Effect of polyelectrolyte ratio and temperature on complexation

The rationale behind this study lies in the hypothesis that PECs can be formed by the electrostatic interactions between the negatively charged EH and the positively charged PEI or CH, and that these PECs can be exploited as drug delivery systems. An explorative screening was initially conducted to understand the effect of the polyelectrolyte ratio and temperature on the complexation process, and to identify the conditions for the fabrication of insoluble PECs with a macroscopically gel-like or solid-like behaviour. For the EH/PEI complexation, the working solutions were pH-adjusted to 8, to maximize EH charge density. In fact, EH, as other lignosulphonates polymers, is negatively charged at all practical pH values due to the presence of the sulphonate groups, but importantly, its charge density increases by increasing the pH due to the progressive deprotonation of carboxyl groups (pKa ca. 2.7) and phenolic hydroxyl groups (pKa = 6–11) [10, 11, 26]. This working pH value of 8 could not be used for the EH/CH complexation, due to the fact that CH is only soluble below its pKa of ca. 6.5. For this reason, EH/CH complexation was carried out at pH 5, a condition under which both polymers are soluble and charged [10, 11, 26].

The effect of the polyelectrolytes ratio (EH:PEI and EH:CH ratios from 1:3 up to 9:1) and temperature (from − 20 °C up to 90 °C) was studied, and the visual outcomes (liquid-, gel-, or solid-like behaviour of the polymeric mixtures) are summarized in Table 1.

Table 1 Effect of polyelectrolytes ratio and temperature on the formation of EH/PEI and EH/CH complexes. Prototypes selected for follow up experiments are shown with a bold, underlined symbol

All polymeric mixtures presented a brown appearance due to the presence of EH, which is a brown polymer due to the high content of lignin [10]. Overall, EH seemed to form more stable complexes with PEI compared to CH. Indeed, the majority of EH/PEI mixtures presented a gel-like or solid-like behaviour, except for those prepared with EH:PEI ratios of 1:3 (at 4 and 23 °C), 5:1 and 9:1 (at all temperatures < 90 °C), which presented a liquid-like behaviour, likely due to the unbalanced charge stoichiometry which triggered macro-phase separation. This is the separation of the system in distinct phases on a large, macroscopic scale, characterized by regions with predominantly polyanionic (EH-rich domains) or polycationic (PEI-rich domains) character, leading to a system with lost structural integrity and a liquid-like behaviour, due to the lack of homogeneously distributed electrostatic interactions. The charge density of EH and PEI at pH 8 is 2.2 meq/g and 4 meq/g, respectively. It follows that a EH:PEI mass ratio of approximately 2:1 is necessary to achieve a charge stoichiometry of 1:1, which is the ratio under which insoluble PECs are usually formed [26]. This explains why mixtures prepared with a EH/PEI ratio close to 2:1 had a better macroscopic outcome at any temperature (gel-like or solid-like behaviour, indicated in the table with + and ++, respectively) compared to mixtures prepared with more unbalanced ratios. Overall, when using these unbalanced EH:PEI ratios (i.e., 1:3, 5:1 and 9:1), the increase of temperature resulted in the formation of homogeneous gel-like mixtures, likely because at higher temperatures the mobility of the polymeric chains increases, allowing a better mixing and limiting the macro-phase separation. It needs to be noted that for samples treated at –20 °C, starting solutions were quickly mixed at room temperature, subsequently frozen at –20 °C, and finally thawed to room temperature for the evaluation of the gel-like behaviour. It follows that the samples likely underwent a partial complexation at room temperature, which was subsequently slowed down or temporarily stopped during freezing and finally reactivated during thawing. Mixtures with a EH:PEI ratio of 2:1 (at 23 °C) and of 9:1 (at 90 °C) were selected for follow up experiments. The first one was chosen because it was the only solid-like mixture among those prepared at the most convenient temperature of 23 °C. The second one was chosen since it was the only mixture with a gel-like behaviour among those prepared with the highest EH:PEI ratio, which is considered beneficial to obtain a high loading of aromatic, positively charged drugs, such as methylene blue tested in this study.

In contrast to EH/PEI mixtures, the majority of EH/CH mixtures presented a liquid-like behaviour at almost all tested temperatures, likely due to the much lower polymer concentration used, which in turn was dictated by the lower CH solubility, compared to that of PEI. Only when operating at 90 °C, gel formation was visible for mixtures with EH:CH ratios of 1:3 and 1:2. This finding is in line with the results reported by Fredheim et al., who observed compact precipitates during complexation of chitosan and lignosulfonate only at 100 °C, whereas at 50 and 4 °C the complexation resulted in fluffy aggregates [26]. This highlights that the complexation process is entropy-driven, as usually found for polyanion/polycation complexation, where an increase in entropy due to the release of counter-ions associated with the polymers drives the establishment of polymer-polymer electrostatic interactions [26, 28]. To facilitate the formation of more stable, gel-like EH/CH complexes, the addition of AG as a thickening agent was studied. AG is a galactose-based polysaccharide extracted from red algae and it is composed of agarose and agaropectin, usually in a 70/30 ratio [29]. AG is a negatively charged polysaccharide due to the presence of sulphate and pyruvate groups in agaropectin, and it is currently studied for the formation of hydrogels [30]. The most stable complex was observed when using an agar concentration of 20 g/L, which corresponds to a EH:CH:AG ratio of 1:1:2. Therefore, this prototype was selected for follow up experiments, together with the mixture having a EH:CH ratio of 1:2, which was the mixture with the highest EH content that presented a gel-like behaviour.

In summary, the selected prototypes were those made with the following EH/PEI or EH/CH ratios under the specified temperature conditions: (i) EH:PEI 2:1 at 23 °C, further referred to as EP 2:1; (ii) EH:PEI 9:1 at 90 °C, further referred to as EP 9:1, (iii) EH:CH 1:2 at 90 °C, further referred to as EC 1:2; (iv) EH:CH:AG 1:1:2 at 90 °C, further referred to as EC 1:1 + A. Finally, as a proof-of-concept, the mixture with a EH:PEI ratio of 1:1 at 23 °C was selected for the qualitative assessment of gelation and injectability, as described in the next section.

Qualitative assessment of gelation and injectability

Gelation and injectability of a mixture with a EH:PEI ratio of 1:1 at 23 °C was qualitatively assessed, as its behaviour is representative of all other mixtures of the study with a gel-like behaviour (marked with + in Table 1). The mixture had a brown, gelling appearance and did not flow under the action of the gravitational force when the vial was turned upside-down, which confirms the formation of a stable hydrogel network. The hydrogel was easily taken up using a positive displacement pipette and, after releasing it on a plastic Petri dish, it retained a drop-like three-dimensional shape (Fig. 2A). High shape fidelity of the generated drop was maintained also after turning the Petri dish upside-down (Fig. 2B). This observation indicates that the hydrogel has a shear-thinning behaviour, i.e., it is able to flow through the tip of a pipette under the applied shear due to the temporary break-down of the physical network. Importantly, the maintenance of the three-dimensional shape after injection demonstrates that the physical network is quickly re-established when the shear stress is removed (recovery phase) [31, 32]. This feature has a high interest in the biomedical field since hydrogels with shear-thinning/fast recovery behaviour are widely investigated as injectable and 3D printable biomaterials in drug delivery and tissue engineering [33,34,35].

Fig. 2

Photographs of a hydrogel made of a mixture with a EH:PEI ratio of 1:1. Shape-stable drop, generated after injecting 100 µL of hydrogel on a Petri dish (A). Shape stability is maintained also after turning the Petri dish upside-down

Thermal properties

DSC analysis is a valuable tool for gaining knowledge on polymer interactions within PECs, by studying, for example, the migration of thermal decomposition peaks associated with depolymerization and pyrolytic reactions of polyelectrolytes [36]. DSC curves for the selected samples are shown in Fig. 3 (separated DSC curves with numerical intervals on the Y-axis are provided in Fig. S1 in the Supplementary Data). The first thermal effect is an endothermic event in the temperature range of 30–150 °C. EC 1:2 and EC 1:1 + A prototypes showed clear peak temperatures at 101 and 113 °C in this region, respectively, whereas for both PEI-containing samples much broader peaks were observed. This thermal event is attributed to the evaporation of water, as previously observed in DSC thermograms of hydrophilic polymers, such as PEI and CH, that are able to attract a significant amount of water [36]. DSC curves showed a more pronounced water evaporation for the EC groups compared to the EP groups, suggesting that in the presence of chitosan, PECs have a higher moisture content. This is in line with our previous findings on the role of water on multilayers based on EH and CH or EH and PEI through electrostatic interactions [11]. In that study, for each of the systems (EH/CH and EH/PEI), we calculated the total adsorbed mass (polymer and associated water) by Quartz Crystal Microbalance with Dissipation (QCM-D) and the amount of adsorbed polymer (without associated water) by Stagnation Point Adsorption Reflectometry (SPAR). Results demonstrated that EH/CH multilayers showed more adsorbed mass per unit area compared to EH/PEI films as measured by QCM. However, SPAR data indicated that EH/CH multilayers contained a much lower polymer content than EH/PEI, revealing that EH/CH system has a higher capacity to interact with water.

Fig. 3

DSC curves over the entire temperature ramp (30–625 °C) for four selected PECs prototypes made of a EH:CH ratio of 1:2 (EC 1:2), a EH:CH:AG ratio of 1:1:2 (EC 1:1 + A), a EH:PEI ratio of 2:1 (EP 2:1), and a EH:PEI ratio of 9:1 (EP 9:1)

In the temperature range of 200–400 °C, polymer decomposition reactions occur. For all samples, broad exothermal peaks were observed in this region, with a visible peak temperature at 379 °C for EC 1:2 sample. These exothermal peaks are attributed to chitosan degradation, involving saccharide dehydration, depolymerization, and the decomposition of amino groups, as well as to the decomposition of sulphonic groups in EH that occurs with SO2 release. Peak temperatures for the decomposition of pure chitosan are usually reported between 279 and 311 °C [36,37,38,39,40,41,42], and the decomposition of sulfonic groups in lignosulfonate is reported around 350 °C [43]. Importantly, the higher peak temperature observed for EC 1:2 sample (379 °C) suggests the participation of both polymers in the complexation reaction. The absence of sharp decomposition peaks in the region 200–400 °C for EP 2:1, EP 9:1 and EC 1:1 + A samples, likely indicates that much stronger polymer interactions were present in these samples compared to EC 1:2 sample.

Surface morphology

SEM photographs for selected PECs made of EH and PEI or CH are presented in Fig. 4. PECs made of EH and PEI, i.e., EP 2:1 and EP 9:1 are shown in Fig. 4A-D. The matrix of EP 2:1 sample was composed of relatively homogeneous polyhedral-shaped spikes, having a face length of approximately 100 nm. On the other hand, EP 9:1 sample did not have any well-defined microscopic structure, likely due to the large excess of EH with respect to PEI.

Fig. 4

SEM photographs of dried PECs made of EH and PEI, i.e., EP 2:1 (A, B) and EP 9:1 (C, D); and made of EH and CH, i.e., EC 1:2 (E, F) and EC 1:1 + A (G, H). Scale bar indicates 10 μm in A, C, E, G and 1 μm in B, D, F, H

PECs composed of EH and CH assembled in the form of micro-sized flakes. More specifically, EC 1:2 (Fig. 4E-F) formed flakes of irregular shape, which fused into large microparticles with a diameter of approximately 5 μm. In contrast, flakes in the EC 1:1 + A sample (Fig. 4G-H) were organized into elongated microstructures with a length of approximately 1–10 μm. The morphological differences between the EC 1:2 and EC 1:1 + A samples are mainly attributed to the addition of agar, which facilitates the assembly of CH:EH flakes into elongated structures. Likely, CH:EH flakes assemble along the agar molecules, forming long microfibrils and an overall more homogeneous matrix compared to that of the agar-free sample, i.e., EC 1:2.

Stability of PECs

The stability data of the four selected PECs in pure water, acidic solution (pH 1.2) and PBS (pH 7.4) are summarized in Table 2. EP 2:1 was the most stable complex in all three fluids, exhibiting the lowest lignin leaching (3.1–8.3%) and weight loss (≤ 8.5%) compared to the other samples of the study. This minimal EH leaching is explained by the fact that the EH:PEI mass ratio of 2:1 used in this sample corresponds approximately to a 1:1 charge stoichiometry, which is the ratio under which insoluble PECs are usually formed, as explained in Section “Effect of polyelectrolyte ratio and temperature on complexation” In EP 2:1 samples, EH leaching seemed somewhat pH-dependent, showing the lowest value at pH 1.2, likely due to the high protonation degree of PEI at this pH. An increase of the EH:PEI ratio to 9:1 resulted in a 4-11-fold increase in EH leaching, reasonably due to the loss of unbound EH, that was used in large excess.

Table 2 Stability and EH leaching of PECs in pure water, at pH 1.2 (simulated gastric fluid) and pH 7.4 (simulated intestinal fluid) (± is the standard deviation, n = 3)

Both chitosan-containing complexes (EC 1:2 and EC 1:1 + A) had higher weight loss in all fluids compared to PEI-containing complexes, and the agar-free sample EC 1:2 showed the highest EH leaching of all samples of the study (80.1–95.6%). This confirms the low stability of PECs in the absence of agar due to the low polymer concentration used, as explained in Section “Effect of polyelectrolyte ratio and temperature on complexation”. In contrast, the addition of agar significantly reduced EH leaching (23.7–28.4% for EC 1:1 + A sample), likely due to the increase of the matrix viscosity which reduced EH diffusion. As observed for EP 2:1 sample, also for CH-containing samples, the acidic conditions reduced EH leaching, likely due to the high charge density of chitosan at pH 1.2. Weight loss values for EC 1:1 + A samples under the different pH conditions were higher (62.3–81.4%) compared to those found for EC 1:2 samples (35.4–69.7%), despite the lower EH leaching, which suggests a certain agar release during the washing. For all samples, when comparing weight loss values in PBS with those in water, values in PBS were systematically lower than those in water. This does not directly indicate a higher stability of PECs in PBS, since weight values used for the weight loss calculation refer to the total dry mass including the content of potentially absorbed salts.

Taken altogether, EH leaching depended on pH, PECs composition and matrix viscosity, being low pH, 1:1 charge ratio and high viscosity the most appropriate conditions to limit EH loss. For all pH conditions, the yield followed the trend EP 2:1 > EP 9:1 > EC 1:2 > EC 1:1 + A, indicating that PEI-containing samples were overall more stable than CH-containing samples and that a 1:1 charge stoichiometry positively affects the yield.

Loading and release of MB

Loading of MB in PECs made of EH and PEI or CH was performed by mixing MB in the EH solution to allow drug-polymer interactions prior to PECs formation. Indeed, MB is a positively charged, polycyclic aromatic compound (Fig. 5) that can interact with EH through several types of interactions. Firstly, the positive charges of MB can interact with the negative groups of EH (i.e., phenolic, carboxyl and sulphonic groups). Secondly, H-bonds can occur between the N atoms of MB and the hydroxyl groups of EH. Finally, π-π stacking interactions can be established between the several aromatic rings of EH and those of MB (Fig. 5) [27]. As a consequence of these interactions, the values of loading efficiency were high for all four formulations (89.6–97.9%). The slightly higher encapsulation efficiency of EC 1:1 + A sample compared to EC 1:2 is likely attributed to the higher amount of negatively charged polymers available for ionic interactions with positively charged MB. Indeed, EC 1:1 + A sample contained a higher amount of EH, as well as a slightly negative agar, which was not present in EC 1:2 sample. Moreover, compared to EC 1:2 sample, EC 1:1 + A sample had a much lower EH leaching which provided a higher amount of EH available for interactions with MB. The MB loading of PECs varied between 3.2 and 14.9%, based on the different PECs compositions (Table 3). Considering that the same amount of MB was used for all formulations during loading and that this amount was almost equally entrapped in terms of efficiency, the difference in the MB loading derives from the fact that different formulations have different final masses of loaded PECs, due to different starting polymer concentrations and different weight losses. In line with this, PEI formulations had lower MB loading values than CH formulations due to their higher starting polymer concentration and lower weight loss. Weight loss data reported in Table 2 correlates well with the MB loading data reported in Table 3, as samples that had the lowest weight loss had the lowest MB loading due to the highest mass of loaded PECs. Weight loss increases with the trend EP 2:1 < EP 9:1 < EC 1:2 < EC 1:1 + A (Table 2), which is the same trend followed by the increase of the MB loading (Table 3).

Fig. 5

Schematic illustration of some of the possible interactions between methylene blue and EH hemicellulose-rich lignosulphonate

Table 3 Values of loading efficiency (LE%) and percent loading (L %) of MB, calculated according to Eqs. 1 and 2 for PECs composed of EH and PEI or CH (+ agar)

Release curves of all MB-loaded PECs during the first 30 h under the two pH conditions tested (simulated gastric and intestinal fluids) are shown in Fig. 6A and B (cumulative and distributive release expressed as mg of released MB out of the dry mass of PECs), whereas curves representing the cumulative release as a function of the square root of time are presented in Fig. 6C (pH 1.2) and 6D (pH 7.4). Cumulative release expressed as % of MB released over the total MB entrapped is reported in Figure S2 (Supplementary Data). Release curves of all agar-free samples were characterized by an initial burst release phase during the first 3.5 h when the gastric transit at pH 1.2 was simulated, followed by a much slower release kinetics that reached a plateau during the subsequent hours when pH was kept at 7.4 (Fig. 6A). In contrast, EC 1:1 + A continuously released MB during the entire time frame. More in detail, for this formulation, two different burst phases were visible, i.e., an initial one at pH 1.2 and a second one corresponding to the pH change (Fig. 6B). During the first 3.5 h, the cumulative release of all PECs scaled linearly with the square root of time, indicating a first-order release kinetics (Fig. 6C). By comparing the slopes of the fitting lines, it is clear that the release rate decreases with the trend EC 1:1 + A > EC 1:2 > EP 2:1 > EP 9:1 (corresponding slope values 0.96, 0.61, 0.16 and 0.05). During the second phase at pH 7.4, all agar-free samples had a very low release rate (Fig. 6D, slope values 0.01–0.07). In contrast, EC 1:1 + A showed a similar release rate to that shown during the acidic phase (Fig. 6D, slope 0.95 vs. 0.96), which highlights the ability of this formulation to release MB in a sustained fashion. Agar is a well-known hydrogel-forming component, and therefore, in EC 1:1 + A sample, it likely contributed to a much higher water uptake compared to the other samples, which in turn facilitated MB diffusion through the polymeric matrix [30]. Remarkably, this sample had the highest MB loading and the greatest amount of MB released over time, which makes it a more promising prototype compared to the others. Taken altogether, in EC 1:1 + A sample, AG acted as a viscosity-enhancer leading to three main effects: (i) acquisition of the gel-like behaviour, (ii) reduction of the EH leaching (Table 2), (iii) increased and sustained MB release. This highlights that AG has an important role. However, it is noted that while AG is needed to boost the release, EH remains a key-component for the PECs formation and the drug retention and loading. In fact, given the chemical structures of AG and EH, as well as the establishment of interactions with the drug, it is reasonable that AG participates in drug retention only to a minor extent compared to EH. Indeed, EH is highly negative and has aromatic rings which are responsible for the retention of positively charged drugs with aromatic rings, like MB, due to the establishment of electrostatic and π-π stacking interactions. In contrast, AG is composed of a mixture of agarose and agaropectin. The first is a neutral linear polymer consisting of repeating units of D-galactose and 3,6-anhydro-L-galactose, whereas the latter is a more heterogeneous and branched fraction of agar, which contains additional substituents, such as negatively charged sulphate and pyruvate groups. However, since agaropectin accounts for only a small fraction of AG, the overall effect is a week negative charge. Given that AG is only slightly negative and does not have aromatic rings in its chemical structure, it can participate in the drug retention only to a minor extent. It can be hypothesized that AG participates in drug retention also by reducing EH leaching and, hence, by providing more EH available for the interactions with the drug. All these observations lead to the conclusion that agar is an important component for the stabilization of PECs and drug release, whereas EH ensures PECs formation and strong interactions with the drug. It follows that the release kinetics is a result of the influence of both components on the entire system. Comparing the two PECs made of EH and PEI, EP 9:1 sample which had the highest loading, resulted in the lowest MB release over time, which indicates that strong interactions are established between MB and the large excess of EH used in this sample. Taken altogether, we defined the conditions to formulate drug-loaded PECs based on a novel hemicellulose-rich lignin and able to control the release of a model drug in the gastrointestinal tract. Our findings are relevant to the current research trends on the topic, since the gastrointestinal controlled release of numerous drugs, which are currently administered intravenously, is still an unmet challenge. An example of this active research subject is the development of novel oral formulations of paclitaxel against colon cancer, where important challenges are related to paclitaxel high lipophilicity, efflux, and affinity for intestinal and liver degrading enzymes [44,45,46,47,48,49]. Our future studies will focus on the encapsulation in these novel PECs of anticancer drugs such as paclitaxel, which will benefit from π-π stacking interactions similar to those between EH and methylene blue, exploited in this study. In the context of anti-cancer therapy, the high potential of the proposed drug delivery system should also be considered because of the anti-oxidant effect of lignin. Lignin has an intrinsic scavenging effect against reactive oxygen species (ROS), and lignin fibers can adsorb carcinogens and toxins in the colon [50, 51]. This ability has been associated with a reduction of colon cancer relapse that is mainly linked to oxidative stress. This opens up new possibilities for the development of lignin-based drug delivery systems that may synergically improve the effect of chemotherapeutics.

Fig. 6

MB release kinetics from PECs. Cumulative (A) and distributive (B) release of MB (mg/g of dry PEC). Cumulative release as a function of the square root of time at pH 1.2 (C) and pH 7.4 (D)