pores, which collectively add significantly to the surface area but are less visually prominent in SEM images, aligning with the reported correlation between larger surface area and smaller pore sizes [13, 15].

The weight loss percentage in DMEM medium after 3 days indicated that the drug-loaded scaffolds (Scaf-1 and Scaf-2) experienced relatively faster weight loss compared to their corresponding blank scaffolds (BScaf-1 and BScaf-2), due to the release of 5-FU (Online Resource 1). Additionally, scaffolds prepared with 0.1% GA in 0.2M CaCl₂ as a crosslinking solution (BScaf-2 and Scaf-2) exhibited comparatively slower weight loss compared to those prepared with 0.2M CaCl₂ alone (BScaf-1 and Scaf-1, Table 1). This is might be due to the effect of GA as a crosslinking agent for gelatin, which enhances its mechanical properties and slows its biodegradation [24, 25].

The current study is an extension of a previously published work, in which the biocompatibility of the scaffold, including MTT assays, was thoroughly investigated and reported [3].

Figure 4 shows the studied scaffolds loaded with cells. The cells adhered to the scaffolds in the form of spindle shape-like structures which confirmed the biocompatibility of the scaffolds with the cells. The cells tended to cover the entire region of the presoaked drug-loaded scaffolds, indicating their mature development. Furthermore, the cells appeared to grow and disseminate across the porous surface, ensuring a constant supply of nutrients. While the soaked blank scaffolds in Fig. 4 showed round-shaped cells on the surface, most of them spread randomly. The previous observations proved that cell growth on the presoaked drug-loaded scaffolds was markedly better than on the blank scaffold. This might be attributed to the larger pore size of the presoaked drug-loaded scaffolds compared to the blank ones. Although the average pore size of the studied scaffolds (both the drug-loaded and blank ones) was not within the range previously reported as optimal for cell attachment, the results indicated the predominant effect of the surface area as well as the biodegradability of the used hydrogel-forming polymers alginate and gelatin. Several studies have reported that scaffolds with smaller pores have larger surface areas, which provide increased sites for cellular attachment [12,13,14,15]. It is worth noting that the pore size of the studied scaffolds is greatly affected by the biodegradability of the used hydrogel-forming polymers alginate and gelatin [18, 19].

SEM images of the blank scaffolds (BScaf-1 and BScaf-2) and the drug-loaded scaffolds (Scaf-1 & Scaf-2) after loading with cells

The biodegradable biocompatible polymers used are well tolerated in the body and mimic the natural bone matrix, thus helping bone regeneration and healing. This role could allow the body to attain homeostasis again after tumor excision surgeries. With time, after acting as structural support, the body replaces the scaffolds with normal bone cells integrating the HA into its structure.

After soaking, the average pore sizes of the drug-loaded and blank scaffolds appeared more comparable based on SEM imaging. However, several factors beyond average pore size likely contributed to the observed differences in cell attachment. Specifically, the drug-loaded scaffolds exhibited greater pore interconnectivity and more uniform surface topography, which can enhance cell spreading and infiltration. Additionally, the release of 5-FU during the presoaking process may have induced localized microstructural changes or surface softening not easily captured by SEM, creating a more favorable environment for cell attachment. Residual effects from drug release may also have influenced surface chemistry, enhancing protein adsorption and cell adhesion.

The type of crosslinking solution and the zinc content both play a role in influencing the scaffold’s structure and biological performance. The use of glutaraldehyde (GA) as a crosslinker in Scaf-2 and BScaf-2 likely contributed to the formation of a denser, more chemically stable hydrogel network compared to ionic crosslinking with calcium ions in Scaf-1 and BScaf-1. This denser network may have enhanced pore stability and shape retention during soaking, indirectly supporting improved cell attachment by preserving scaffold architecture. Additionally, increasing the zinc content in HA was associated with a reduction in crystallite size (as shown in Table 2), which can enhance the scaffold’s surface reactivity and promote better protein adsorption and cell–material interactions. Zinc also has recognized biological functions, including stimulating osteoblastic activity and promoting cellular adhesion, which may contribute positively to cell compatibility. While the current study focused primarily on scaffold morphology and adhesion behavior, these combined effects of crosslinking chemistry and zinc incorporation likely played a role in the enhanced cell attachment observed in the drug-loaded formulations.

Evaluation of the long-term stability of scaffolds is indispensable in assessing their performance as implantable drug delivery systems. Understanding the stability profile of these systems is crucial for ensuring their efficacy and safety over time, especially in clinical applications. Stability assessments involve monitoring various parameters such as physical integrity, chemical composition, and drug release kinetics under different storage conditions. By conducting stability studies, formulation parameters can be optimized and the stability can be enhanced [56, 57]. These studies can be accomplished utilizing X-ray diffraction (XRD) and Fourier-transform infrared (FT-IR) spectroscopy, instrumental techniques [58, 59]. XRD provides valuable information about the crystallinity of materials and changes in crystal lattice parameters and phase composition. By analyzing diffraction patterns obtained from XRD experiments, alterations in the crystallinity and phase transitions of the scaffold matrix or drug molecules during storage can be detected [58]. Similarly, FTIR spectroscopy enables the identification of functional groups and chemical bonds present in the scaffold formulation. By monitoring changes in FTIR spectra over time, the stability of scaffold-drug interactions and any chemical modifications or degradation phenomena can be assessed.

Long-term scaffold stability is critical for clinical translation, particularly in implantable drug delivery systems. Ensuring unchanged crystallinity (XRD), chemical functionality (FT-IR), and drug release profile after one year confirms suitability for real-world storage and use, ensuring efficacy and safety [41, 57, 59].

After storage for one year at 30 °C ± 2 °C and 65 ± 5% relative humidity in well-closed glass containers, the studied scaffolds were evaluated for their physical and chemical stability as well as 5-FU release profiles.

The stored hydrogel scaffolds were inspected visually for any changes in color and appearance. All the studied scaffolds maintained their yellowish-white to yellowish-brown color after storage. The thickness of the studied scaffolds was also evaluated, and no significant differences were observed in the scaffold thickness as a result of storage (p-value > 0.05, Online Resource 2).

Figure 5 shows the release profiles of 5-FU from the freshly prepared and stored scaffolds. The results showed that the release rate of 5-FU from Scaf-1 (cf. Table 1) was similar before and after storage as depicted by the values of the difference ( f 1) and similarity (f2) factors ( f 1 = 9.57& f 2 = 63). While, the release rate of 5-FU from Scaf-2 (cf. Table 1) was slightly decreased after storage, however, this decrease was not statistically significant ( f 1 = 10 & f 2 = 58).

5-FU Cumulative % released from the fresh and one-year stored scaffolds; Scaf-1 and Scaf-2

The difference factor ( f 1) calculates the percentage difference at each time point and evaluates the relative error between them [41, 60]. The two release profiles are identical when f 1 equals zero and their similarity is indicated at values below 15 [61]. While, the similarity factor (f2) is a function of mean differences between the two release profiles [41, 62]. A value of 50 means a 10% average difference at all given time points and the similarity of the two profiles is indicated above this value [60, 63].

It is worth mentioning that 5-FU release from different types of scaffolds could be modulated to be either fast or extended utilizing different types of polymers and/or different cross-linking techniques [18, 19]. Burst and fast release were reported within 1–2 h [2, 34] while extended release over longer periods was also reported [33, 64, 65]. The drug release from various scaffold types might be customized according to the personalized needs, such as obtaining a sustained drug release for long-term therapy or filling the gap following surgical tumor removal and promoting normal tissue regeneration to prevent relapse [2,