Rhamnolipids as epithelial permeability enhancers for macromolecular therapeutics

Accepted Manuscript Rhamnolipids as epithelial permeability enhancers for macromolecular therapeutics Diego Romano Perinelli, Driton Vllasaliu, Giulia Bonacucina, Benedetta Come, Stefania Pucciarelli, Massimo Ricciutelli, Marco Cespi, Rosangela Itri, Francesco Spinozzi, Giovanni Filippo Palmieri, Luca Casettari PII: DOI: Reference:

S0939-6411(17)30355-7 http://dx.doi.org/10.1016/j.ejpb.2017.07.011 EJPB 12567

To appear in:

European Journal of Pharmaceutics and Biopharmaceutics

Received Date: Revised Date: Accepted Date:

16 March 2017 10 July 2017 20 July 2017

Please cite this article as: D. Romano Perinelli, D. Vllasaliu, G. Bonacucina, B. Come, S. Pucciarelli, M. Ricciutelli, M. Cespi, R. Itri, F. Spinozzi, G. Filippo Palmieri, L. Casettari, Rhamnolipids as epithelial permeability enhancers for macromolecular therapeutics, European Journal of Pharmaceutics and Biopharmaceutics (2017), doi: http:// dx.doi.org/10.1016/j.ejpb.2017.07.011

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Rhamnolipids as epithelial permeability enhancers for macromolecular therapeutics Diego Romano Perinellia, Driton Vllasaliub, Giulia Bonacucinaa, Benedetta Comec, Stefania Pucciarellid, Massimo Ricciutellia, Marco Cespia, Rosangela Itrie, Francesco Spinozzic, Giovanni Filippo Palmieria and Luca Casettarif* a

School of Pharmacy, University of Camerino, via Gentile III da Varano, 62032 Camerino, MC, Italy b School of Pharmacy, University of Lincoln, Green Lane, Lincoln, LN6 7DL, UK. c Department of Life and Environmental Science, Polytechnic University of Marche, Ancona, Italy d School of Biosciences and Veterinary medicine, University of Camerino, via Gentile III da Varano, 62032 Camerino, MC, Italy e Instituto de Física da Universidade de São Paulo, IFUSP, Rua do Matão, Travessa R, 187, 05508-090 São Paulo, Brazil. f Department of Biomolecular Sciences, School of Pharmacy, University of Urbino, Piazza del Rinascimento, 6, 61029 Urbino (PU), Italy

ABSTRACT The use of surfactants as drug permeability enhancers across epithelial barriers remains a challenge. Although many studies have been performed in this field using synthetic surfactants, the possibility of employing surfactants produced by bacteria (the so called biosurfactants”) has not been completely explored. Among them, one of the most well characterized class of biosurfactants are rhamnolipids. The aim of the study was to investigate the effect of rhamnolipids on the epithelial permeability of fluorescein isothiocyanate-labelled dextrans 4kDa and 10kDa (named FD4 and FD10, respectively) as model for macromolecular drugs, across Caco-2 and Calu-3 monolayers. These cell lines were selected as an in vitro model for the oral and respiratory administration of drugs. Before performing permeability studies, rhamnolipids mixture was analysed in terms of chemical composition and quantification through mass analysis and HPLC. Cytotoxicity and transepithelial electrical resistance (TEER) studies were also conducted using Caco-2 and Calu-3 cell lines. A dose-dependent effect of rhamnolipids on TEER and FD4 or FD10 permeability across both cell lines was observed at relatively safe concentrations. Overall, results suggest the possibility of using rhamnolipids as absorption enhancers for macromolecular drugs through a reversible tight junction opening (paracellular route), despite more investigations are required to confirm their mechanism of action in term of permeability. Keywords: biosurfactants; Caco-2; Calu-3, cytotoxicity; TEER. 1

INTRODUCTION The use of biosurfactants, as well as surfactants from renewable sources, for drug delivery applications has become increasingly appealing in recent years [1,2]. One such class of surfactants is that of rhamnolipids. These have attracted significant attention as promising surface-active excipients for cosmetic and pharmaceutical applications, owing to their good emulsification, foaming and wetting properties, as well as their excellent surface activity [3]. Rhamnolipids are biosurfactants belonging to the class of glycolipids, made up of one (mono-rhamnolipids) or two (di-rhamnolipids) rhamnose moieties, linked to a large variety of 3-(hydroxyalkanoyloxy)alkanoic acids (generally from C8 to C16 carbon chain). They are predominantly produced from different strains of Pseudomonas aeruginosa by a fermentation process, which gives rise, after purification, to a mixture of mono-and di-rhamnolipids with different lengths of the hydrophobic tails. As with other biosurfactants, rhamnolipids offer advantages over known synthetic surfactants, including a potentially low toxicity and high level of environmental biodegradability, in addition to favourable intrinsic biological properties such as antimicrobial activity against several Gram positive and Gram negative bacteria, and fungi [4–6]. Rhamnolipids have already been investigated as an alternative to synthetic surfactants in several pharmaceutical formulations, including microemulsions [7], nanoemulsions [8] or topical formulations administered to the skin [9]. A key application of surfactants in the pharmaceutical field is their ability to act as drug absorption enhancers. Therapeutically acceptable absorption of therapeutic biomacromolecules such as proteins and peptides across the mucosal surfaces of the intestinal and the respiratory system remains a challenge and injection-mediated administration remains the default option for these therapeutics. Several classes of synthetic surfactants have been shown to increase mucosal drug permeability in vitro and in vivo, including polysorbates [10,11], sodium dodecyl sulphate [12] sodium caprate [13,14], through different selective or non-specific mechanisms [15]. The interaction of surfactants with the epithelial tissue is thought to modulate the barrier property, e.g. through a tight junction opening mechanism. However, key to the safe and effective use of these

2

materials is an acceptable toxicity profile, which has often been raised as an issue with the use of some surfactants. Research to date demonstrates that rhamnolipids possess a favourable toxicological profile [16,17] and, being secreted from Pseudomonas aeruginosa as virulence factors to promote the infiltration across epithelia [18], could display an improved performance (at optimal doses) as permeability enhancers in comparison to synthetic surfactants. Rhamnolipids have previously been evaluated as permeability enhancers [19,20]. Specifically, the effect of different concentrations of rhamnolipid mixtures on the permeability of small molecules such as phenol red, propranolol [19] and [14C] mannitol across Caco-2 (intestinal epithelial) monolayers has been investigated [20]. However, it is not known (not reported in the literature) whether these materials can enhance the epithelial permeability of macromolecules. The aim of this study is to investigate the effect of a rhamnolipid mixture on the epithelial permeability of macromolecules. Fluorescently-labelled dextrans of two different molecular weights, namely 4kDa and 10kDa (FD4 and FD10), were used as model macromolecular drugs. The commercial mixture of rhamnolipids was initially analysed to identify, through mass analysis and high-performance liquid chromatography (HPLC), the main components. This was followed by assessment of rhamnolipid effects on cytotoxicity, transepithelial electrical resistance (TEER) and permeability in Caco-2 and Calu-3 polarised monolayers as models for the intestinal and airway epithelia, respectively.

MATERIALS Rhamnolipids from Pseudomonas aeruginosa (90% purity) were purchased from Sigma-Aldrich (Poole, UK). Caco-2 and Calu-3 cells were obtained from the American Type Culture Collection (ATCC, Manassas, Virginia) and used at passages 32–36 for Calu-3 and 62-68 for Caco-2. Hank’s Balanced Salt Solution (HBSS) with sodium bicarbonate and without phenol red, trypsin, antibiotic/antimycotic solution and Foetal Bovine Serum (FBS) were all obtained from SigmaAldrich (Poole, UK). Dulbecco's Modified Eagle's Medium (DMEM) was obtained from Thermo

3

Fisher Scientific (Waltham, USA) and supplemented with penicillin, streptomycin and amphotericin at final concentrations of 100 units/mL, 100 µg/mL and 0.25 µg/mL, respectively, and FBS at 10% (v/v). Transwell® permeable supports of 12 mm diameter, 0.4 m pore size, were obtained from Corning Life Sciences (Tewksbury, MA, USA). MTS reagent (commercially known as “CellTiter 96 AQueous One Solution Cell Proliferation Assay”) was purchased from Promega (Madison, Wisconsin). LDH assay kit (Pierce LDH Cytotoxicity Assay Kit) was purchased from Thermo Scientific (Waltham, MA, USA). Fluorescein isothiocyanate (FITC)-labelled dextran of approximate average MW of 4 kDa (FD4) and 10kDa (FD10) were purchased from Sigma-Aldrich (Poole, UK).

METHODS Mass analysis For mass analysis, approximately 2.5 mg of the rhamnolipid mixture was dissolved in 1 mL of water and analysed by direct injection in an electrospray ionisation (ESI) mass apparatus (HP 1100 LC/MSD, Agilent), equipped with a single quadrupole detector. The sample (1 μL) was analysed in negative mode at different fragmentor voltages (0, 30, 60, 90 V).

HPLC analysis HPLC runs were performed using a HPLC system (HP1100, Agilent Technologies) equipped with a photodiode array detector (DAD). Rhamnolipids were analysed after derivatization using 2bromoacetophenone [21]. The separation of rhamnolipids was achieved using a reverse phase liquid chromatograpy with a C18 Discovery column (5 µm, 15 cm x 4.6 mm) (Supelco, USA). The elution was made using water/acetonitrile mixture at 70:30 ratio. Flow rate was 0.8 ml/min. Estimation of the relative amount (as a percentage) of the mono-and di-rhamnolipids in the mixture was performed by the ratio of the integrated area over time relative to the chromatographic peaks measured at 248 nm (ChemStation software, Agilent Technologies).

4

Surface tension measurements Surface tension measurements were performed at 37°C by tensiometry, using the Du-Noüy ring method (DCA-100 contact angle tensiometer; First Ten Angstrom, USA). A stock solution of rhamnolipids was prepared in HBSS and then diluted to obtain different concentrations. Surface tension of each solution was measured, controlling the temperature of the vessel (Lauda E300 circulating thermostat). The critical micelle concentration (CMC) and the surface tension at CMC (γCMC) were determined from the breakpoint of the surface tension versus log surfactant concentration (mM). Measurements were performed in triplicates.

Dynamic light scattering Dynamic light scattering measurements were performed using a Malvern Zetasizer NanoS instrument (Malvern, Worchestershire, UK). Counts (Kcps) of different concentration of rhamnolipids solutions were recorded as previously reported [22]. CMC was determined by the straight-line interception method. Measurements were performed in triplicates at 37°C.

Cytotoxicity assay Caco-2 and Calu-3 cells were seeded on 96-well plates at 10,000 cells per well and cultured in DMEM for 48 h. Prior to the assay, culture medium was removed and replaced with different concentrations of rhamnolipids dissolved in HBSS. Triton X-100 (0.1% v/v in HBSS) and HBSS were used as the positive and negative control, respectively. Cells were incubated (at 37°C, 5% CO2) with samples and controls for 3 h. Samples and controls were then removed and the tests were subsequently conducted according to the manufacturers’ instructions for both MTS and LDH assay, with at least four repeats for each sample. The absorbance of the formazan product in both cytotoxicity assays was measured at 490 nm using a plate reader (Tecan M200 Pro). The EC50 values (concentration of surfactants inducing 50% cell death from MTS and concentration of

5

surfactants inducing 50% of LDH release from LDH assays) were calculated using a nonlinear regression (Prism version 5.0, GraphPad Inc., USA) as follows:

Eq 1

where Top and Bottom are plateaus in the units of the Y axis.

TEER Caco-2 and Calu-3 cells were detached from the flasks and seeded on filter inserts (Transwell®) at 100,000 cells/cm2. Cells were then cultured in DMEM to confluence for 21 days (Caco-2) or 14 days (Calu-3) using liquid-covered conditions (LCC). Culture medium was changed every 48 h. Only monolayers with a TEER ≥800 Ωcm2 for Calu-3 cells and TEER ≥1000 Ωcm2 for Caco-2 were used in these experiments. Prior to sample application, culture medium was replaced with HBSS. Baseline TEER was recorded following 30 min equilibration in HBSS. Different concentrations of rhamnolipids (15, 30, 60, 90 μg/mL) in HBSS were selected based on the results obtained from the cytotoxicity assays. These were applied to the apical side of the cell monolayers, while HBSS was applied in the basolateral side. Cells were incubated with the samples or HBSS as a control for 3 h. TEER was measured every 30 min. Following this, samples were removed and cells washed extensively with PBS. Culture medium (DMEM) was then added to both sides of the cell monolayers. A further measurement of TEER was taken 24 h following sample application to establish TEER reversibility. The change in TEER was reported as a percentage relative to the baseline value. TEER was measured using an EVOM Voltohmmeter (World Precision Instruments, USA), equipped with a pair of chopstick electrodes. Background resistance due to the filter (∼100 to 110 Ωcm2) was deducted from all the measurements. All experiments were performed in four replicates.

6

Permeability assays Cell monolayers with TEER ≥800 Ωcm2 (Calu-3) or ≥1000 Ωcm2 (Caco-2) were used in these studies; FD4 and FD10 were used as model macromolecular drugs. Prior to sample application, culture medium was removed and the cell layers washed with PBS. Cells were then equilibrated in HBSS for 30 min. Rhamnolipid solution at the same concentrations as those tested in TEER experiments and FD4 or FD10 at a final concentration of 0.5 mg/mL in HBSS were then applied to the apical side of the cells. HBSS was added in the basolateral side. Basolateral solution was sampled (100 μl volumes) at 30, 60, 90, 120, 150 and 180 min after sample application and the sample volume replaced with fresh HBSS. The amount of FD4 and FD10 permeating the cell monolayers in 3 hours was quantified by spectrofluorimetry, using a Tecan M200 Pro plate reader. After the final sampling, cell monolayers were then washed with PBS and TEER measured to ensure an intact cell layer integrity during the permeability experiments. FD4 and FD10 permeability is expressed as the apparent permeability coefficient (P app), calculated using the following equation:

Eq.2

Papp, apparent permeability (cm/s); ΔQ/Δt, permeability rate (amount of FD traversing the cell layers over time); A, diffusion area of the layer (cm2); C0, apically added FITC-dextran concentration. The experiment was conducted in four replicates.

Statistical analysis All statistical comparisons were made through a one-way ANOVA test followed by Dunnett test (Prism version 5.0, GraphPad Inc., USA). P values < 0.05 were considered as statistically significant.

7

RESULTS Chemical characterization of rhamnolipid mixture The negative ESI mass spectrum obtained after direct injection of the rhamnolipid mixture clearly shows two main molecular ions (M-1) with a larger relative abundance. These ions have 503 m/z and 649 m/z, denoting the predominance in the mixture of rhamnolipids with molecular weight of 504 and 650 daltons. According to the literature, these molecular weights can be attributed to monorhamnolipids (Rha-C8-C12 and Rha-C10-C10) and di-rhamnolipids (Rha-Rha-C8-C12 or Rha-Rha C10-C10) [23]. The other ions in the range between 450 m/z and 700 m/z have a relative abundance less than 20% and they can be attributed to the presence in the mixture of smaller amounts of additional rhamnolipids such as Rha-C8-C10 (475 m/z), Rha-C10-C12:1 (529 m/z), Rha-C10-C12 (531 m/z), Rha-Rha C10-C8 (621 m/z) (Figure 1). Considering the fragmentation ions, the most abundant are 333 m/z and 187 m/z, which can be ascribed to Rha-C10 (334 MW) and 3hydroxydecanoic acid (188 MW), suggesting that the main rhamnolipids of the mixture are RhaC10-C10 (504 MW) and Rha-Rha-C10-C10 (650 MW). This was also confirmed by repeating the mass spectra at increasing fragmentor voltages (from 0 to 120 V) in order to increase the fragmentation of the molecules. In this case, an increase in the relative percentage of 333 m/z ion (Rha-C10) was observed, confirming the presence, as main components of the mixture, of rhamnolipids bearing a C10-hydrocarbon chain (SF1). Mass analysis also confirmed the purity of the rhamnolipid mixture as declared by the manufacturer (90%). An estimation of the relative amounts (%) of the mono- and di-rhamnolipids in the mixture after derivatization was achieved through HPLC. Figure 2 reports the obtained chromatogram, which presents three distinctive eluted groups of peaks. The shaper peak with the lowest retention time is ascribed to the excess of the derivatization agent (2-bromo acetophenone). The other two groups of peaks are referred to mono(3.8 min of retention time) and di-rhamnolipids (7.9 min of retention time) in the mixture detected as phenacyl esters. These peaks were identified according to literature and the polarity of compounds [24]. An additional proof of their identification can be given by UV spectra related to

8

the peak (SF2). From the ratio of the integrated area the relative percentage of the components were calculated as 63-67% of the di-rhamnolipids and 37-33% of the mono-rhamnolipids. H-NMR and C-NMR spectra of the rhamnolipid mixture were also recorded and can be found in the supplementary materials (SF3 and SF4).

Surface tension and CMC determination Figure 3 shows the results obtained from tensiometry and dynamic light scattering (counts) analysis. These techniques were employed to determine the critical micelle concentration (CMC) of the rhamnolipid mixture. Both plots display two evident breakpoints of the experimental data, which can be considered as the CMC of the surfactant mixtures. In the surface tension vs concentration plot, the plateau indicates the concentration at which the air-water interface is completely saturated by the rhamnolipids. On the other hand, in the counts vs concentration plot, the increase in the scattered light to the detector (counts) is associated with the formation of mixed rhamnolipid surfactant micelles and/or aggregates [25]. The CMC values calculated by the two techniques were found to be 881±7.0 µg/mL by tensiometry and 112.0±7.0 µg/mL by DLS. Surface tension at CMC (γCMC) was 27.3±1.0 mN/m.

Cytotoxicity assays The toxicity of rhamnolipids towards human epithelial cell lines, Caco-2 and Calu-3, was evaluated by MTS and LDH assays. A concentration-dependent effect on cell toxicity against the tested cell lines can be observed with both cytotoxicity assays (Figure 4). The EC50 values, calculated by a non-linear regression fitting, from the MTS assay were 47±2 µg/mL for Caco-2 and 54±2 µg/mL for Calu-3 cells. Higher values were obtained from the LDH assay, namely 105.6±2 µg/mL in Caco-2 and 93.25±2 µg/mL in the Calu-3 cell line.

TEER

9

A concentration-dependent decrease in TEER was observed in both Caco-2 and Calu-3 cell monolayers after exposure to different concentrations of the rhamnolipid mixture (Figure 5). The lowest TEER (around 50% of the baseline value for Caco-2 or around 60% for Calu-3) was measured following the application of the highest concentration of rhamnolipid solution (90 µg/mL). The decrease in TEER was persistent while cells remained exposed to rhamnolipid samples (3 hours). The reversibility of the effect was confirmed by measuring TEER after 24h from the removal of rhamnolipid solutions. A complete TEER recovery was observed in both cell lines with all the tested sample concentrations.

Permeability assay The effect of different concentrations of rhamnolipids on FD4 and FD10 permeability across Caco2 and Calu-3 cell monolayers is shown in Figure 6. A concentration-dependent effect was observed, with 15 µg/mL and 30 µg/mL concentrations not producing a statistically significant effect, while larger doses of 60 µg/mL and 90 µg/mL notably increasing the apparent permeability values of both macromolecules. Generally, the calculated apparent permeability coefficients were higher for FD4 compared to FD10, as it would be expected considering the molecular weights of these materials. A comparable effect on the macromolecular permeability across Caco-2 and Calu-3 was found with the lower two concentrations, as well as the largest tested dose (90 μg/mL), which led to a 6.5-fold enhancement of FD4 permeability and a 3.2-fold increase of FD10 permeability in both cell lines. With the 60 µg/mL dose, a statistically significant permeability increasing effect was only apparent in Caco-2 monolayers.

DISCUSSION The modulation of epithelial permeability to permit mucosal absorption of biotherapeutics in order to achieve therapeutically acceptable bioavailability in a safe manner is one of the key challenges in macromolecular drug delivery. Several classes of compounds have been examined for decades for

10

their potential to achieve this objective, with limited success in terms of translation to clinic. These compounds predominantly include surfactants and polymers, evaluated as permeability enhancers for application in different routes of administration (e.g. oral or respiratory) both in vitro and in vivo [15][26][27]. Unlike other glycolipid-based surfactants, which have been extensively investigated as absorption enhancers [28][29], rhamnolipid biosurfactants have not been explored sufficiently in this field of application. In fact, to the best of our knowledge, currently there are only two studies reporting the evaluation of rhamnolipids as permeability enhancers [19] [20]. These studies examined the effect of rhamnolipid mixtures on the permeability of small molecules across Caco-2 monolayers. Therefore, no literature reports are currently available on the effect of rhamnolipids on other epithelial cells/models or indeed their potential as absorption enhancers for macromolecular drugs. Because of their production by a bacterial fermentation process, rhamnolipids are obtained as a mixture of different congeners and the elucidation of the properties of individual constituent compounds is not straightforward. This could lead to misleading results in the literature because the composition of the surfactant mixture can have a large impact on the biological properties (e.g. cytotoxicity or permeability-enhancing effect). For this reason, in the present study, prior to testing cytotoxicity and the effect on permeability, chemical identification of the main components and estimation of the relative percentage of mono- and di-rhamnolipids were performed by ESI mass analysis and HPLC (the supplier of the rhamnolipid mixture did not disclose the composition; only the purity of the material). A number of studies have reported the mass spectra of rhamnolipid mixtures produced from different bacterial species [23,30,31] and the two main components of the mixture in this study were identified as Rha-C10C10 (MW 503) and, RhaRha-C10C10 (MW 650). This therefore indicates a prevalence in the mixtures of surface active molecules composed of one or two rhamnose residues and two 3-hydroxy decanoic acid units. Moreover, HPLC separation revealed that di-rhamnolipids are more abundant than mono-rhamnolipids. Surface and aggregation properties of the rhamnolipid mixture were investigated through tensiometry and dynamic light scattering. The analyses were performed in HBSS and at 37°C to enable a comparison with the data 11

obtained from the biological assays. The main parameter calculated from these techniques is the CMC of the rhamnolipid mixture. It is useful to know the CMC in this context as cytotoxicity is dependent on this physicochemical parameter. Generally, concentrations of non-ionic surfactants close to, or higher than, the CMC value are cytotoxic when tested by in vitro assays [32][33]. The calculated CMC for the rhamnolipid mixture was found to be comparable to the range of values reported in the literature [34], despite the fact that a direct comparison is not always reliable owing to the different chemical composition of rhamnolipid mixtures and testing conditions. A study by Ikizler and colleagues reported a CMC value of 1×10−4 M (50.4 µg/mL) for pure Rha C10-C10 (MW 504) and of 1.5×10−4 M (97.5 µg/mL) for pure Rha-Rha C10-C10 (MW 650) [35]. The calculated CMC value for the analysed mixture was close to that of pure Rha-Rha C10-C10, confirming the prevalence of di-rhamnolipids in the mixtures. CMC was also determined by DLS and was found to be somewhat higher than that calculated from tensiometry, as previously observed for other surfactants [22,36]. In terms of cytotoxicity, work by Jiang et al. reported the cytotoxicity of a mixture of pure mono(Rha-C10 was the prevalent congener) and a mixture of pure di-rhamnolipids (Rha-Rha C10-C10 was the prevalent congener) in FBS-free or FBS-containing medium (10% v/v) on different cancer or normal human cells [17]. The toxicological profile of the mixture investigated in the present work was found to be comparable to that reported for pure di-rhamnolipids on Caco-2 cells in FBSfree medium (IC50 around 50 µg/mL). Higher calculated IC50 values were apparent with the LDH assay (around 90-100 µg/mL), which can be attributed to the properties of rhamnolipids to intercalate in biological membranes [37]. Selection of rhamnolipid concentrations employed for TEER and permeability studies were informed by the results of the cytotoxicity assays. A notable effect on TEER (reduction by around 45-60%) was observed on both Caco-2 and Calu-3 cell monolayers, following the application of concentrations up to 90 µg/mL. A similar effect on TEER (decrease by similar levels, namely 42%) was also reported by Wallace et al. in Caco-2 cell monolayers [20] following the application of a 100 µg/mL dose. 12

No data is available in the literature regarding the effect of rhamnolipids on airway Calu-3 cells, with respect to TEER and permeability. In fact, the potential usefulness of biosurfactants such as lipopeptides (e.g. surfactin) or glycolipids (e.g. rhamnolipids, sophorolipids or mannosylerythritol lipids) for mucosal delivery of macromolecular drugs remains to be fully explored. As observed with other classes of surfactants [38], a significant increase in FD4 and FD10 permeability was only observed at a concentration close to IC50 values. The largest extent of permeability enhancement, apparent at the highest tested concentration (90 µg/mL), reflects the effects on TEER (highest reduction at this dose). Permeability enhancement effect was strongly dependent on the molecular weight of the macromolecules, with rhamnolipids displaying a larger effect on FD4 permeability enhancement compared to FD10 in both Caco-2 and Calu-3 monolayers. The molecular weight effect is confirmed by a previous study with mannitol as a smaller permeant, which was associated with a 13-fold increase in permeability across Caco-2 monolayers following the application of a 100 µg/mL rhamnolipid mixture of composition comparable to that analysed in this study [20]. According to the TEER results, a possible absorption enhancing mechanism of action via the tight junctions (paracellular route) could be suggested for rhamnolipids, although more studies would have to be conducted to confirm this achievement. An involvement of the transcellular route in the macromolecular drug absorption (e.g. transcytosis mediated by nano-size vesicles) cannot be excluded. Overall, this work demonstrates, for the first time, that rhamnolipids show potential as safe and effective excipients that improve mucosal absorption of macromolecular therapeutics.

CONCLUSIONS This study demonstrates that rhamnolipids are capable of increasing the permeability of macromolecular drugs across in vitro models of the intestinal and airway epithelium (Caco-2 and Calu-3 cell monolayers, respectively) in a concentration-dependent manner. Data suggest that the permeability-increasing effect may be attributed to reversible opening of epithelial tight junctions, although more studies are required to confirm this finding. The work therefore points to the

13

potential of biosurfactant rhamnolipids as safe and effective systems for improving the absorption (and hence bioavailability) of macromolecular therapeutics following mucosal administration.

References

[1]

R. Marchant, I.M. Banat, Biosurfactants: a sustainable replacement for chemical surfactants?, Biotechnol. Lett. 34 (2012) 1597–1605. doi:10.1007/s10529-012-0956-x.

[2]

K.K. Sekhon Randhawa, P.K.S.M. Rahman, Rhamnolipid biosurfactants-past, present, and future scenario of global market, Front. Microbiol. 5 (2014) 454. doi:10.3389/fmicb.2014.00454.

[3]

T.T. Nguyen, D.A. Sabatini, Characterization and emulsification properties of rhamnolipid and sophorolipid biosurfactants and their applications., Int. J. Mol. Sci. 12 (2011) 1232–44. doi:10.3390/ijms12021232.

[4]

A. Abalos, A. Pinazo, M.R. Infante, M. Casals, F. García, A. Manresa, Physicochemical and Antimicrobial Properties of New Rhamnolipids Produced by Pseudomonas aeruginosa AT10 from Soybean Oil Refinery Wastes, Langmuir. 17 (2001) 1367–1371. doi:10.1021/la0011735.

[5]

E. Haba, A. Pinazo, O. Jauregui, M.J. Espuny, M.R. Infante, A. Manresa, Physicochemical characterization and antimicrobial properties of rhamnolipids produced byPseudomonas aeruginosa 47T2 NCBIM 40044, Biotechnol. Bioeng. 81 (2003) 316–322. doi:10.1002/bit.10474.

[6]

L. Magalhães, M. Nitschke, Antimicrobial activity of rhamnolipids against Listeria monocytogenes and their synergistic interaction with nisin, Food Control. 29 (2013) 138– 142. doi:10.1016/j.foodcont.2012.06.009.

[7]

T.T.L. Nguyen, A. Edelen, B. Neighbors, D.A. Sabatini, Biocompatible lecithin-based microemulsions with rhamnolipid and sophorolipid biosurfactants: Formulation and potential applications, J. Colloid Interface Sci. 348 (2010) 498–504. doi:10.1016/j.jcis.2010.04.053.

[8]

B. Lu, M. Vayssade, Y. Miao, V. Chagnault, E. Grand, A. Wadouachi, et al., Physicochemical properties and cytotoxic effects of sugar-based surfactants: Impact of structural variations., Colloids Surf. B. Biointerfaces. 145 (2016) 79–86. doi:10.1016/j.colsurfb.2016.04.044.

[9]

T. Stipcevic, A. Piljac, G. Piljac, Enhanced healing of full-thickness burn wounds using dirhamnolipid, Burns. 32 (2006) 24–34. doi:10.1016/j.burns.2005.07.004.

[10] D. Dimitrijevich, A.J. Shaw, A.T. Florence, Effects of Some Non-ionic Surfactants on Transepithelial Permeability in Caco-2 Cells, J. Pharm. Pharmacol. 52 (2000) 157–162. doi:10.1211/0022357001773805. [11] E.S. Swenson, W.B. Milisen, W. Curatolo, Intestinal Permeability Enhancement: Efficacy, Acute Local Toxicity, and Reversibility, Pharm. Res. 11 (1994) 1132–1142. doi:10.1023/A:1018984731584. [12] E.K. Anderberg, P. Artursson, Epithelial Transport of Drugs in Cell Culture. VIII: Effects of Sodium Dodecyl Sulfate on Cell Membrane and Tight Junction Permeability in Human Intestinal Epithelial (Caco-2) Cells, J. Pharm. Sci. 82 (1993) 392–398. 14

doi:10.1002/jps.2600820412. [13] E.K. Anderberg, T. Lindmark, P. Artursson, Sodium Caprate Elicits Dilatations in Human Intestinal Tight Junctions and Enhances Drug Absorption by the Paracellular Route, Pharm. Res. 10 (1993) 857–864. doi:10.1023/A:1018909210879. [14] S.M. Krug, M. Amasheh, I. Dittmann, I. Christoffel, M. Fromm, S. Amasheh, Sodium caprate as an enhancer of macromolecule permeation across tricellular tight junctions of intestinal cells, Biomaterials. 34 (2013) 275–282. doi:10.1016/j.biomaterials.2012.09.051. [15] M.A. Deli, Potential use of tight junction modulators to reversibly open membranous barriers and improve drug delivery, Biochim. Biophys. Acta - Biomembr. 1788 (2009) 892–910. doi:10.1016/j.bbamem.2008.09.016. [16] N. Christova, B. Tuleva, A. Kril, M. Georgieva, S. Konstantinov, I. Terziyski, et al., Chemical Structure and In Vitro Antitumor Activity of Rhamnolipids from Pseudomonas aeruginosa BN10, Appl. Biochem. Biotechnol. 170 (2013) 676–689. doi:10.1007/s12010013-0225-z. [17] L. Jiang, C. Shen, X. Long, G. Zhang, Q. Meng, Rhamnolipids elicit the same cytotoxic sensitivity between cancer cell and normal cell by reducing surface tension of culture medium, Appl. Microbiol. Biotechnol. 98 (2014) 10187–10196. doi:10.1007/s00253-0146065-0. [18] L. Zulianello, C. Canard, T. Kohler, D. Caille, J.-S. Lacroix, P. Meda, Rhamnolipids Are Virulence Factors That Promote Early Infiltration of Primary Human Airway Epithelia by Pseudomonas aeruginosa, Infect. Immun. 74 (2006) 3134–3147. doi:10.1128/IAI.01772-05. [19] L. Jiang, X. Long, Q. Meng, Rhamnolipids enhance epithelial permeability in Caco-2 monolayers, Int. J. Pharm. 446 (2013) 130–135. doi:10.1016/j.ijpharm.2013.02.003. [20] C.J. Wallace, S.H. Medina, M.E.H. ElSayed, Effect of Rhamnolipids on Permeability Across Caco-2 Cell Monolayers, Pharm. Res. 31 (2014) 887–894. doi:10.1007/s11095-013-1210-5. [21] T. Schenk, I. Schuphan, B. Schmidt, High-performance liquid chromatographic determination of the rhamnolipids produced by Pseudomonas aeruginosa, J. Chromatogr. A. 693 (1995) 7–13. doi:10.1016/0021-9673(94)01127-Z. [22] D.R. Perinelli, L. Casettari, M. Cespi, F. Fini, D.K.W. Man, G. Giorgioni, et al., Chemical– physical properties and cytotoxicity of N-decanoyl amino acid-based surfactants: Effect of polar heads, Colloids Surfaces A Physicochem. Eng. Asp. 492 (2016) 38–46. doi:10.1016/j.colsurfa.2015.12.009. [23] E. Déziel, F. Lépine, S. Milot, R. Villemur, Mass spectrometry monitoring of rhamnolipids from a growing culture of Pseudomonas aeruginosa strain 57RP, Biochim. Biophys. Acta Mol. Cell Biol. Lipids. 1485 (2000) 145–152. doi:http://dx.doi.org/10.1016/S13881981(00)00039-1. [24] B. Vijaya, N.R. Jayalakshmi, K. Manjunath, Isolation and partial characterization of a biosurfactant produced by Pseudomonas aeruginosa PAVIJ from contaminated soil., Res. J. Pharm. Biol. Chem. Sci. 5 (2014) 881–895. [25] Ö. Topel, B.A. Çakır, L. Budama, N. Hoda, Determination of critical micelle concentration of polybutadiene-block-poly(ethyleneoxide) diblock copolymer by fluorescence spectroscopy and dynamic light scattering, J. Mol. Liq. 177 (2013) 40–43. doi:http://dx.doi.org/10.1016/j.molliq.2012.10.013. [26] D. Vllasaliu, L. Casettari, R. Fowler, R. Exposito-Harris, M. Garnett, L. Illum, et al., Absorption-promoting effects of chitosan in airway and intestinal cell lines: A comparative 15

study, Int. J. Pharm. 430 (2012) 151–160. doi:10.1016/j.ijpharm.2012.04.012. [27] F. McCartney, J.P. Gleeson, D.J. Brayden, Safety concerns over the use of intestinal permeation enhancers: A mini-review, Tissue Barriers. 4 (2016) e1176822. doi:10.1080/21688370.2016.1176822. [28] A. Szuts, P. Szabó-Révész, Sucrose esters as natural surfactants in drug delivery systems - A mini-review, Int. J. Pharm. 433 (2012) 1–9. doi:10.1016/j.ijpharm.2012.04.076. [29] S. Lucarini, L. Fagioli, R. Campana, H. Cole, A. Duranti, W. Baffone, et al., Unsaturated fatty acids lactose esters: cytotoxicity, permeability enhancement and antimicrobial activity, Eur. J. Pharm. Biopharm. 107 (2016) 88–96. doi:10.1016/j.ejpb.2016.06.022. [30] J. Pereira, E. Gudiña, M. Dória, M. Domingues, L. Rodrigues, J. Teoxeira, et al., Characterization by electrospray ionization and tandem mass spectrometry of rhamnolipids produced by two Pseudomonas aeruginosa strains isolated from Brazilian crude oil, Eur. J. Mass Spectrom. 18 (2012) 399. doi:10.1255/ejms.1194. [31] E. Déziel, F. Lépine, D. Dennie, D. Boismenu, O.A. Mamer, R. Villemur, Liquid chromatography/mass spectrometry analysis of mixtures of rhamnolipids produced by Pseudomonas aeruginosa strain 57RP grown on mannitol or naphthalene, Biochim. Biophys. Acta - Mol. Cell Biol. Lipids. 1440 (1999) 244–252. doi:10.1016/S1388-1981(99)00129-8. [32] Â.S. Inácio, K.A. Mesquita, M. Baptista, J. Ramalho-Santos, W.L.C. Vaz, O. V Vieira, In vitro surfactant structure-toxicity relationships: implications for surfactant use in sexually transmitted infection prophylaxis and contraception., PLoS One. 6 (2011) e19850. doi:10.1371/journal.pone.0019850. [33] K. Ekelund, K. Osth, C. Påhlstorp, E. Björk, S. Ulvenlund, F. Johansson, Correlation between epithelial toxicity and surfactant structure as derived from the effects of polyethyleneoxide surfactants on caco-2 cell monolayers and pig nasal mucosa., J. Pharm. Sci. 94 (2005) 730–44. doi:10.1002/jps.20283. [34] I.E. Kłosowska-Chomiczewska, K. Mędrzycka, E. Hallmann, E. Karpenko, T. Pokynbroda, A. Macierzanka, et al., Rhamnolipid CMC prediction, J. Colloid Interface Sci. 488 (2017) 10–19. doi:10.1016/j.jcis.2016.10.055. [35] B. İkizler, G. Arslan, E. Kipcak, C. Dirik, D. Çelenk, T. Aktuğlu, et al., Surface adsorption and spontaneous aggregation of rhamnolipid mixtures in aqueous solutions, Colloids Surfaces A Physicochem. Eng. Asp. (2016). doi:10.1016/j.colsurfa.2016.06.056. [36] D.R. Perinelli, M. Cespi, L. Casettari, D. Vllasaliu, M. Cangiotti, M.F. Ottaviani, et al., Correlation among chemical structure, surface properties and cytotoxicity of N-acyl alanine and serine surfactants, Eur. J. Pharm. Biopharm. 109 (2016) 93–102. doi:10.1016/j.ejpb.2016.09.015. [37] A. Ortiz, J.A. Teruel, M.J. Espuny, A. Marqués, Á. Manresa, F.J. Aranda, Effects of dirhamnolipid on the structural properties of phosphatidylcholine membranes, Int. J. Pharm. 325 (2006) 99–107. doi:10.1016/j.ijpharm.2006.06.028. [38] S. Shubber, D. Vllasaliu, C. Rauch, F. Jordan, L. Illum, S. Stolnik, Mechanism of Mucosal Permeability Enhancement of CriticalSorb® (Solutol® HS15) Investigated In Vitro in Cell Cultures, Pharm. Res. 32 (2015) 516–527. doi:10.1007/s11095-014-1481-5.

16

Figure 1. ESI negative mass spectrum of rhamnolipid mixture (Fragmentor voltage 30 V).

Figure 2. Chromatogram obtained from the HPLC separation of mono- and di-rhamnolipids after derivatization as phenacyl esters.

Figure 3. Surface tension vs different concentration of rhamnolipids in HBSS measured by force tensiometry at 37°C (A). Counts vs different concentration of rhamnolipids in HBSS measured by dynamic light scattering at 37°C (B).

Figure 4. Effect of rhamnolipids on cell viability (MTS assay) (A) and on membrane damage (LDH assay) (B) of Caco-2 and Calu-3 cells. Data are presented as the mean ± SEM (n = 4).

Figure 5. Effect of different concentrations of rhamnolipids (applied in HBSS at 15, 30, 60 and 90 µg/ml) on transepithelial electrical resistance (TEER) of Caco-2 (A) and Calu-3 (B) monolayers. Data are presented as the mean ± SEM (n = 4).

Figure 6. Effect of different rhamnolipid concentrations (15, 30, 60, 90 µg/ml, applied in HBSS) on the apparent permeability of FITC-labelled dextran of 4kDa (FD4) and 10kDa (FD10) across Caco2 (A) and Calu-3 (B) cell monolayers. Data are presented as the mean ± SEM (n = 4). ****P<0.0001; ***0.0001

17

18

19

20

21