Zeta potential changing nanoemulsions based on phosphate moiety cleavage of a PEGylated surfactant

The aim of this study was to develop and evaluate a zeta potential changing nanoemulsion (NE) containing polyoxyethylene (9) nonylphenol monophosphate ester (PNPP) as emulsi ﬁ er. 2% (v/v) of PNPP and 0.18% (5 μ M) of cetyltrimethylammonium bromide (CTAB) were incorporated into the li- pophilicNEpreconcentrate(PEG-40castoroil,glyceryltricaprylate/tricaprate,propyleneglycoldicaprylocaprate, propyleneglycolmonolaurate,highlypuri ﬁ eddiethyleneglycolmonoethylether;20/20/10/30/20,v/v).Afterdi-lutingthelipophilicpreconcentrate,resultingNEwasanalyzedregardingdropletsize,polydispersityindex(PDI), storagestabilityandzetapotentialchangeafterincubationwithintestinalalkalinephosphatase(IAP).Phosphate releaseduetocleavagebyIAPwasquanti ﬁ edbymalachitegreen assay and toxicity aswellascellular uptake be- havioronCaco-2cellswasdetermined.TheNEcontainingPNPPandCTABdisplayedadropletsizeof113nmand aPDIof0.20.Itwasstableover4h.Azetapotentialchangefrom − 33.7to+8.2mVwasobservedduetocleavage of phosphate groups by isolated IAP. PNPP could be identi ﬁ ed as non-toxic up to concentrations of 0.01% (v/v). Furthermore, an enhanced cellular uptake of the NE changing its zeta potential on Caco-2 cells was determined. Therefore, PNPP seems to be a promising tool for the development of zeta potential changing NE. © 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http:// creativecommons.org/licenses/by/4.0/).


Introduction
Nanoemulsions (NE) are isotropic dispersions of two non-miscible liquids with droplet sizes in the nanometric range. In case of o/w NE an oily phase is dispersed in an aqueous medium and stabilized by appropriate surfactants. They can be prepared by "high-energy" techniques such as ultrasonication and high-pressure homogenization or alternatively be designed as self-emulsifying systems. As the oily droplets showed high potential as nanocarriers for oral drug delivery, the pharmaceutical industry discovered NE as useful formulation tool [1][2][3]. However, upon oral administration NE have to overcome several barriers before facilitating systemic absorption of encapsulated drugs [4]. The efficiency of NE to take these hurdles strongly depends on their zeta potential. To achieve enhanced mucus permeation properties a negative zeta potential is favored [5,6], as the mucus contains large amounts of mucins with anionic sialic-and sulfonic acid moieties immobilizing positively charged carriers due to electrostatic attraction [7,8]. After crossing the mucus barrier, a positive zeta potential is preferred to enhance cellular uptake via ionic attraction of these nanocarriers with the negatively charged cell surface [9][10][11]. Consequently, anionic nanocarriers with the ability to shift their zeta potential to positive once having reached the absorption membrane hold great promise as delivery systems.
Due to the incorporation of phosphorylated surfactants in NE a zeta potential change from negative to positive values can be achieved via cleavage of the phosphate moieties by the membrane bound enzyme intestinal alkaline phosphatase (IAP) [12][13][14][15]. For an efficient cleavage by IAP, however, the accessibility of the phosphate moieties on the surface of the oily droplets is crucial. As adequate emulsification properties and stability of NE in general require the addition of strong PEGylated emulsifiers, the formation of a PEG corona around the lipophilic core is evident. When the phosphate moiety is bound directly to the hydrophobic tail of the surfactant anchoring it in the oily core of the droplets, its arrangement at the oil-PEG interface and a consequently limited access for IAP due to the shielding effect of the PEG corona is likely. The resulting zeta potential change is minor. In order to overcome this PEG corona, a hydrophilic linker between the phosphate group and the lipophilic tail of the surfactant seems useful. Wolf et al. synthesized such a surfactant, namely N,N′-bis(polyoxyethylene) oleylamine bisphosphate (POAP) and achieved a zeta potential shift of Δ21.6 mV underlining the potential of such surfactants [16].
Although a proof-of-concept could be provided for this type of phosphorylated surfactant, its tertiary amine substructure turned out to be too toxic for further developments. Combining the required cationic moiety with the phosphate substructure on the same molecule was obviously a too ambitious concept. Furthermore, two phosphorylated PEG chains on the same surfactant might hinder the accessibility of IAP as well.
The purpose of this study was therefore to develop a zeta potential changing system based on a different phosphorylated surfactant exhibiting a PEG spacer. To achieve this goal, polyoxyethylene (9) nonylphenol monophosphate ester (PNPP), as illustrated in Fig. 1, was incorporated in a NE. The resulting formulation should demonstrate a reduced cytotoxicity compared to previous systems including substances such as POAP due to the lack of a positive charged molecule moiety. Furthermore, the single-chain hydrophilic PEG linker of PNPP might enable a more efficient phosphate cleavage by IAP leading to a higher zeta potential change in comparison to already established systems.

Purification of PNPP
In order to remove residues of phosphoric acid contaminating PNPP flash column chromatography (silica gel 60, 400-220 mesh, Carl Roth, Germany) was conducted. Therefore, a solvent mixture consisting of isopropanol, 25% ammonia solution and water was utilized (7/2/1, v/ v/v) and subsequently evaporated under vacuum. Thereafter, precipitated silica was removed via centrifugation giving PNPP as a viscous yellowish liquid with a yield of 76%. All further experiments were performed with this purified PNPP.

Preparation of nanoemulsions
The lipophilic phase of NE without PNPP (LPNE Blank ) was prepared by homogenizing components as listed in Table 1. For LPNE containing PNPP (LPNE PNPP ), 2% (v/v) of PEG-40 castor oil were replaced with PNPP. Furthermore, the cationic surfactants cetylpyridinium chloride (CPC), cetyltrimethylammonium bromide (CTAB), n-dodecyltrimethylammonium bromide (DTAB) as well as the esterquat bis-(isostearoyl/oleoyl isopropyl) dimonium methosulfate (EQ) were incorporated in a concentration of 5 μM into LPNE PNPP by vortexing and treatment with ultrasound until a homogenous phase was obtained. As CTAB turned out to be the most promising cationic surfactant, it was also tested in concentrations of 1 μM, 15 μM, 20 μM as well as 30 μM. As all formulations displayed self-emulsifying properties, emulsions were prepared by diluting the lipophilic phase 1:100 (v/ v) with demineralized water and shaking at 300 rpm and 37°C for 1 min. NE were incubated at 37°C under shaking at 300 rpm and characterized regarding droplet size after 0 as well as 4 h (NanoBrook 90Plus PALS, Brookhaven Instruments, USA).

Zeta potential change analysis
Zeta potential change was induced by the addition of 5 μL·mL −1 of intestinal alkaline phosphatase resulting in an enzymatic activity of 5 U·mL −1 to LPNE having been diluted 1:100 (v/v) with demineralized water. Thereafter, zeta potential of samples shaken at 300 rpm and 37°C was analyzed at 0 and 4 h via the particle analyzer applying a voltage of 4 V and a field frequency of 2 Hz. Emulsions of LPNE Blank serving as reference were prepared and investigated as described above. LPNE containing 5 μM of CTAB was additionally investigated at 1, 2 and 3 h.

Phosphate release studies
For phosphate release studies, a malachite green (MLG) assay was conducted based on an already described method [16,17]. In brief, 1.25 mL of an aqueous ammonium molybdate tetrahydrate solution (7.5%, w/v), 0.1 mL of Tween 20 (11%, v/v) in water as well as 5 mL of a malachite green oxalate solution (0.12%, w/v) prepared with 2.25 M sulfuric acid were combined in order to obtain the MLG reagent. Afterwards, the amount of inorganic phosphate within samples was determined by mixing 20 μL of 1 M sulfuric acid, 180 μL of test sample as well as 50 μL of the MLG reagent in a 96-well plate and subsequent measurement of the absorbance at 620 nm after shaking at 150 rpm for 10 min. For calculating inorganic phosphate concentrations, a calibration curve (R 2 = 0.997) was prepared within a concentration range of 0.5 to 50 μM using an emulsion prepared with LPNE Blank as dilution medium to address the influence of surfactants on the analysis [18].  2.4.1. Phosphate release studies catalyzed by isolated phosphatase As LPNE PNPP containing 5 μM of CTAB (LPNE PNPP+CTAB ) turned out to be the most promising formulation after zeta potential analysis, the released phosphate after enzyme addition was investigated. Therefore, 5 μL·mL −1 of IAP with an enzymatic activity of 0.5 U·mL −1 were added to LPNE PNPP+CTAB having been diluted 1:1000 (v/v) with demineralized water. While incubating the NE at 37°C and 300 rpm, aliquots of 180 μL were withdrawn at 0, 1, 2, 3 as well as 4 h and analyzed via MLG assay. LPNE Blank having been treated in the same way as LPNE PNPP served as reference.

Caco-2 cells induced phosphate release
Phosphate release of LPNE Blank as well as LPNE PNPP+CTAB was investigated on Caco-2 cells as they represent an intestinal cell line expressing IAP [19,20]. Cells were cultured in Eagles minimum essential medium (MEM) supplemented with 10% fetal bovine serum as well as 1% penicillin-streptomycin after seeding them in 24-well plates with a density of 2 × 10 5 per well. Cells were incubated at 90% relative humidity, 5% CO 2 as well as 37°C and MEM was changed every 48 h. After 14 days cell monolayers were washed trice with HEPES buffered saline pH 7.4 (HBS, 20 mM HEPES, 5.5 mM glucose, 137 mM NaCl, 5 mM KCl, 1 mM CaCl 2 ) and treated with 500 μL of HBS containing 1% (v/v) Phosphatase Inhibitor Cocktail 2 (PIC) or 500 μL of pure HBS for 30 min. Then, 1.1 mL of LPNE PNPP+CTAB diluted 1:1000 (v/v) with HBS or with HBS containing 1% (v/v) PIC were transferred into each well, whereby HBS and HBS with 1% (v/v) PIC served as references. While incubating cells at 37°C, aliquots of 180 μL were taken at predetermined time points and immediately mixed with 20 μL of 1 M sulfuric acid to inhibit any further enzymatic activity. Afterwards, the amount of released phosphate was determined by MLG assay.

Cytotoxicity studies on Caco-2 cells
In order to investigate the impact of NE prepared from LPNE Blank as well as LPNE PNPP+CTAB or plain PNPP on viability of Caco-2 cells, a resazurin assay was conducted [21]. Cells having been cultured in 24well plates as described above were washed twice with HBS. 500 μL of LPNE Blank and LPNE PNPP+CTAB both having been diluted 1:1000 (v/v) with HBS or PNPP dissolved in HBS in a concentration range from 0.002% to 0.02% (v/v) were transferred into the wells. As positive control HBS and as negative control 2% (v/v) Triton X100 in HBS were employed. Following an incubation period of 4 h, cells were washed twice with HBS. Afterwards, 300 μL of a 44 mM resazurin solution prepared with MEM were added to each well and after a reaction time of 3 h at 37°C the fluorescence intensity of the supernatants was analyzed via a microplate reader (Spark®, Tecan Trading AG, Switzerland) using an excitation wavelength of 540 nm and an emission wavelength of 590 nm. Subsequently, cell viability was calculated based on the following equation: Cellular uptake of emulsified LPNE PNPP+CTAB with and without phosphate cleavage was investigated by confocal laser scanning microscopy (Leica TCS SP8). Caco-2 cells were seeded in an 8-well plate and used for experiments after confluency was attained. Cells were incubated for 30 min with OptiMEM (OM) or OM containing 1% (v/v) PIC. In the following, cells were treated for 2 h with LPNE PNPP+CTAB labelled with 0.1% (w/w) Lumogen® F yellow 083 and diluted 1:10,000 (v/v) with OM or with OM containing 1% (v/v) PIC. Subsequently, cells were washed with preheated OM thrice and nuclei were stained utilizing Hoechst 33528 (1 μg·mL −1 ). Equal confocal settings were applied for recording all fluorescence images and image post-processing was conducted by ImageJ whereby the yz-and xz-projections were obtained from 5 xy-images of an image stack taken at 0.2 μm z-step length. Furthermore, spectral unmixing was used in order to erase fluorescence bleed-through between detection channels and 2D image filtering was applied via a Gaussian filter.

Statistical data analysis
Results are presented as means ± SD. One-way ANOVA followed by Tukey post hoc test with a level of significance of 0.05 was utilized for statistical data evaluation (IBM SPSS Statistics, version 22).

Evaluation without cationic surfactants
As depicted in Table 2, NE displayed stable droplet sizes over 4 h and polydispersity indexes (PDIs) ≤ 0.2 indicating monodisperse systems. For NE with PNPP (NE PNPP ), however, a larger droplet size of about 40 nm compared to NE without PNPP (NE Blank ) was found. This increase might be explained by a different emulsification behavior of LPNE PNPP compared to LPNE Blank , as 2% (v/v) of the surfactant PEG-40 castor oil were replaced with PNPP. Regarding zeta potential profiles of both formulations upon incubation with IAP, a strong increase of Δ33.4 mV was observed for NE PNPP whereas NE Blank showed no significant change within 4 h ( Table 2). The measured shift amplitude was even more than 1.5-fold higher compared to the so far most pronounced change in zeta potential reported in the literature [16] demonstrating the potential of PNPP.

Evaluation of different cationic surfactants
Despite this change in zeta potential, the formulation NE PNPP still exhibited a slightly negative potential of −6.7 mV after 4 h. Thus, in order to reach a positive zeta potential, different cationic surfactants were incorporated in LPNE PNPP in a concentration of 5 μM for comparison reasons. In case of all surfactants, a minor decrease in droplet size in a range of Δ10 to 20 nm in relation to NE PNPP was analyzed, as shown in Table 3. Furthermore, droplet size remained stable for 4 h and PDIs were below 0.2 for all tested formulations. Accordingly, the emulsification characteristics of NE were altered only to a minor extent by incorporating various cationic surfactants.
Comparing the IAP-mediated zeta potential change, no statistical difference between CPC, CTAB and DTAB could be found, whereas this change was less pronounced in case of EQ. As the relative highest shift in zeta potential with Δ41.9 mV and the significant (p ≤ 0.05) most positive zeta potential of +8.2 mV was achieved with CTAB, it was used for more detailed studies.

Evaluation of different concentrations of cationic surfactant
CTAB was incorporated into LPNE PNPP in various concentrations to investigate the influence on droplet size and zeta potential change. With increasing concentrations of CTAB droplet size decreased likely caused by a concentration dependent influence of the cationic surfactant on the emulsification characteristics of LPNE PNPP . Correspondingly, the shift in zeta potential was attenuated, as for 1 μM CTAB the zeta potential increased by Δ46.5 mV, whereas a concentration of 30 μM resulted in a zeta potential change of only Δ13.2 mV (Table 4). An explanation for this decrease might be an inhibitory effect of CTAB in higher concentrations on IAP [22]. Nevertheless, usage of different CTAB concentrations facilitated a zeta potential adjustment and provides therefore flexibility for further developments with PNPP.
Selecting a CTAB concentration for following studies with LPNE PNPP was based on three aspects. First, the initial zeta potential should be negative to enable sufficient mucus permeation [5,23].
Second, the surface charge of droplets upon contact with IAP should shift to positive to support cellular uptake [9][10][11]. Third, the concentration of CTAB should be kept at a minimum, as cationic surfactants are well-known for their toxicity [24,25]. Consequently, LPNE PNPP containing 5 μM of CTAB (LPNE PNPP+CTAB ) was chosen for further experiments, as this formulation displayed highly negative charged droplets (−33.7 mV) and a positive zeta potential (+8.2 mV) upon incubation with IAP.

Zeta potential change over time
In the following zeta potential change of LPNE PNPP+CTAB over time was evaluated during incubation with IAP. As illustrated in Fig. 2 (left axis, ♦), a fast increase in zeta potential for LPNE PNPP +CTAB was observed within the first hour. The process then slowed down most likely caused by a reduced amount of substrate available for enzymatic cleavage. After 4 h, an overall shift in zeta potential of Δ41.9 mV was determined for LPNE PNPP+CTAB . In comparison to already established zeta potential changing NE based on different phosphorylated surfactants as listed in Table 5, the accomplished shift in zeta potential was nearly twice as high as the highest so far achieved zeta potential shift for NE highlighting the great potential of PNPP.

Phosphate release studies catalyzed by isolated enzyme
In order to confirm that the zeta potential change of NE PNPP+CTAB is caused by and correlates with the cleavage of phosphate groups from PNPP, phosphate release of NE PNPP+CTAB was investigated over 4 h. Within the first hour, as shown in Fig. 2 (right axis ○), 18.7 μM of Table 4 Droplet size, polydispersity index (PDI) as well as zeta potential after 0 and 4 h of nanoemulsions with PNPP containing different amounts of cetyltrimethylammonium bromide (CTAB). Emulsions were prepared by diluting lipophilic preconcentrates 1:100 (v/v) with demineralized water and adding 5 U•ml −1 of intestinal alkaline phosphate. Indicated values are means (n = 3) ± SD.   phosphate was released corresponding to 64% of the overall phosphate release of 29.3 μM after 4 h. As these outcomes were in good accordance with the zeta potential change of NE PNPP+CTAB , it can be concluded that the shift in zeta potential is caused by cleavage of the phosphate group on PNPP by IAP. Overall, 92% of phosphate having been added to the formulation was cleaved off explaining why a higher zeta potential change could be achieved for NE with PNPP in comparison to POAP, as for this surfactant only 50% of phosphate moieties were cleaved off within the same period [16]. The higher phosphate release and corresponding zeta potential change determined for NE PNPP+CTAB might be explained by the PEG spacer of PNPP. It seemed due this hydrophilic linker the incorporated phosphate groups were located outside the PEG corona resulting in an improved accessibility of IAP in comparison to substance like phosphorylated tyrosine-octadecylamide synthesized by Salimi et al. [13]. Comparing PNPP to POAP, the double-tail structure of POAP could cause a steric hindrance of IAP and therefore a more pronounced phosphate release and zeta potential change was determined for NE containing the single-chain PNPP [16].

Caco-2 cell induced phosphate release
Results of phosphate release studies conducted with NE PNPP+CTAB on Caco-2 cells are illustrated in Fig. 3. With a concentration of 48.7 μM after 4 h, a pronounced phosphate release from NE PNPP+CTAB was observed. This value was significantly higher (p ≤ 0.001) compared to 24.9 μM released from the NE PNPP+CTAB in presence of PIC. However, the released amount of phosphate determined for HBS (12.5 μM) and HBS with PIC (14.3 μM) was lower than the phosphate concentration analyzed for NE PNPP+CTAB in presence of PIC.
A likely explanation for the difference between these control samples is a phosphate cleavage in case of NE PNPP+CTAB applied with PIC by unspecific esterases being not or only partially inhibited by PIC. Furthermore, phosphate release from perishing cells caused by a time dependent cytotoxicity of NE PNPP+CTAB might also be responsible for this observation [28]. Nevertheless, NE PNPP+CTAB without PIC released 23.8 μM more phosphate compared to the reference NE PNPP+CTAB with PIC. Accordingly, upon oral administration a sufficient phosphate release at the intestinal epithelium can be anticipated providing the intended change in zeta potential. Table 5 Overview of zeta potential changing nanoemulsion systems based on phosphate bearing compounds. Intestinal alkaline phosphatase (IAP).

Phosphate bearing compound
Chemical structure Zeta potential values before and after incubation with IAP (mV) Maximum achieved change in zeta potential (ΔmV) Ref.

Cell viability studies on Caco-2 cells
A resazurin assay was utilized to evaluate the impact of NE with and without PNPP+CTAB as well as unmodified PNPP on Caco-2 cell viability. As illustrated in Fig. 4, NE PNPP+CTAB was significantly (p ≤ 0.05) more toxic in comparison to NE Blank likely caused by the incorporated cationic surfactant CTAB that is known for its cytotoxic potential [29]. Nevertheless, with a cell viability of 70.0%, NE PNPP+CTAB still demonstrated a low cytotoxic potential [30]. However, in comparison to other zeta potential changing systems [13,14,16,27], the cationic charge is not located on the same surfactant from that the phosphate is cleaved off. Thus, cationic surfactants displaying a lower cytotoxicity compared to CTAB, such as in particular biodegradable surfactants based on arginine [31] or lysine [32] could be utilized as alternatives for zeta potential changing systems based on PNPP.
Moreover, unformulated PNPP displayed no cytotoxicity up to concentrations of 0.01% (v/v) and at 0.02% (v/v) cell viability still remained at 54.3% allowing the application of higher amounts of PNPP in formulations than having been used within this study [30]. Furthermore, PNPP possessed enhanced cell compatibility in comparison to POAP, which exhibited a pronounced cytotoxic effect already at a concentration of 0.01% (v/v) and provoked almost complete die-off at 0.02% (v/v) likely caused by its cationic substructure [16].

Cellular uptake studies on Caco-2 cells
Confocal imaging was applied to evaluate the cellular uptake of NE PNPP+CTAB including a fluorescent dye on Caco-2 cells. Fig. 5 presents a top view (A), an enlarged top view (B) and a side view (C) of the cells after treatment with NE PNPP+CTAB . According to these images, a considerable higher cellular uptake for NE PNPP+CTAB was found when applied without PIC. Additionally, C proves that the dye was taken up by cells and not attached on their surface. After cleavage of phosphate groups of PNPP by IAP expressed on Caco-2 cells, NE droplets should demonstrate a positive zeta potential as shown in Section 3.2 [19,20]. As the cell membrane exhibits a dense negative charge resulting from components such as amino acids with negatively charged side chains, anionic phospholipids and heparan sulfates, positively charged carriers are able to interact more tightly with the cell surface. Thus, cell membranes are depolarized leading to an enhanced cellular uptake of these carriers. Therefore, the obtained results are in good accordance with outcomes of previous studies [6,[9][10][11][33][34][35][36]. For example, Nazir et al. recently described a 2-fold enhanced uptake for zeta potential changing NE on Caco-2 cells in the absence of IAP inhibitors [27].

Conclusion
Zeta potential changing NE based on phosphate moiety cleavage seem to be a promising approach to improve the oral bioavailability of active pharmaceutical ingredients. In order to further improve their potency, the single-chain phosphorylated surfactant PNPP with a hydrophilic PEG linker was evaluated to address the issue of steric enzyme hindrance caused by a PEG corona around the oily droplet core. By incorporating PNPP and the cationic surfactant CTAB, a NE with a distinctly improved zeta potential change from negative to positive in comparison to previous systems was generated. Furthermore, a more efficient cleavage of phosphate moieties was achieved and an enhanced  Based on these results, PNPP is the so far most promising tool for developing zeta potential changing NE containing phosphorylated substances. Therefore, follow-up studies investigating the properties of such a system upon encapsulation of suitable drugs would be of great interest in the future.

Funding
Funding for this study was received by the Österreichischen Forschungsförderungsgesellschaft (FFG, Austria, grant number 7677912/856696).