3.1 Agarose-collagen hydrogel: preparation and characterization
Three agarose-collagen (A-C) weight ratios were used for preparation of the hydrogels used in this paper, i.e. 0.5%-0.02%, 0.25%-0.02%, and 0.125%-0.02%, respectively. The collagen amount (equal to 0.2 mg/mL) was kept constant in all the hydrogel formulations. This concentration, close to that used in other studies in which blended hydrogels were prepared, is sufficient to provide the anchoring sites to the embedded cells. On the other hand, the amount of agarose was varied from 0.5 to 0.125% to evaluate the impact on the biophysical properties of the hydrogel and consequently their effect on the cellular spheroids development. To prepare the hydrogels and, at the same time, to prevent the denaturation of collagen and the cell damage and death, the agarose solution was first heated until boiling (100 °C) to dissolve the polysaccharide, and then cooled down until it reached 45 °C. At this point, since gelation of pure agarose occurs at around 36 °C, it was rapidly mixed with collagen and the cell suspension. The preparation sketch is reported in Figure 1S.
The physical properties of the A-C hydrogels were characterized. The study of the hydrogels surface structure and topology through SEM techniques shows that decreasing the agarose percentage, the porosity of the hydrogels increases and the whole structure appears less compact (Figure 1). In 0.125%-0.02% A-C hydrogels, indeed, the higher percentage of water makes the dry structure very brittle displaying pores that are quite uniform and interconnected with a mean size of 71 ± 14 µm (Figure 1e and f). Although the measure of the pore sizes of a dry structure cannot be considered realistic, it provides an estimation of the 3D organization of the hydrogels. In 0.25%-0.02% and 0.5%-0.02% A-C hydrogels, the number of pores is smaller, while the average pore size appears to be slightly larger (81 ± 21 and 87 ± 25 µm, respectively, Figure 1a-d). The higher turgidity and compactness of the hydrogels with higher percentage of agarose is also evident at a macroscopic level, as shown in the bottom panel of Figure 1S. Indeed, while 0.25%-0.02% and 0.5%-0.02% A-C hydrogels appear self-standing, that with 0.125%-0.02% weight ratio looks softer and less compact.
FTIR spectroscopy was performed on the hydrogels to investigate whether collagen, though the low percentage amounts used, was detectable on the outer surface of the hydrogels, thus confirming its distribution within the hydrogel texture. To this aim, the analysis was performed directly on the synthetized hydrogels deposited on the ATR crystal, without any further manipulation. Pure agarose hydrogel (0.25%) was recorded as reference (light green line in Figure 2a), showing the typical signals at 988 and 1076 cm-1, relative to the C-H bending and to the C-O stretching of the glycosidic bonds, and two broad peaks at 1656 and 3421 cm-1, characteristic of the stretching of the H-O-H bound water and of the O-H hydrogen bonded carbohydrate hydroxylic groups, respectively (Oza, Prasad, & Siddhanta, 2012). Two broad peaks were also observed in the pure collagen hydrogel at about 3328 and 1645 cm-1 for the amide C=O and N-H stretching, respectively (pink line in Figure 2a). Additional smaller signals were detected at 1051 cm-1for the C-OH stretching vibrations of carbohydrate moieties attached to the protein (Petibois, Gouspillou, Wehbe, Delage, & Déléris, 2006). The resulting FTIR spectrum of the blend 0.25%-0.02% A-C hydrogel (red curve of Figure 2a) showed all the peaks characteristic of the pure compounds, i.e. smaller signals at 989 and 1079 cm-1, with a small side-bump at 1045 cm-1, and much broader peaks at 1649 and 3464 cm-1. Interestingly, while the last two peaks had similar intensity in pure collagen, a much higher intensity of the signal at lower frequencies was observed in pure agarose as well as in the blend hydrogel. As expected, no significant differences were observed in the spectra of the other two blend hydrogels (data not shown).
Another critical feature of a hydrogel is the capability to absorb and retain water. Figure 2b) shows the swelling behavior of the three types of hydrogels. The hydrogels containing 0.5% and 0.25% agarose show a similar trend with a swelling ratio respectively of 20 and 25 times at t0 and a maximum swelling of 27 and 30 times after 24 h in PBS. These data are in accordance with those reported in the literature. On the other hand, the ability to absorb water of the hydrogel with 0.125% agarose is considerably lower, with a swelling ratio of 5.2 times at t0 and 7.1 at t24, respectively. Thus, there appears to be a critical threshold of agarose percentage below which the physical properties of the hydrogel are dramatically altered.
A degradation test was carried out to investigate the stability of the three formulations over time. Panels c) and d) of Figure 2 show the percentage residual weight of the hydrogels kept at 37 °C in PBS and DMEM up to two weeks. 0.5%-0.02% and 0.25%-0.02% A-C hydrogels were shown to be quite stable to degradation and lost around 10 and 15% of their weight after 14 days of incubation in PBS and DMEM with 10% FBS, respectively. On the other hand, 0.125%-0.02% A-C hydrogels showed a maximum of degradation close to 40% after 2 weeks in both incubation media.
In parallel, the amount of collagen potentially released was estimated via a protein quantification assay. To this aim, the A-C hydrogels were kept in PBS at 37 °C and the volume of PBS was collected and renewed every 24 h up to 14 days. Figure 2S shows the results of the BCA assay carried out with the three types of hydrogel. Collagen release was detected in all hydrogels but to a higher extent and with a quicker trend in that with a lower agarose content. The overall percentage amount of collagen release after 2 weeks is equal to 50%, 27%, and 8% of the whole collagen present in the hydrogel containing 0.125%, 0.25%, and 0.5% agarose, respectively. This loss, together with the less compact texture of the 0.125%-0.02% A-C hydrogel, would explain the higher degradation of this matrix.
A further parameter that deserves investigation is the capability of a hydrogel to allow the diffusion throughout the matrix of biomolecules and nutrients, such as growth factors and serum proteins having molecular weight of several tens of kDa. To this aim, a diffusion test was performed by using two fluorescent probes, FITC-conjugated hyaluronic acid and Transferrin-TRITC, with a MW of around 10 and 80 kDa, respectively. Once loaded onto the three hydrogel formulations, the variation of concentration of the feeding solution was monitored over time and reported versus the diffusion in pure PBS. The results show that hyaluronic acid (Figure 2e) diffuses rapidly into the softest hydrogel reaching 100% diffusion after 72 h, while the other two hydrogels show a similar trend although with a more gradual diffusion that decreases by increasing the percentage of agarose, reaching a maximum of 82% and 75% after one week, respectively. On the other hand, in the case of transferrin-TRITC the diffusion is much slower: in the 0.125%-0.02% A-C hydrogel the diffusion reaches almost 48% after 24 h, while at the same time point it is closed to 10% in the case of the other two blends. (Figure 2f) Noteworthy, after one week the diffusivity does not exceed 50% even in the softest matrix.
The structural properties of the A-C samples were also evaluated by mean of compression test under physiological-like conditions (PBS, 37 °C). Figure 3a) shows how the agarose concentration deeply affects the hydrogels’ mechanical properties. In fact, the increase in agarose concentration corresponds to an increase in the compressibility modulus. As expected, the 0.5%-0.02% ratio shows the highest E modulus (5.7 ± 0.5 kPa), followed by 0.25%-0.02% (1.6 ± 0.4 kPa) (p = 0.0004) and 0.125%-0.02% (0.7 ± 0.2 kPa) (p = 0.0001). A minor but significant difference was found between 0.25%-0.02% and 0.125%-0.02% (p = 0.03). In order to correlate degradation resistance to structural stability over time, the A-C samples’ mechanical performances were evaluated after 0, 1, 4, and 8 days of incubation in physiological-like conditions (PBS, 37 °C, in humified atmosphere with 5% CO2). While 0.125%-0.02% hydrogels could not be tested over time due to their low consistence, the 0.5%-0.02% and 0.25%-0.02% blends retained their structural integrity until the 8th day of measure. As shown in Figure 3b, no significant changes in the E modulus were registered in the case of 0.5%-0.02% and 0.25%-0.02% hydrogels. These data are in accordance with the degradation tests in which a minimum weight loss was recorded.