Protein encapsulation within poly-lactide-co-glycolide (PLGA) microspheres is a simple and straightforward procedure. To date, empirical approaches have been employed for protein stabilization during the first step of the double emulsion process (W₁/O/W₂).

In this step high pressures, temperature gradients and shear forces enhance the probability of oxidations to occur, leading to free radicals production and protein damages. When shear forces are considered, their net result is an increase in the organic solvent surface area which triggers protein adsorptions, aggregate precipitations and conformational damages [1,2,3,4,5,6,7,8]. BSA addition to diphtheria toxoid (Dtxd) or to tetanus toxoid (Ttxd) [9,10] was proposed to avoid antigen aggregation. The underlying supposition is that BSA competes with the antigen for the adsorption at the CH₂Cl₂/H₂O interface. Evidently, this can occur. However, it is not reasonable to add heterologous proteins to pharmaceutical formulations intended for human usage. Another proposition was to increase the protein concentration, maintaining the same interfacial CH₂Cl₂/H₂O area. Good results were obtained for carbonic anhydrase and BSA [11], lysozyme [12], and human growth hormone [13]. With insulin [14] the increase in concentration induces gel formation that leads to peptide stabilization. Other ideas were the co-dissolution of PLGA with non ionic surfactants [1,15,16,17,18,19,20,21] and protein chemical modifications with polyethylene glycol (PEG) [22]. It is known, from literature, that the Hoffmeister ions salts are able to change proteins solubilities (both, in water or at interface). Based on these facts it was decided to used them as a simple solution to avoid protein precipitations, aggregations or conformational modifications during the first phase (interfacial contact with CH₂Cl₂) of the W/O/W PLGA microsphere preparation. The salt can easily be removed after microsphere precipitation and washing. We discuss the mechanisms by which protein conformation is maintained and damages are avoided.

The Dtxd solubility was enormously affected by the emulsification both in the presence of CH₂Cl₂ and in the absence of salts (Figure 1), when 33% of the molecule precipitated. The kosmotropics NaH₂PO₄ and NaCl, practically did not alter protein recovery in all concentrations studied. It was interesting to note that the chaotropic MgCl₂ precipitated at least 80% of Dtxd in all concentrations used. In contrast, the other chaotropic, KSCN solubilized Dtxd and 95% of the protein was recovered after emulsification in the presence of the organic solvent (Figure 1, 30 – 70 mM KSCN).

In order to know the nature and extent of these Dtxd losses, the soluble Dtxd fractions after emulsification in the presence of salts were applied on a gel filtration HPLC column (Figure 2). The dimer (2 x Dtxd) concentration was practically the same in all salt concentrations used (Figure 2A). There is a large reduction in Dtxd monomers when the emulsification was done in the presence of MgCl₂ (Figure 2B). The increase in FA and FB (Fragments A and B) occurred in concert with the decrease of Dtxd monomers when Dtxd was emulsified in the presence of SCN^-, suggesting a significant Dtxd monomer hydrolysis. The decrease in monomers in the presence of Mg²⁺ and PO₄^2- (Figure 2B) did not coincide with the dimer, FA or FB increases (Figure 2A). Probably, in these situations insoluble aggregates (containing more than 2 Dtxd monomers) could be formed, since the Dtxd dimer was still soluble (Figure 2A).

The lesser immunological identity of these fractions was observed when the emulsification was done in the presence of MgCl₂ (Figure 3). The other salts, practically, did not alter the immunological identity of the soluble Dtxd fractions (Figure 3).

The extent of these conformational damages was analyzed by circular dichroism (Figure 4). Changes in the α−helical content were observed at θ_222nm (Figure 4) and for S-S dihedral angle conformations, at θ_260nm (Figure 5). The α−helical content in the presence of NaH₂PO₄ increased in all the concentrations studied (Figure 4). Oscillations on θ_222nm in the lower NaCl concentration were observed. The biggest negative Cotton effect was observed in the Dtxd emulsified in the presence of MgCl₂. In the presence of KSCN the Dtxd α−helical content remained stable (Figure 4).

To facilitate the discussion, a schematic representation of Dtxd is shown in Scheme 1. It is important to note that diphtheria toxin has two S-S bridges (C₁₈₆-C₂₀₁ and C₄₆₁-C₄₇₁). One of them is an inter-peptide linkage (FA-FB, C₁₈₆-C₂₀₁) and the other is an intra-peptide linkage (FB, C₄₆₁-C₄₇₁). The dihedral S-S angle (linkage between the FA and FB) suffered conformational alterations during the simulation of the first emulsification during the microsphere preparation method (Figure 5). Probably, the observed dihedral changes are related with the C₁₈₆-C₂₀₁ bridge because of the steric hindrance and constraints on the other S-S linkage (C₄₆₁-C₄₇₁). It is known that angular dihedral variations ≥ 120° are related with a negative Cotton effect. In contrast, angular dihedral variations ≤ 60° induce positive Cotton effects [25,26].

The lowest S-S dihedral conformational change was observed when Dtxd was emulsified in the presence of NaCl (Figure 5). From 10-40 mM NaH₂PO₄ the S-S dihedral change was ≥ 120° and above 50 mM NaH₂PO₄ remained ≤ 60°. The biggest Dtxd S-S variations were observed in the presence of MgCl₂. In the presence of MgCl₂, the Dtxd opened its structure (the S-S changed from ≤ 60°, a positive Cotton effect, to ≥ 120° a negative Cotton effect). Between 50 and 70 mM KSCN, the S-S dihedral Dtxd angle also changed its conformation, but the variations were close to zero and corresponded to a closed conformation.

It was observed in the fluorescence spectra that all MgCl₂ concentrations studied produced the highest increase in the F350 nm/F330 nm ratio (that corresponds to the degree of tryptophan, W, exposition to the media). The NaCl and the NaH₂PO₄ induced an intermediate W exposition. The lowest W exposition was observed in the presence of KSCN (Figure 6).

Taking all the previous results into consideration, Dtxd microencapsulations within MS-PLGA in the presence of KSCN were performed and under this condition, the encapsulation efficiency raised from 53.84% to 64.26 (when the microencapsulation of Dtxd was done within the end methylated PLGA of 12 kDa) and 23.2% (from 46.41% to 69.63%) within the free –COOH PLGA (12 kDa).

It was verified in this work that simple molecules like Hoffmeister salts could be the best solution to protect the Dtxd protein from aggregation and conformational damages during the emulsification with CH₂Cl₂. Salts are easier to remove than exogenous proteins. The problem we faced was solved because high quantities of Dtxd were solubilized by SCN^- and consequently increased the encapsulation efficiency.

The enhancement of Dtxd protein solubility in the presence of SCN^- contrasted with other literature results, where lisozyme was precipitated in the presence of this chaotropic [26]. When the solvent is changed (ethyl acetate, for example) the SCN^- decreased Dtxd solubility [27], meaning that the protein partition depends on several factors like dielectric constant.

The problem approached in this work was not simple neither trivial: a ternary system composed by the protein Dtxd, the water and the organic solvent where a fourth element would be added - the salt, under a 24,000 rpm agitation! It would be an oversimplification to opt for one of the theories [28,29,30,31] to explain the Hoffmeister effect on protein solubility. The protein solubility change [30,31] in the presence of kosmotropic salts which stabilize solutes, by increasing the order of water or in the presence of chaotropic salts which create weaker hydrogen bonding thus decreasing the order of water and increasing its surface tension. Nowadays it is known that the Hoffmeister series mechanism of action is related to ion specific phenomena [29]. But, changes on water structure after salt additions do not explain the ion specific phenomenon [28,32]. Here the chaotropics SCN^- and Mg²⁺ were used. In addition to their charge difference it is known that SCN^- is poorly hydrated and Mg²⁺ is strongly hydrated [28,32]. The Dtxd was in water, pH 6.5, and containing salt in a condition of 2 pH units above its pI [23] with a net negative charge. Therefore, it is reasonable to assume that an ionic interaction between Dtxd (negatively charged) and Mg²⁺ did not favor protein solubility. In this salt condition the protein could precipitate after organic solvent contact, maximized by the agitation at 24000 rpm. So, here, the Hoffmeister series would be understood as an ion specific phenomenon [28,33]. The kosmotropic salts, NaCl and NaH₂PO₄ did not alter the Dtxd aggregation caused by agitation. The aggregation caused by the presence of CH₂Cl₂ was not associated to the increase in Dtxd dimmer (the dimmer is soluble) but with the formation of large aggregates. Dtxd was not hydrolyzed in FA and FB. The Dtxd soluble fraction immunologically recognized was lost after emulsification in the presence of MgCl₂. This Dtxd not immunologically recognized conformation corresponded to a protein more folded and more exposed to the environment. This exposed protein showed its hydrophobic residues to the media. This fact corroborates the expressed hypothesis that Mg²⁺ interacted strongly with Dtxd. The SNC^- protected Dtxd from unfolding, S-S dihedral opening and immunological damages. The Dtxd solubility increased 95% in the presence of SNC^-, corresponding to a molecule in its native conformation. This technological gain was used to prepare a Dtxd/PLGA formulation in the presence of 30 mM SNC^-. The recovery increased in 10.42% (from 53.84% to 64.26%) the microencapsulation of Dtxd within the end methylated PLGA (12 kDa) and 23.2% (from 46.41% to 69.63%) within the free –COOH PLGA (12 kDa).

In conclusion, the aim of this work was achieved, which was to obtain the maximum protein solubility after emulsification in the presence of CH₂Cl₂ to begin its PLGA microencapsulation within ideal conditions. This was a technological breakthrough because a simple solution such as salt addition avoided the usage of heterologous protein.