Thermoresponsive hydrogels with low toxicity from mixtures of ethyl(hydroxyethyl) cellulose and arginine-based surfactants

International Journal of Pharmaceutics 436 (2012) 454–462

Contents lists available at SciVerse ScienceDirect

International Journal of Pharmaceutics journal homepage: www.elsevier.com/locate/ijpharm

Thermoresponsive hydrogels with low toxicity from mixtures of ethyl(hydroxyethyl) cellulose and arginine-based surfactants Maria Teresa Calejo a , Anna-Lena Kjøniksen a,b , Aurora Pinazo c , Lourdes Pérez c , Ana Maria S. Cardoso d , Maria C. Pedroso de Lima d , Amália S. Jurado d , Sverre Arne Sande a , Bo Nyström b,∗ a

School of Pharmacy, Department of Pharmaceutics, University of Oslo, P.O. Box 1068, Blindern, N-0316 Oslo, Norway Department of Chemistry, University of Oslo, P.O. Box 1033, Blindern, N-0315 Oslo, Norway c Department of Chemical and Surfactant Technology, IQAC-CSIC, J. Girona 18-26, 08034 Barcelona, Spain d CNC – Centre for Neuroscience and Cell Biology, University of Coimbra, Portugal b

a r t i c l e

i n f o

Article history: Received 20 April 2012 Received in revised form 10 July 2012 Accepted 12 July 2012 Available online 20 July 2012 Keywords: EHEC Thermoresponsive gel system Rheo-SALS Amino acid-based surfactants Biocompatibility

a b s t r a c t Ethyl(hydroxyethyl) cellulose (EHEC) is known to form hydrogels in water at elevated temperatures in the presence of an ionic surfactant. In this paper, the potential use of arginine-based surfactants is explored considering the production of a low toxicity thermoresponsive hydrogel for pharmaceutical and biomedical applications. The interactions between EHEC and the monomeric surfactant N␣ -lauroyll-arginine methyl ester (LAM) and two gemini surfactants N␣ ,Nω -bis(N␣ -acylarginine) ␣,ω-dialkyl amides were evaluated by Rheo-Small Angle Light Scattering measurements. The complex viscosity of the systems was dependent on surfactant concentration and temperature. Under specific conditions, soft gels of homogeneous structure were produced. The cloud point (CP) of the EHEC–LAM system varied significantly with surfactant concentration, while only moderate CP changes were found in the presence of the gemini surfactants. Finally, the effect of the surfactants on the viability of a human cell line was evaluated. Despite the lower toxicity of LAM, the superior gel forming efficiency of the gemini surfactants at lower concentrations revealed their advantageous suitability as components of a biocompatible thermoresponsive gel system. © 2012 Elsevier B.V. All rights reserved.

1. Introduction The interactions between polymers and surfactants in aqueous solution have been a subject of great interest over the past decades. These systems can show notable and complex properties, significantly caused by structure variations. This makes them interesting in a number of industrial applications, for instance in products for personal care (surface conditioning) and pharmaceutical products, detergents and foams (Kwak, 1998). One remarkable example is the combination of amphiphilic polymers with ionic surfactants, in which striking viscosification effects can be observed (Kwak, 1998; Malmsten, 2002).

Abbreviations: LAM, N␣ -lauroyl-l-arginine methyl ester; C6 (LA)2 and C9 (LA)2 , N␣ ,Nω -bis(N␣ -acylarginine) ␣,ω-dialkyl amides with C6 and C9 spacers, respectively; cac, critical aggregation concentration; cmc, critical micelle concentration; CP, cloud point; CTAB, cetyltrimethylammonium bromide; EC50 , half maximal effective concentration; EHEC, ethyl(hydroxyethyl)cellulose; GP, gel point; LCST, lower critical solution temperature; Rheo-SALS, Rheo-Small-Angle Light Scattering; SDS, sodium dodecyl sulfate. ∗ Corresponding author. Tel.: +47 22855522. E-mail address: [email protected] (B. Nyström). 0378-5173/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ijpharm.2012.07.018

Ethyl(hydroxyethyl) cellulose (EHEC) is a non-ionic amphiphilic polysaccharide that has been widely studied in combination with ionic surfactants such as the anionic sodium dodecyl sulfate (SDS) (Hoff et al., 2001; Kjøniksen et al., 1998, 2005; Lund et al., 2001; Nyström and Lindman, 1995; Nyström et al., 1995) and sodium dodecanoate (SDoD) (Bo et al., 2005; Dal-Bó et al., 2011), and the cationic cetyltrimethylammonium bromide (CTAB) (Lund et al., 2001; Nyström and Lindman, 1995; Nyström et al., 1995). EHEC exhibits a lower critical solution temperature (LCST), above which it phase-separates. Specifically, as the temperature is increased, the polymer becomes more hydrophobic and less water-soluble, inducing the formation of large aggregates that separate from the water phase (Hoff et al., 2001; Kwak, 1998). However, in the presence of ionic surfactants this scenario can be significantly changed. It has been suggested that the surfactant molecules interact with the hydrophobic microdomains of the polymer, forming mixed micellar-like structures that can involve substituents from more than one polymer chain (Kjøniksen et al., 1998, 2005). The formation of these micelle-like clusters can both create new connection points and strengthen already existing connections between polymer chains. In addition, the presence of charges in the surfactant molecules endows a polyelectrolyte character onto the otherwise

M.T. Calejo et al. / International Journal of Pharmaceutics 436 (2012) 454–462

uncharged polymer, causing the matrix to swell due to charge repulsion. As the temperature is increased, a combination of events involving an enhanced connectivity, together with the swelling of the network can lead to an increase in viscosity and, at specific conditions, to a sol–gel transition (Hoff et al., 2001; Kjøniksen et al., 1998, 2005; Lund et al., 2001). Accordingly, the EHEC–ionic surfactant system can be particularly interesting for industrial applications requiring temperatureinduced viscosification of a solution or a mixture. Nonetheless, it cannot be disregarded that the use of surfactants poses a number of concerns related with their inherent environmental toxicity, low chemical and biological biodegradability and poor biocompatibility (Morán et al., 2004). In the pharmaceutical, cosmetic and medical fields, the use of surfactants is considerably restrained, due to the low toxicity required for all preparation components (RuelGariepy and Leroux, 2004). The production of more biocompatible and environment-friendly surfactants is therefore of great interest (Morán et al., 2004). In recent years, the synthesis of surfactants with structures based on natural compounds has attracted much attention (Clapés and Rosa Infante, 2002; Holmberg, 2001; Infante et al., 1997; Morán et al., 2004; Takehara, 1989). Surfactants based on amino acids have been shown to be environment friendly. They exhibit low toxicity and high biodegradability, while still retaining good surface activity and aggregation features (Brito et al., 2009; Infante et al., 1997; Martinez et al., 2006; Sánchez et al., 2007). This is the case for the cationic N␣ -lauroyl-l-arginine methyl ester (LAM), a surfactant synthesized from arginine by the group of Infante et al. (1984). The properties of this surfactant have been extensively investigated over the last decades (Castillo et al., 2004; Martinez et al., 2006; Pinazo et al., 1999; Sánchez et al., 2007). Apart from the good surface activity, this surfactant is also an active antibacterial agent against both Gram-positive and Gram-negative bacteria. This is due to its capacity to adsorb and interact with the negatively-charged bacterial membranes and thereby cause their disruption (Castillo et al., 2004; Infante et al., 1984). The surfactant is significantly less hemolytic in isotonic medium than the commercial cationic counterpart hexadecyltrimethylammonium bromide (HTAB), while an important protective antihemolytic effect was additionally observed at low concentrations (Sánchez et al., 2007). Another essential strategy to minimize the toxic effect of surfactants is to decrease the amount of surfactant used. This can be achieved by using surfactants with improved efficiency and enhanced performance, as is the case for gemini or dimeric surfactants (Infante et al., 2010). These surfactants are composed of two hydrophobic chains and two hydrophilic head groups, linked through a covalently bound spacer. This unique structure usually causes them to have much lower critical micelle concentrations (cmc) than the conventional monomeric surfactants. They are also more effective in adsorbing at the water/air interface and in reducing the surface tension of water (Xia and Zana, 2004). The production of gemini surfactants based on natural compounds such as amino acids would thus provide an enhanced solution toward a ‘Green Chemistry’ with improved properties and efficacy. Based on this concept, a new class of gemini cationic surfactants has been synthesized from arginine. They have a structure that consists of two symmetrical long chain N␣ -lauroyl-l-arginine residues linked by amide covalent bonds to an ␣,ω-alkylidenediamine spacer chain of various lengths (Perez et al., 1996). These N␣ ,Nω -bis(N␣ acylarginine) ␣,ω-dialkyl amides or bis(Args) are effective low toxicity antibacterial agents and upon chemical degradation only non-toxic and ecologically friendly products are generated (Perez et al., 1996). In this work, we investigate the potential use of LAM, the monomeric surfactant, and two bis(Arg) surfactants with distinct spacer length in their ability to interact with EHEC and produce

455

Fig. 1. (a) Structure segment of EHEC. (b) Schematic drawings of the surfactants used. Cn (LA)2 – N␣ N␻ -bis-(N␣ -lauroyl arginine) ␣-␻-alquildiamine, n = 6 [C6 (LA)2 ], n = 9 [C9 (LA)2 ]; LAM – N␣ lauroyl arginine methyl ester hydrochloride.

a thermoresponsive gelling system with enhanced environmental and biocompatibility properties. The rheological properties of the system, as well as the potential toxicity of the surfactants on a human cell line are investigated. The results are discussed having in mind their prospective use in human pharmaceutical or medical applications. 2. Materials and methods 2.1. Materials and solution preparation EHEC (structure segment shown in Fig. 1a), product DVT 89017, was obtained from Akzo Nobel Surface Chemistry AB, Stenungsund, Sweden. The polymer had a number average molecular weight (Mn ) of 80,000, a degree of substitution of ethyl groups DSethyl = 1.9/anhydroglucose unit, and the molar substitution of ethylene oxide groups MSEO = 1.3/anhydroglucose unit. The sample

456

M.T. Calejo et al. / International Journal of Pharmaceutics 436 (2012) 454–462

was polydisperse with a polydispersity index (Mw /Mn ) of about 2. EHEC was extensively dialyzed against pure water (cut-off 8000, membrane from Spectrum Medical Industries) and freeze-dried before use. The arginine-based gemini surfactants C6 (LA)2 and C9 (LA)2 were synthesized by a previously described method (Perez et al., 1996). These compounds are made up of two symmetrical long chain N␣ -lauroyl arginine residues of 12 carbon atoms linked by amide covalent bonds to an alkylenediamine spacer chain of six (C6 (LA2 )) or nine (C9 (LA2 )) methylene groups (Fig. 1b). They were obtained with a purity of 99% by chemical condensation of the single N␣ -lauroyl-l-arginine previously protected, to the corresponding methylendiamine in presence of an activating agent. A final deprotection reaction was carried out to obtain the gemini compounds. More details on the synthetic procedures, purification, NMR and HPLC characterization are given elsewhere (Perez et al., 1996). The lauroyl arginine methyl ester surfactant, indicated as LAM (Fig. 1b), was obtained by a two-step procedure: (a) methylation of the acid group of the commercial l-Arginine and (b) condensation of the lauric acid to the amino group of the arginine methyl ester using an activating agent. More details on the synthesis are reported in the literature (Infante et al., 1998). Surfactant solutions were prepared at different concentrations in water by stirring at room temperature until complete dissolution. When necessary, samples were further diluted with water before use. Each surfactant solution was added to an accurately weighed mass of EHEC in order to produce samples with a constant polymer concentration of 2 wt% (in the semidilute regime, well above the overlap concentration c*) and a varying surfactant concentration. Samples were further stirred for several hours (>12 h) at low speed (50 rpm) to ensure complete homogenization of the samples and avoid the formation of air bubbles. 2.2. Rheo-Small-Angle Light Scattering (Rheo-SALS) Rheological and SALS measurements were performed simultaneously under oscillatory shear in a Paar-Physica MCR 300 rheometer, equipped with a specially designed plate–plate glass system. The plate diameter was 43 mm and the gap between the plates was 0.25 mm. All samples were measured according to a step-wise protocol, in which the temperature was changed from 10 to 44 ◦ C (sample equilibration at each temperature: 30 min; measurements during 11 min). A CCD camera (driver LuCam V. 3.8), placed in a parallel position in relation to the screen, captured two-dimensional pictures of the scattering patterns of the samples. Time exposure for the acquisition of pictures was 200 ms. A detailed description of the Rheo-SALS combined technique is provided in an earlier publication (Zhu et al., 2007). 2.3. Evaluation of rheology data The samples were compared in terms of the complex viscosity profiles acquired under different conditions. The gel points (GP) were assessed by plotting the viscoelastic loss tangent (tan ı = G /G ) against temperature at different frequencies. The gel point is estimated by the intersection point of the curves, indicative of the frequency-independency of tan ı (Chambon and Winter, 1987; Izuka et al., 1992; Winter and Chambon, 1986). The GPs were additionally confirmed by plotting the ‘apparent’ viscoelas  tic exponents n and n (G ∼ ωn ; G ∼ ωn ) against temperature and by finding the crossover where n equals n (Chambon and Winter, 1987; Izuka et al., 1992; Winter and Chambon, 1986). A more detailed description of the methods used for the estimation of the GPs can be found in a previous publication (Kjøniksen et al.,

1998). In the cases where a sol–gel transition was observed, the gel strength parameter (S) was estimated from the following equation: G =

G = Sωn  (1 − n) cos ı tan ı

(1)

where G and G are the storage and loss modulus, respectively, ı is the phase angle between stress and strain, ω is the angular frequency, n is the relaxation exponent, and  (1 − n) is the Legendre gamma function, as described by the model of Izuka and Winter (Izuka et al., 1992). The homogeneity of the gel system will also be discussed in terms of the fractal dimension value (df ), calculated from the model of Muthukumar (1989) using the following equation, for a polydisperse system: n=

d(d + 2 − 2df )

(2)

2(d + 2 − df )

where d represents the space dimension (d = 3). 2.4. Cytotoxicity of arginine-based surfactants HeLa cells (human epithelial cervical carcinoma cell line) were maintained in culture at 37 ◦ C, under 5% CO2 , in Dulbecco’s modified Eagle’s medium–high glucose (DMEM–HG; Sigma, USA), supplemented with 10% (v/v) heat inactivated fetal bovine serum (FBS; Sigma, USA), penicillin (100 U/ml) and streptomycin (100 ␮g/ml). HeLa cells grown in monolayer were detached by treatment with 0.25% trypsin solution (Sigma, USA). The cell viability in the presence of the cationic surfactants was assessed by the modified Alamar Blue assay, which measures the oxido-reductive capacity of cells, through the spectrophotometric analysis of the produced metabolites (Konopka et al., 1996). Briefly, HeLa cells were seeded at a density of 50,000 cells/well in 48-well plates and incubated at 37 ◦ C for 24 h. The surfactant solutions, prepared in Opti-MEM (reduced serum) medium at different concentrations, were added to the cells and left to incubate at 37 ◦ C. After 48 h, the medium was replaced with 300 ␮l of 10% (v/v) Alamar Blue (Sigma, USA) dye in DMEM cell culture medium and the plates were further incubated for 45 min at 37 ◦ C. An aliquot of the supernatant (150 ␮l) was collected from each well and transferred to 96-well plates. The absorbance at 570 and 600 nm (information provided by the supplier) was measured in a SPECTRAmax PLUS 384 spectrophotometer and cell viability was calculated according to the equation: Cell viability (%) =

(A570 − A600 ) of treated cells × 100. (A570 − A600 ) of control cells

(3)

3. Results and discussion 3.1. Rheology of the EHEC–arginine surfactant systems One essential requirement in a polymer-based gelling system is that the polymer chains are close enough to overlap each other. This condition was the reason why all solutions were prepared at a polymer concentration of 2 wt%, i.e., in the semidilute concentration regime. Under these conditions, the mean distance among polymer chains is short, causing them to superimpose each other creating an interpenetrating network (Tanaka, 2011). In the presence of ionic surfactants, it is further expected that the formation of mixed micelles between the hydrophobic groups in EHEC and the amphiphilic surfactants creates the conditions necessary to induce a sol–gel transition as the temperature is increased (Kjøniksen et al., 1998; Nyström and Lindman, 1995; Nyström et al., 1995). Rheo-SALS is a powerful combined technique that allows us to characterize the structure evolution in the EHEC–surfactant system under oscillatory shear through the simultaneous acquisition

M.T. Calejo et al. / International Journal of Pharmaceutics 436 (2012) 454–462

Fig. 2. Effect of surfactant concentration and temperature on the complex viscosity of 2 wt% EHEC solutions. The inset in C9 (LA)2 shows with better resolution the viscosification in the range 1 × 10−3 to 1 × 10−2 mmol kg−1 .

of rheological data and light scattering patterns indicative of the macroscopic features of the system. The complex viscosity profiles of 2 wt% EHEC solutions in the presence of the Arg-derived surfactants are shown in Fig. 2. In the presence of LAM (upper panel in Fig. 2), the viscosity profile is dependent on surfactant concentration and temperature. It can be observed that at the lowest temperatures, the system generally exhibits a low viscosity due to the low connectivity of the samples. The presence of a solution that flows when handled makes it possible to create a liquid formulation that can be easily administered. A small viscosification maximum, observed in the moderate concentration range 10–15 mmol kg−1 (>cmc, which in the case of LAM is ca. 6 mmol kg−1 (Pinazo et al., 1999)) is progressively becoming more prominent as the temperature is raised up to 37 ◦ C. At this temperature, the growth (Cabane et al., 1996; Kjøniksen et al., 2005) of the hydrophobic microdomains of EHEC and the interactions with the surfactant will favor enhanced connectivity between the polymer chains, and ultimately this causes the

457

viscosity to increase. This associating behavior between EHEC and conventional ionic surfactants has been widely addressed before (Kjøniksen et al., 2005; Lund et al., 2001). As the temperature is increased to 40 ◦ C, the viscosity maximum is found to shift to a lower surfactant concentration. This event has been previously observed in the presence of SDS (Kjøniksen et al., 1998) and can be explained by a decrease in the critical aggregation concentration (cac) as a result of the augmented hydrophobicity of EHEC at higher temperatures. The viscosification maxima observed in the presence of LAM were additionally found to result in sol–gel transitions. These transitions were estimated as described in the experimental section. One example for each surfactant, representing the crossover where n equals n is shown in Fig. 3. The transitions took place at different temperatures, depending on the surfactant concentration (Table 1). In the case of LAM, the increase of surfactant concentration from 12 to 15 mmol kg−1 causes the GP to increase, due to a higher solubilization of the polymer’s hydrophobic microdomains in the presence of an excess surfactant (Cabane et al., 1996). As a result, a higher temperature is necessary in order to create the connectivity provided by the hydrophobic “lumps” needed to achieve gelation. In addition, as can be seen from Table 1, the resulting gel shows a remarkably low value of the gel strength parameter (S), suggesting that the solubilization generates smaller hydrophobic domains and thereby weakens the connectivity in the gel network. The low values of df indicate a network with an “open” structure (Kjøniksen et al., 1998). The gel formed at 12 mmol kg−1 showed a higher value of S due to an improved connectivity generated by the larger hydrophobic microdomains. However, this value is lower than the previously estimated value for the EHEC–SDS and EHEC–CTAB systems (Nyström et al., 1995). Even so, the EHEC–LAM system revealed a high value of df and hence the indication of a more homogeneous network (Kjøniksen et al., 1998). Higher gel strength values may additionally be obtained by increasing the polymer concentration, which will promote the development of more entanglements and larger hydrophobic microdomains that favor a higher cross-linking density (Nyström et al., 1995). In the presence of the C6 and C9 gemini surfactants, two concentration regions seem to be observed from the rheology measurements. These regions will be separately analyzed as follows. First of all, the viscosity of the system was very low at 11 and 25 ◦ C, while a clear increase is observed in the range 1 × 10−3 to 1 × 10−2 mmol kg−1 as the temperature is increased. At this stage, the samples are in the vicinity of the critical micelle concentration (cmc; C6 (LA)2 and C9 (LA)2 = ca. 2 × 10−3 and 3 × 10−3 mmol kg−1 , respectively (Pinazo et al., 1999)). As discussed in the case of LAM addition to EHEC, these results are explained by the augmented association between the surfactant molecules and the polymer that becomes more hydrophobic as the temperature is raised. As the optimal balance between the electrostatic repulsion, caused by the surfactant molecules, and the hydrophobic associations is approached, a sol–gel transition is observed. In this concentration range, the GP (Table 1) generally increases with increasing surfactant concentration, as a result of a higher solubilization of the polymer’s hydrophobic moieties. In other words, in the gel region, higher surfactant concentrations hamper to some extent the hydrophobic associations and thus a higher temperature is necessary to achieve enough connectivity to induce gelation. A similar trend has been reported (Kjøniksen et al., 1998) for gelling EHEC/SDS systems. As discussed in the case of LAM, this deteriorates the connectivity of the gel network, which explains the lower gel strength observed as the surfactant concentration increases. However, in this range of surfactant concentrations, the reduction of S is moderate. Accordingly, the fractal dimension (df ) is nearly

458

M.T. Calejo et al. / International Journal of Pharmaceutics 436 (2012) 454–462

Table 1 Gel properties of the thermoresponsive EHEC–arginine surfactant system. System

[Surfactant]/mmol kg−1

Gel point/◦ C

Fractal dimension (df )

Gel strength parameter (S)/Pa sn

2 wt% EHEC, LAM

12 15

36.0 40.5

2.13 1.68

14.2 0.41

2 wt% EHEC, C6 (LA)2

0.002 0.010 0.015 3.00

36.5 40.0 44.0 38.0

2.42 2.42 2.43 1.62

18.3 17.1 14.3 0.02

2 wt% EHEC, C9 (LA)2

0.002 0.010 0.030 2.00 3.00

39.0 43.0 39.0 33.0 33.0

2.45 2.4 2.39 1.42 1.56

18.9 16.4 5.61 0.01 0.03

constant (ca. 2.4), in agreement with the structure of a homogeneous network (Kjøniksen et al., 1998). In this context it is interesting to note that the gel strength values found for the present gel systems are in the same range as reported elsewhere for gels of the polyacrylic acid polymer, commercially known as Carbopol® 971-P NF (Sanz Taberner et al., 2002). Specifically, values roughly between 0.5 and 40 Pa sn were estimated for the Carbopol® gels, depending on pH and polymer concentration. In this article, these gels were considered suitable vehicles in bioadhesive formulations for mucosal application. Ma and co-workers (Ma et al., 2008) prepared Pluronic F127-g-poly(acrylic acid) copolymers as in situ gelling vehicles for ophthalmic drug delivery applications. In this case, higher polymer concentrations were used (4–8%), producing substantially higher values of complex viscosity (dependent on temperature), than those estimated for the systems in the present work. The drug release rates were found dependent on the acrylic acid/Pluronic molar ratio, weight-average molecular mass and the complex viscosity of the copolymers, as well as the polymer concentration (Ma et al., 2008). The values of gel strength estimated for the hydrogels are also lower (despite in the same range of magnitude) than those found elsewhere for commercial vaginal polymer gels, where it is also mentioned that the rheological properties of the gels appear to be well correlated with claimed therapeutic/usage purpose (Neves et al., 2009). In the surfactant concentration range between 1 × 10−2 and 1 × 10−1 mmol kg−1 , a decline of the complex viscosity is detected. This phenomenon is also verified in the presence of LAM and for other EHEC–surfactant systems (Kjøniksen et al., 1998; Lund et al., 2001) and has frequently been associated with the obstruction of the hydrophobic associations within the network due to an excess of surfactant surrounding the hydrophobic groups of EHEC and an enhanced electrostatic repulsion. This causes a connectivity deficiency, even at the highest temperatures. However, in the case of gemini and considerably hydrophobic surfactants, another hypothesis should be considered, which is that the high hydrophobicity of the surfactants is exceeding the repulsion effect of the polymer chains endowed by the bounding of the surfactant to the polymer. This would lead to a more intricate scenario than what is commonly observed for simple surfactants. The enhanced association of the polymer can lead to the formation of a fragmented network, which is anticipated to cause a decrease in the viscosity of the solution. The determination of the cloud points of these surfactants should assist in the establishment of the real situation, as will be discussed later. The second region of rheological significance is observed in the presence of very high concentrations of the gemini surfactants (i.e., ca. 1000 times higher than the cmc). Particularly in the case of C9 (LA)2 , an important temperature-dependent viscosification maximum is observed in this concentration region (Fig. 2). This behavior can be explained in light of the previous knowledge of the aggregation behavior of the gemini surfactants. In an earlier publication,

it was shown that 0.1 wt% (ca. 0.9 mmol kg−1 ) aqueous solutions of C9 (LA)2 formed flat, twisted ribbons that evolved into tightly twisted ribbons and thread-like ribbons or fibers (6–8 nm in diameter) as the concentration was increased to 0.7 wt% (ca. 6 mmol kg−1 ) (Weihs et al., 2005). Such structures have been claimed to be created by the extended alignment of surfactant molecules, associated with strong hydrogen bonding within and between the amphiphilic molecules and the surrounding water molecules (Fuhrhop et al., 1993; Weihs et al., 2005). This high anisotropic behavior is also known for causing a high viscosity (Tadros, 2010), as observed in our work. The observed viscosification at high concentrations of C9 (LA)2 (Fig. 2) might hence not be specifically related with the interactions with EHEC, but rather a result of the self-assembling properties of the gemini surfactant in the bulk. This is additionally supported by the particularly low gel strength and df values estimated for the highest concentrations of C9 (LA)2 (Table 1). A slight tendency for a similar viscosification is found at the highest concentrations of C6 (LA)2 . An apparent GP was found at 3 mmol kg−1 , similarly revealing a low value of S and df . However, in this case even higher concentrations should be necessary in order to achieve the full effect of the self-aggregation behavior of the surfactant molecules. In this respect, the differences between the two gemini surfactants might be related to the spacer length. Specifically, the high hydrophobicity and flexibility of the spacer in C9 (LA)2 is said to penetrate into the aggregated core, reducing the distance between the polar head groups and forming a compact molecule with low surface area. By doing so, this surfactant is able to form closely packed structures at lower concentrations than for the case of the C6 spacer homologue (Pérez et al., 1998; Weihs et al., 2005).

3.2. Cloud point determination and Rheo-SALS As discussed previously, polymers such as EHEC produce transparent solutions below the LCST. Above the LCST, the samples become turbid as a result of a macroscopic phase-separation. In the case of EHEC, the LCST marks the temperature at which extensive hydrophobic associations take place, forming large lumps or aggregates that are responsible for the observed turbidity. The presence of ionic surfactants can to some extent prevent the onset of these associations and thereby help to stabilize the network (Hoff et al., 2001; Kjøniksen et al., 1998, 2005; Lund et al., 2001; Nyström and Lindman, 1995; Nyström et al., 1995). The cloud point (CP), defined as the temperature at which the solution becomes turbid and heterogeneous (Teraoka, 2002), can thus be a valuable parameter when comparing the stability or aggregation status of an EHEC sample. During Rheo-SALS measurements, the scattered intensities of the incident light were determined in the presence of varying surfactant concentrations. In a plot of the scattered intensity versus temperature, one can observe that when the CP is reached, the scattered intensity dramatically increases, as a result of formation

M.T. Calejo et al. / International Journal of Pharmaceutics 436 (2012) 454–462

459

Fig. 4. Effect of the concentration of the gemini and monomeric surfactants on the cloud point of semidilute (2 wt%) aqueous EHEC solutions. The dashed line shows an extrapolation into the expected profile observed for LAM concentrations >12 mmol kg−1 .

Fig. 3. Method for the determination of gel points, based on finding the crossover where n equals n . In these examples, the lowest concentration of each surfactant shown to induce gelation of semidilute EHEC solutions is represented. Gel points are indicated by the arrows.

of association complexes. The estimated CP values for EHEC in the presence of LAM and the gemini surfactants are shown in Fig. 4. In the presence of low LAM concentrations (<5 mmol kg−1 ), the value of CP drops as the surfactant concentration is increased, and a plateau region is evolved in the interval from 5 mmol kg−1 to 10 mmol kg−1 . The decrease of CP at low surfactant concentrations has previously (Kjøniksen et al., 2005; Lund et al., 2001) been observed for EHEC/SDS solutions, and this effect was attributed to enhanced polymer–surfactant interactions as the critical aggregation concentration of the system was approached. Surfactant concentrations up to 5 mmol kg−1 are below the cmc (6 mmol kg−1 ) (Pinazo et al., 1999) of the surfactant. The hypothesis is that in the plateau region mixed polymer–surfactant micelles are formed, and

the pronounced drop of the CP in the domain 10–12 mmol kg−1 is attributed to the formation of large polymer–surfactant complexes because of the complicated structure and hydrophobicity of the LAM surfactant. At this stage the strong interactions between hydrophobic moieties of the surfactant and hydrophobic microdomains of the polymer generate large association complexes. We should note that the observed CP of approximately 26 ◦ C is much lower than the cloud point of 32 ◦ C (Lund et al., 2001) observed for 2 wt% EHEC solutions without surfactant. It should be mentioned that the turbidity of the LAM solutions without polymer is very low and practically independent of the level of LAM addition over the considered concentration range. This suggests that it is the strong polymer–surfactant interactions that give rise to the low CP values. At a higher level of surfactant addition, the gel network is evolved and the surfactant is more evenly distributed through the network. As a consequence, the number of large polymer–surfactant junction zones decreases and CP increases to values higher than 44 ◦ C. In this case, the determination of the exact CP value was precluded due to experimental limitations, since the highest temperature at which Rheo-SALS measurements could be performed was 44 ◦ C. Up until this temperature, the scattered intensity of EHEC samples in the presence of the highest surfactant concentrations remained mostly unchanged, indicating a higher CP than 44 ◦ C. The rise of CP at high surfactant concentrations indicates that the hydrophobic microdomains are highly solubilized and that the polymer becomes more hydrophilic. The binding of surfactant to the polymer may also give rise to electrostatic stabilization. This suggests that even higher temperatures are necessary in order to cause phase-separation of the EHEC–surfactant mixture. This hypothesis is further supported by the cloud point behavior

460

M.T. Calejo et al. / International Journal of Pharmaceutics 436 (2012) 454–462

Fig. 5. 2D SALS patterns of 2 wt% aqueous EHEC solutions in the presence of the Arg-derived surfactants. The indicated temperatures correspond to the gel points.

found previously (Kjøniksen et al., 2005; Lund et al., 2001) for EHEC in the presence of high levels of SDS addition. The CP of EHEC in the presence of the gemini surfactants was less affected by the increasing surfactant concentrations than in the case of LAM. The general trend that is observed for the gemini surfactants (C9 (LA)2 and C6 (LA)2 ) in the surfactant interval from 0.001 mmol kg−1 to 1 mmol kg−1 is a moderate decrease of the cloud point from about 37 ◦ C to 33 ◦ C. Both gemini surfactants have low cmc values. There are some fluctuations in the CP values in this interval but they are virtually within the experimental uncertainty. In the low concentration range where viscosification and gelation events were observed, i.e., < 0.03 mmol kg−1 (Fig. 2 and Table 1), the change in temperature was only 2 ◦ C. In this concentration range and at the gel point, the 2D scattering images of different samples of C6 (LA)2 are very similar, indicating a comparable turbidity (Fig. 5). The same is observed among different low concentrations of C9 (LA)2 , even though a higher scattering intensity is found in this case, due to the higher GPs and the highest hydrophobicity of the polymer at these temperatures. One might also notice that in the concentration region 0.01–0.1 mmol kg−1 the general decrease in CP is associated with a decrease in viscosity (Figs. 2 and 4), as discussed previously. This means that the high surfactant concentrations cause the hydrophobic association effects to surpass the repulsion between ions (as hypothesized before), leading to a reduced solubility of EHEC and causing it to associate at lower temperatures. Such observation is not commonly observed in EHEC–surfactant systems and is probably a very specific characteristic of the association of the long-chain, long-spacer gemini surfactants with the nonionic polymer. The CP features observed at concentrations higher than 1 mmol kg−1 should also be discussed in light of the specific of the supramolecular association patterns of the surfactants, particularly observed in the case of C9 (LA)2 . The fact that this surfactant is involved in the formation of ribbons or fiber-like aggregated structures in this concentration range (Weihs et al., 2005) suggests that the surfactant is poorly available to interact with the polymer, which explains the unchanged CP with increasing surfactant concentration. Conversely, in this concentration range, the C6 -spacer surfactant is still available to interact with the polymer, since it self-aggregates at even higher concentrations (Weihs et al., 2005). The presence of high amounts of this surfactant are hence able to solubilize the hydrophobic microdomains of EHEC, causing the CP to increase and explaining the low intensity scattering pattern

at 3.00 mmol kg−1 of C6 (LA)2 (Fig. 5) and the low viscosity values observed in Fig. 2. 3.3. Cell viability in the presence of the Arg-derived surfactants One of the most critical aspects of the industrial use of surfactants is concerned with their potential toxicity, making their use in medical and pharmaceutical applications limited. The in vitro cytotoxicity of the Arg-derived surfactants was thus evaluated in a human cell line (HeLa) to investigate their suitability in the production of the EHEC-based thermoresponsive gel system, intended for human use. Factors such as the spacer and alkyl chain length, the specific polar head group and the counterions are known to affect cell viability, while different cells can also react differently to the presence of the foreign compounds (Almeida et al., 2011). The results for cell viability in the presence of increasing surfactant concentrations are shown in Fig. 6. The table in the inset additionally shows the estimated vales of the half maximal effective concentration (EC50 ), i.e., the concentration of surfactant that induces a reduction of cell viability to 50%. As shown in Fig. 6, a clear dose–response relationship is observed in all cases. The value

Fig. 6. Cell viability in the presence of Arg-derived surfactants, tested in the concentration range 0.001–15 mmol kg−1 . Values were measured by the modified Alamar Blue assay and taken as a percentage of the untreated control cells (mean ± standard deviation; n = 3). The table in the inset shows the median effective concentrations of the surfactants (EC50 ) estimated from the dose–response curves.

M.T. Calejo et al. / International Journal of Pharmaceutics 436 (2012) 454–462

of EC50 followed the sequence LAM > C9 (LA)2 > C6 (LA)2 . The lower toxicity observed for the monomeric surfactant is probably related with its lower hydrophobicity, due to the presence of a single alkyl chain. This specific correlation between the hydrophobicity of the Arg-based surfactants and their toxicological potential has been addressed in an earlier publication (Pérez et al., 2002). Considering the similar structure of the gemini surfactants, one can anticipate that any differences between the two should be due to the spacer length and hydrophobicity. One interesting finding is that the EC50 of C6 (LA)2 was lower (i.e., higher toxicity) than that of C9 (LA)2 , despite the shorter spacer length of the former and, hence, lower expected hydrophobicity. The reasons for this might be related to the actual conformation of the surfactants in the aqueous medium. As discussed above, it has been argued that long hydrophobic spacers such as n = 9 become flexible allowing penetration into the hydrophobic core and thereby avoiding contact with the hydrophilic environment (Pérez et al., 1998; Weihs et al., 2005). Accordingly, the head groups may come closer to each other and the area of the molecule decreases. Overall, the reduction in the exposure of the hydrophobic moiety to the aqueous environment might explain a lower cell membrane disturbance in the presence of C9 (LA)2 as compared to C6 (LA)2. These results were also generally in agreement with the tendencies found in environmental toxicity experiments, where a lower toxicity was found for LAM compared to the gemini surfactants (Pérez et al., 2002). In Photobacterium phosphoreum, C6 (LA)2 was also revealed to be more toxic than the C9 homologue, and a similar trend was found between them in the presence of Daphnia magna. Quite importantly, the monomeric surfactant has revealed a lower eye and skin irritation potential than the conventional SDS (as established by in vitro and in vivo assays) (Martinez et al., 2006) and it is less hemolytic than HTAB (Sánchez et al., 2007), even though no information in this respect has been published for the gemini surfactants. Finally, some lines of discussion should be given in view of the intended application of the surfactants. In spite of LAM’s higher biocompatibility, very high concentrations of this surfactant are needed in order to induce a sol–gel transition in semidilute EHEC solutions at elevated temperatures. These concentrations (12 mmol kg−1 ) are well above the EC50 of the surfactant and a significant cytotoxicity is expected under these conditions. Even so, the low EC50 of LAM, namely in comparison with some conventional surfactants, makes this surfactant a promising candidate for other applications requiring lower surfactant concentrations. The gemini surfactants were shown to be significantly more effective in the interactions with EHEC, and as a result very low concentrations are needed in order to induce sol–gel transitions. As such, these surfactants may turn out to be promising candidates for pharmaceutical and medical applications involving thermoresponsive gelling systems, since they are able to induce gelation at concentrations lower than the estimated EC50 values (0.002 mmol kg−1 for C6 (LA)2 and ≤0.03 mmol kg−1 for C9 (LA)2 – Table 1). As a matter of fact, cell viability at the minimum surfactant concentration inducing gelation of EHEC was estimated as higher than 90% in the presence of the gemini surfactants, but only 8% in the presence of LAM. In addition, preliminary experiments revealed no toxic effects of 2 wt% EHEC solutions on HeLa cells, and the presence of the polymer actually seems to contribute to a significantly lower toxicity of the surfactants (data not shown). This finding is explained by the association of the surfactant molecules in the form of mixed micelles with the polymer’s hydrophobic groups, and consequently by the lower availability of free surfactants to interact with the cells. Finally, it should be pointed out that in situations where GP > body temperature, adjustments in polymer concentration or

461

in the formulation are expected to lead to more adequate results for the intended application. 4. Conclusions In this work, EHEC was combined with arginine-derived surfactants and the rheological and turbidity properties were evaluated. Both LAM and the two gemini surfactants were found to induce gelation of semidilute EHEC solutions as the temperature is increased. Particularly low concentrations of the gemini surfactants were established to cause the sol–gel transition. Considering a pharmaceutical or medical application, these systems are suitable candidates for the administration as a liquid solution (at low temperatures), undergoing in situ gelation at body temperature. The systems combining EHEC and the gemini surfactants were found to be particularly suitable for human use, due to the low toxicity of the surfactants at the concentrations inducing polymer gelation, as assessed by toxicity studies on HeLa cells. After the encouraging results from the in vitro experiments, further studies are under way to confirm the suitability of the EHEC–Arg surfactant thermoresponsive hydrogels for in vivo applications. It is well-known that the complex in vivo environment, involving a high ionic strength medium rich in digestive enzymes and the activation of immune defense mechanisms can all influence the stability of the system. Bearing in mind the production of a long-lasting drug delivery system, we have also been working on the combination of these hydrogel systems with drug-loaded polymeric microparticles. In this context, the presence of an in situ gelling system can both contribute to keep the microparticles in the target tissue and as an additional barrier for drug release. Acknowledgments M.T. Calejo, A.-L. Kjøniksen, S.A. Sande and B. Nyström gratefully acknowledge financial support from the Norwegian Research Council, Project Number 190403. This project is part of the SiteDel activities at the University of Oslo. A. Pinazo and L. Pérez would like to acknowledge financial support from the project ‘Spanish Plan Nacional CTQ2009-14151-CO01’. A.M. Cardoso, M.C. P. Lima and A.S. Jurado would like to acknowledge the grants PTDC/QUI-BIQ/103001/2008 and PTDC/QUI-QUI/115212/2009 funded by Portuguese Foundation for Science and Technology and FEDER/COMPETE. A.M. Cardoso is a recipient of a fellowship from the Portuguese Foundation for Science and Technology (SFRH/BD/63288/2009). References Almeida, J.A.S., Faneca, H., Carvalho, R.A., Marques, E.F., Pais, A.A.C.C., 2011. Dicationic alkylammonium bromide gemini surfactants. Membrane perturbation and skin irritation. PLoS One 6, e26965. Bo, A.D., Schweitzer, B., Felippe, A.C., Zanette, D., Lindman, B., 2005. Ethyl (hydroxyethyl) cellulose–sodium dodecanoate interaction investigated by surface tension and electrical conductivity techniques. Colloids Surf. A 256, 171–180. Brito, R.O., Marques, E.F., Silva, S.G., do Vale, M.L., Gomes, P., Araújo, M.J., RodriguezBorges, J.E., Infante, M.R., Garcia, M.T., Ribosa, I., 2009. Physicochemical and toxicological properties of novel amino acid-based amphiphiles and their spontaneously formed catanionic vesicles. Colloids Surf. B 72, 80–87. Cabane, B., Lindell, K., Engström, S., Lindman, B., 1996. Microphase separation in polymer+ surfactant systems. Macromolecules 29, 3188–3197. Castillo, J.A., Pinazo, A., Carilla, J., Infante, M.R., Alsina, M.A., Haro, I., Clapés, P., 2004. Interaction of antimicrobial arginine-based cationic surfactants with liposomes and lipid monolayers. Langmuir 20, 3379–3387. Chambon, F., Winter, H.H., 1987. Linear viscoelasticity at the gel point of a crosslinking PDMS with imbalanced stoichiometry. J. Rheol. 31, 683–697. Clapés, P., Rosa Infante, M.Í., 2002. Amino acid-based surfactants: enzymatic synthesis: properties and potential applications. Biocatal. Biotransform. 20, 215–233. Dal-Bó, A.G., Laus, R., Felippe, A.C., Zanette, D., Minatti, E., 2011. Association of anionic surfactant mixed micelles with hydrophobically modified ethyl (hydroxyethyl) cellulose. Colloids Surf. A 380, 100–106.

462

M.T. Calejo et al. / International Journal of Pharmaceutics 436 (2012) 454–462

Fuhrhop, J.H., Spiroski, D., Boettcher, C., 1993. Molecular monolayer rods and tubules made of ␣-(l-lysine), ␻-(amino) bolaamphiphiles. J. Am. Chem. Soc. 115, 1600–1601. Hoff, E., Nyström, B., Lindman, B., 2001. Polymer–surfactant interactions in dilute mixtures of a nonionic cellulose derivative and an anionic surfactant. Langmuir 17, 28–34. Holmberg, K., 2001. Natural surfactants. Curr. Opin. Colloid Interface Sci. 6, 148–159. Infante, M.R., Pérez, L., Morán, M.C., Pons, R., Mitjans, M., Vinardell, M.P., Garcia, M.T., Pinazo, A., 2010. Biocompatible surfactants from renewable hydrophiles. Eur. J. Lipid Sci. Technol. 112, 110–121. Infante, M.R., Pérez, L., Pinazo, A., 1998. Novel Cationic Surfactants from Arginine. Marcel Dekker Inc., New York. Infante, M.R., Pinazo, A., Seguer, J., 1997. Non-conventional surfactants from amino acids and glycolipids: structure, preparation and properties. Colloids Surf. A 123, 49–70. Infante, R., Dominguez, J.G., Erra, P., Julia, R., Prats, M., 1984. Surface active molecules: preparation and properties of long chain n␣ acyl-l-␣-amino-␻-guanidine alkyl acid derivatives. Int. J. Cosmet. Sci. 6, 275–282. Izuka, A., Winter, H.H., Hashimoto, T., 1992. Molecular weight dependence of viscoelasticity of polycaprolactone critical gels. Macromolecules 25, 2422–2428. Kjøniksen, A.L., Knudsen, K.D., Nyström, B., 2005. Phase separation and structural properties of semidilute aqueous mixtures of ethyl (hydroxyethyl) cellulose and an ionic surfactant. Eur. Polym. J. 41, 1954–1964. Kjøniksen, A.L., Nyström, B., Lindman, B., 1998. Dynamic viscoelasticity of gelling and nongelling aqueous mixtures of ethyl (hydroxyethyl) cellulose and an ionic surfactant. Macromolecules 31, 1852–1858. Konopka, K., Pretzer, E., Felgner, P.L., Düzgünes, N., 1996. Human immunodeficiency virus type-1 (HIV-1) infection increases the sensitivity of macrophages and THP1 cells to cytotoxicity by cationic liposomes. Biochim. Biophys. Acta Mol. Cell Res. 1312, 186–196. Kwak, J.C.T., 1998. Polymer–Surfactant Systems. Mercel Dekker, Inc., New York. Lund, R., Lauten, R.A., Nyström, B., Lindman, B., 2001. Linear and nonlinear viscoelasticity of semidilute aqueous mixtures of a nonionic cellulose derivative and ionic surfactants. Langmuir 17, 8001–8009. Ma, W.D., Xu, H., Wang, C., Nie, S.F., Pan, W.S., 2008. Pluronic F127-g-poly (acrylic acid) copolymers as in situ gelling vehicle for ophthalmic drug delivery system. Int. J. Pharm. 350, 247–256. Malmsten, M., 2002. Surfactants and Polymers in Drug Delivery. Marcel Dekker Inc. Martinez, V., Corsini, E., Mitjans, M., Pinazo, A., Vinardell, M.P., 2006. Evaluation of eye and skin irritation of arginine-derivative surfactants using different in vitro endpoints as alternatives to the in vivo assays. Toxicol. Lett. 164, 259–267. Morán, M.C., Pinazo, A., Pérez, L., Clapés, P., Angelet, M., García, M.T., Vinardell, M.P., Infante, M.R., 2004. Green amino acid-based surfactants. Green Chem. 6, 233–240. Muthukumar, M., 1989. Screening effect on viscoelasticity near the gel point. Macromolecules 22, 4656–4658.

Neves, J., Da Silva, M.V., Gonc¸alves, M.P., Amaral, M.H., Bahia, M.F., 2009. Rheological properties of vaginal hydrophilic polymer gels. Curr. Drug Deliver. 6, 83–92. Nyström, B., Lindman, B., 1995. Dynamic and viscoelastic properties during the thermal gelation process of a nonionic cellulose ether dissolved in water in the presence of ionic surfactants. Macromolecules 28, 967–974. Nyström, B., Walderhaug, H., Hansen, F., Lindman, B., 1995. Rheological behavior during thermoreversible gelation of aqueous mixtures of ethyl (hydroxyethyl) cellulose and surfactants. Langmuir 11, 750–757. Pérez, L., Garcia, M.T., Ribosa, I., Vinardell, M.P., Manresa, A., Infante, M.R., 2002. Biological properties of arginine-based gemini cationic surfactants. Environ. Toxicol. Chem. 21, 1279–1285. Pérez, L., Pinazo, A., Rosen, M.J., Infante, M.R., 1998. Surface activity properties at equilibrium of novel gemini cationic amphiphilic compounds from arginine, Bis (Args). Langmuir 14, 2307–2315. Perez, L., Torres, J.L., Manresa, A., Solans, C., Infante, M.R., 1996. Synthesis, aggregation, and biological properties of a new class of gemini cationic amphiphilic compounds from arginine, bis (Args). Langmuir 12, 5296–5301. Pinazo, A., Wen, X., Pérez, L., Infante, M.R., Franses, E.I., 1999. Aggregation behavior in water of monomeric and gemini cationic surfactants derived from arginine. Langmuir 15, 3134–3142. Ruel-Gariepy, E., Leroux, J.C., 2004. In situ-forming hydrogels—review of temperature-sensitive systems. Eur. J. Pharm. Biopharm. 58, 409–426. Sánchez, L., Martínez, V., Rosa Infante, M., Mitjans, M., Pilar Vinardell, M., 2007. Hemolysis and antihemolysis induced by amino acid-based surfactants. Toxicol. Lett. 169, 177–184. Sanz Taberner, T., Martin-Villodre, A., Pla-Delfina, J., Herráez, J.V., 2002. Consistency of Carbopol 971-P NF gels and influence of soluble and cross-linked PVP. Int. J. Pharm. 233, 43–50. Tadros, T.F., 2010. Rheology of Dispersions: Principles and Applications. Wiley-VCH, Weinheim, Germany. Takehara, M., 1989. Properties and applications of amino acid based surfactants. Colloids Surf. 38, 149–167. Tanaka, F., 2011. Polymer Physics: Applications to Molecular Association and Thermoreversible Gelation. Cambridge University Press, Cambridge, UK. Teraoka, I., 2002. Polymer Solutions – An Introduction to Physical Properties. John Wiley & Sons, Inc., New York. Weihs, D., Danino, D., Pinazo-Gassol, A., Perez, L., Franses, E.I., Talmon, Y., 2005. Selfaggregation in dimeric arginine-based cationic surfactants solutions. Colloids Surf. A 255, 73–78. Winter, H.H., Chambon, F., 1986. Analysis of linear viscoelasticity of a crosslinking polymer at the gel point. J. Rheol. 30, 367–382. Xia, J., Zana, R., 2004. Gemini Surfactants: Synthesis, Interfacial and Solution-Phase Behavior and Applications. Marcel Dekker Inc., New York, USA. Zhu, K., Jin, H., Kjøniksen, A.L., Nyström, B., 2007. Anomalous transition in aqueous solutions of a thermoresponsive amphiphilic diblock copolymer. J. Phys. Chem. B 111, 10862–10870.