Technological stability of solid lipid nanoparticles loaded with phenolic compounds: Drying process and stability along storage

Journal of Food Engineering 196 (2017) 1e10

Contents lists available at ScienceDirect

Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng

Technological stability of solid lipid nanoparticles loaded with phenolic compounds: Drying process and stability along storage
bora A. Campos a, Ana Raquel Madureira a, *, Bruno Sarmento b, c, d, De Maria Manuela Pintado a, Ana Maria Gomes a a
rio Associado, Escola Superior de Biotecnologia, Universidade Cato
lica Portuguesa/Porto, CBQF e Centro de Biotecnologia e Química Fina, Laborato ~o Vital Apartado 2511, 4202-401 Porto, Portugal Rua Arquiteto Loba b CICS e Department of Pharmaceutical Sciences, Institute of Health Sciences-North, CESPU, Rua Central de Gandra, 1317, 4585-116 Gandra, Portugal c ~o e Inovaça ~o em Saúde, 4150-180 Porto, Portugal I3S e Instituto de Investigaça d INEB e Institute of Biomedical Engineering, University of Porto, Rua do Campo Alegre, 823, 4150-180 Porto, Portugal

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 April 2016 Received in revised form 7 October 2016 Accepted 8 October 2016 Available online 11 October 2016

Solid lipid nanoparticles (SLNs) have been widely studied and tested as vehicles for natural compounds. Recently, Witepsol and Carnauba SLNs were shown to be effective systems for the entrapment of rosmarinic acid (RA) and herbal extracts. In the present work, the improvement of stability and bioactivity of these systems was studied. Thus, the freeze-drying of SLNs produced with Witepsol and Carnauba waxes loaded with RA and herbal extracts (sage and savory) were tested. The use of three different cryoprotectants (glucose, mannitol and trehalose) at two different concentrations (5 and 10%, w/v) were evaluated. Furthermore, the prepared SLNs were stored under different conditions (atmosphere, temperature, absence or presence of light) and in different packaging materials, over 365 days. The effect on the SLNs physical stability and bioactivity was assessed. The most suitable cryoprotectant was mannitol at 10% (w/v) for all formulations tested. The solid state of SLNs, with storage at room temperature, in glass flasks, protected from light and under N2 controlled atmosphere were the best storage conditions in which the SLNs bioactivity was maintained during 365 days. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Rosmarinic acid SLNs Witepsol Carnauba wax Freeze-drying Long-term stability

1. Introduction Rosmarinic acid (RA) is an important bioactive polyphenolic compound with several biological activities, such as antioxidant, anti-inflammatory, anti-mutagenic, anti-bacterial, antiviral, among others (Nunes et al., 2015). Salvia sp. and Satureja montana (sage and savory, respectively) medicinal herbs are rich in RA, besides other phenolic compounds present in minor quantities (e.g. caffeic acid, rutin). The use of natural extracts obtained from these herbs lowers the production costs at industrial scale and facilitates the dealing with regulatory issues. Nanotechnology enables the development of nanometer carriers to the delivery of these type of compounds (Paulino et al., 2011). Nevertheless, some difficulties are encountered in the industrial manufacturing due to the poor long-term physical and chemical stability, as well as the choice of pharmaceutical and food

* Corresponding author. E-mail address: [email protected] (A.R. Madureira). http://dx.doi.org/10.1016/j.jfoodeng.2016.10.009 0260-8774/© 2016 Elsevier Ltd. All rights reserved.

grade materials (Blasi et al., 2007). Lipid NPs seem to possess favorable features to overcome the mentioned issues (Blasi et al., 2007; Wissing et al., 2004). Lipid NPs produced with lipids that remain solid at the body temperature (known as solid lipid nanoparticles, SLN®) generally possess higher stability than most emulsions due to their solid state. However, particle stability depends on a number of different features, such as particle size, polydispersity, particle shape, surface charge, and storage conditions, among others. In an aqueous medium, the poor stability of these systems is a barrier for industrial application, but the reduction of water content in aqueous samples could improve SLNs stability (Abdelwahed et al., 2006b; Chacon et al., 1999). Freeze-drying process allows the dehydration of samples, which may improve and increase stability throughout storage (Abdelwahed et al., 2006b). Furthermore, the production of a powder can facilitate the processing and storage. But, the crystallization of ice may induce mechanical stress on SLNs, which may possibly lead to their destabilization. Hence, in order to enable better handling and stability of certain samples it is necessary to add special excipients, thus contributing to the protection of these

2

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10

fragile systems (Abdelwahed et al., 2006b). Cryoprotectants can be used to protect from the stress of freezing, and to preserve the native structure of SLNs (Abdelwahed et al., 2006b). The most popular cryoprotectants are sugars such as trehalose, sucrose and glucose, or the sugar-alcohol mannitol, which, at concentrations of 5e15%, prevent the aggregation of SLNs and protect them against the mechanical injury induced by this process (Franks, 2013; Pikal, 1999). The aim of this work was to assess the long-term physical and chemical stability of two lipid-based SLN types loaded with RA or with sage and savory extracts in fresh and dehydrated (by freezedrying) states. 2. Materials and methods 2.1. Materials The lipid matrices tested were Witepsol H15 (Sasol, Hamburg, Germany) and Carnauba yellow no. 1 wax (Sigma-Aldrich Chemistry, St. Louis, Missouri, USA). The surfactant Tween 80 (poly-

CI% ¼

In the case of the freeze-dried RA-SLNs, these were first reconstituted with phosphate buffer (PBS; 0.1 M; pH 7.0) at a final concentration of 70 mg/mL. The particle size (PS), polydispersity index (PI) and zeta potential (ZP) were measured using a dynamic light scattering (DLS) from ZetaPALS, Zeta Potential Analyzer (Holtsville, New York, USA). All analyses were carried out in triplicate with an angle of 90 at 25  C. 2.3.2. Evaluation of thermal properties The thermodynamic behavior of the SLNs was determined by using Differential Scanning Calorimetry (DSC-60, Shimadzu, Columbia, USA). Briefly, 3 mg of SLN were placed on an aluminum pan. Thermal behavior was determined in the range 20e100  C at a heating rate of 10  C/min. Enthalpy values and optimal melting temperatures were calculated by the equipment software (ta60 version 2.10, DSC software, Shimadzu, Columbia, USA). The crystallinity indexes (CI%) of SLNs were calculated according to (Kheradmandnia et al. (2010)) using the following equation:

Melting enthalpy ðSLN dispersionÞðJ=gÞ  100 Melting enthalpy ðbulk material without RAÞðJ=gÞ: Concentration of lipid phaseð%Þ

sorbate), rosmarinic acid (RA) and the three cryoprotectants tested (glucose, trehalose and mannitol) were purchased from SigmaAldrich Chemistry (St. Louis, Missouri, USA). The extracts of sage and savory were produced as described by Campos et al. (2015). Peptone (Sigma-Aldrich Chemistry, St. Louis, Missouri, USA), Plate Count Agar (PCA) (Merck, Darmstadt, Germany), Rose Bengal agar (RBA) and chloramphenicol supplement (Lab M, Lancashire, UK) were used for the microbiological analyses. Folin-Ciocalteu reagent (Merck, Darmstadt, Germany), sodium carbonate and gallic acid (Sigma-Aldrich Chemistry, St. Louis, Missouri, USA) were used for the evaluation of total phenolic compounds. 2.2. Production of solid lipid nanoparticles The Witepsol SLNs were produced in duplicate, according to Campos et al. (2014) and the Carnauba wax SLNs were produced according to Madureira et al. (2015). Briefly, SLNs were prepared with RA at the final concentration of 0.15 mg/mL by the hot melt ultrasonication method. Lipid matrices were used at 1% (w/v) and the surfactant at 2% (v/v). 2.3. Evaluation of the freeze-drying ideal cryoprotectant In order to optimize the SLNs freeze-drying process, studies were performed in duplicate only for RA-SLNs. Three different cryoprotectants, namely glucose, mannitol and trehalose were studied at two different concentrations, i.e. 5 and 10% (v/v). The freeze-drying process was performed using a vacuum freeze drier (Model FT33, Armefield, UK), under a vacuum pressure of 100 militorr; the temperature in the freezing chamber was
46  C and the temperature in the sample chamber was 15  C. Samples without cryorprotectant were also processed as control. Freezedried SLNs were evaluated for their visual appearance, viscosity and free powder, as well as, physical and chemical stability. 2.3.1. Particle size and zeta potential analyses The physical properties of the two types of SLNs were assessed.

(1)

All raw materials used in the formulations were evaluated individually in triplicate and in combination. 2.4. Assessment of solid lipid nanoparticles long-term stability Evaluation of long-term stability studies was performed on SLNs loaded with RA and herbal extracts (sage and savory). A design study to measure the chemical and thermodynamic stability in different physical states was performed. Thus, during 1 year, samples were taken at 0, 15, 30, 90, 180 and 365 days, however only the time points where the major differences were found are shown. In Table 1 are described the different conditions tested for SLNs storage, which generated three studies (A, B and C). Different features were evaluated such as, antioxidant capacities, percentage of encapsulated RA, RA release percentage, SLN powder color, water activity and microbiological analyses to evaluate the occurrence of eventual contaminations. 2.4.1. SLNs color SLNs color was evaluated with a portable CR-400 Chroma Meter (Minolta, Osaka, Japan). The CIELab color scale was used to determine the lightness (L), redness (þa*)/greenness (
a*) and yellowness (þb*)/blueness (
b*) of the samples. SLNs samples were

Table 1 Experimental design of the processing parameters applied to SLNs. Storage conditions

Packaging material Temperature Light Atmosphere

Physical state

Plastic Glass ±25  C (Room) ±4  C (Refrigerated) Presence Absence Room Nitrogen

A (Liquid)

B (Solid)

✓

✓ ✓

C (Solid) ✓ ✓

✓ ✓ ✓ ✓

✓ ✓ ✓

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10

measured in on the surface of a white standard plate with color coordinates L ¼ 97.59, a ¼
0.07 and b ¼ 1.89. SLNs colors were expressed as the total difference in color, DE, calculated with the formula:

DE ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðLSLNs
Lcontrol Þ2 þ ðaSLNs
acontrol Þ2 þ ðbSLNs
bcontrol Þ2 (2)

For each sample, three sections were measured and three readings were taken for each. 2.4.2. Water activity The water activity (aw) was measured using a HygroLab 2 (from Rotronic, Bassersdrof, Germany) at room temperature (25 ± 1  C). Freeze-dried SLNs (ca. 0.5 g) were placed on the sample holder of the water activity device; a sealed system was formed by placing the water activity probe on top of the sample holder. The probe was equipped with a small fan to circulate air inside the sample container, a thin film capacitance sensor able to measure RH from 0 to 100 ± 1.5%, and a platinum resistance temperature detector with a precision of ±0.3  C. When aw became constant (which usually took less than 1 h), its value was recorded. Calibration resorted to six saturated solutions of known aw (viz. LiCl ¼ 0.114, MgCl2 ¼ 0.329, K2CO3 ¼ 0.443, Mg(NO3)2 ¼ 0.536, NaBr ¼ 0.653 and KCl ¼ 0.821). The tests were run in triplicate. 2.4.3. Microbiological analysis Microbiological analysis of the samples was performed to assess possible contaminations throughout storage. Samples were taken every 90 days (0, 90, 180, 270 and 365 d) throughout storage. Briefly, 5 g of each sample were weight and diluted in 45 g of sterile peptone water, the mixtures were homogenized for 3 min at 260 rpm in a Stomacher Lab Blender 400 (Sewer Medical, London, UK). All samples suffered serial decimal dilutions with peptone water (0.1%, w/v) and were plated in duplicate on two different media, namely Plate Count Agar (PCA) for mesophilic cell counts, incubated at 37  C for 24 h and Rose Bengal Agar (RBA) with chloramphenicol for yeasts/fungi contaminations, incubated at 30  C for 7 days. Colony counts were performed and the decision to end the experiment was dependent on the appearance of sample contamination. 2.5. Quantification of total polyphenols 2.5.1. Samples preparation Samples of 70 mg/mL of freeze-dried SLNs were rehydrated using phosphate buffer solution (PBS; 0.1 M, pH 7.0). To evaluate and quantify the total polyphenols, the samples were filtered by cut-off of 3 kDa (Amicon® Ultra-4, Millipore; Billerica, MA, USA) by centrifugation at 3000 rpm, at 4  C. The resulting supernatant was removed and analyzed for free RA concentration. 2.5.2. HPLC The quantification of the polyphenols release of the samples for the stability studies was performed using high performance liquid chromatography (HPLC). The chromatographic analysis was performed using a Beckman & Coulter 168 series HPLC system interfaced with a Photo Diode Array UV/Vis detector (PDA 190e600 nm) (Beckman & Coulter; Fullerton, California, USA). Separation was done in a reverse phase column coupled with a guard column containing the same stationary phase (COSMOSIL 5C18-AR-II Packed Column - 4.6 mmI.D.  250 mm; Dartford, UK). Chromatographic separation of phenolic compounds was carried out

3

with mobile phase A - water, methanol and formic acid (92.5:5:2.5) - and mobile phase B e methanol, water and formic acid (92.5:5:2.5) - under the following conditions: gradient elution starts at 40% mobile phase B until 2 min at a continuous flow of 0.5 mL/min, between 2 and 17 min with 50% of mobile phase B, lastly between 17 and 25 min the gradient change with 80% of mobile phase B. Data acquisition and analysis were accomplished using Karat32 software. Detection was performed at wavelengths ranging from 280 to 320 nm. The peaks were obtained at 320 nm (phenolic acids) and were analyzed by comparison of retention time and spectra with pure polyphenols. Quantification was performed using calibration curve of the pure RA. Three independent analyses were performed for each experiment. 2.5.3. Folin-Ciocalteu method The total polyphenol content of the collected samples from the release studies was measured using a modified Folin-Ciocalteu colorimetric method described previously by Gao et al. (2000). Aliquots of 50 mL of samples and control (distilled water) were mixed with 50 mL of Folin-Ciocalteu reagent (0.25 N), 1 mL of sodium carbonate (1 N) and 1.4 mL of distilled water, added in this exact order. The total polyphenol content was determined after 1 h of incubation at room temperature (25  C). The absorbance of the resulting blue color was measured at 750 nm by colorimetry using an UVmini 1240 UVeVis spectrophotometer (Shimadzu, Washington, Maryland, USA). A standard curve was performed using different concentrations of gallic acid. All measurements were performed in triplicate. 2.5.4. Antioxidant capacity - ABTS assay The total antioxidant activity was performed by the ABTS radical cation decolorization assay. For this method the free radicalscavenging capacity was determined as described by Re et al. (1999). ABTS (Sigma-Aldrich Chemistry, St. Louis, Missouri, USA) was dissolved in water at a final concentration of 7 mM. ABTS radical cation (ABTS$þ) was produced by reacting ABTS stock solution with 2.45 mM potassium persulfate (Merck, Darmstadt, Germany) (final concentration) and kept in the dark at room temperature (25 ± 2  C) for 12e16 h before use. The radical maintains a stable form for more than two days when stored in the dark. Before analysis, ABTS$þ was filtered using a 0.22 mm (Orange Scientific, Braine-l’Alleud, Belgium) and diluted with redistilled water to an absorbance of 0.700 (±0.02) at 734 nm with an UVmini 1240 UVeVis spectrophotometer (Shimadzu, Washington, Maryland, USA). A blank was taken with distilled water (A0). Different volumes were tested. After addition of 1.0 mL of diluted ABTS$þ solution (A734 nm ¼ 0.700 ± 0.02) to 10 mL of sample, the absorbance was read exactly 6 min after initial mixture. Since the inhibition percentage (IP) must be limited to a range between 20 and 80%. Samples were diluted whenever necessary. Inhibition % values were calculated using the equation below and then converted into g/L of ascorbic acid equivalent, through a calibration curve prepared using standard solutions of ascorbic acid (Sigma-Aldrich Chemistry, St. Louis, Missouri, USA). All assays were performed in triplicate, considering three different replicates of the analyzed sample.

IP ¼

  OD diluted ABTS$þ
OD sample    100 OD diluted ABTS$þ

(3)

2.6. Statistical analysis All experimented results were analyzed by two-way analysis of

4

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10

a)

b)

d)

c)

Fig. 1. Freeze-dried RA-SLNs without cryoprotectant. a)b) WitepsolSLNs and c)d) Carnauba wax SLNs.

a)

c)

b)

Fig. 2. Freeze-dried RA-SLNs a) without cryoprotectant, b) with mannitol at 5% (w/v) and c) with mannitol at 10% (w/v).

variance (ANOVA), using Bonferroni’s post-hoc test at a significance level of 5%; GraphPad Prism, v. 500 software was used for this purpose (GraphPad Software, San Diego, CA, USA). 3. Results and discussion 3.1. Effects of freeze-drying process on SLNs and selection of best conditions The evaluation of freeze-dried SLNs evaluation involved macroscopic characterization, reconstitution time, measurements of PS, PI and ZP after freeze-drying. A critical analysis of freezedried products included the observation of the final volume and the appearance of the resulting cake. One of the desired characteristics of a freeze-dried product is that the resulting cake must remain intact occupying the same volume as the original frozen mass. However, these characteristics were not observed for the freeze-dried samples without cryoprotectant addition. The cake collapsed and the samples became darker, sticky and melted instantly when handled, as can be seen in Fig. 1.

To overcome this problem, three different cryoprotectants, namely glucose, mannitol and trehalose at two different concentrations, i.e. 5 and 10% (v/v), were added to the liquid SLNs and submitted to freeze-drying. Once more, the visual appearance, viscosity and free powder of SLNs were evaluated. In general, all tested formulations containing the different cryoprotectants presented free white powders and absence of viscosity, as can be seen in Fig. 2 for both type of SLNs with mannitol at 5 and 10% (w/v). Reconstitution time can be the first indicator of the stability of freeze-dried SLNs, since this parameter can give information on their stability, i.e. a long reconstitution time is associated with low stability and collapsed samples. In this study, the samples were reconstituted instantaneously in PBS (0.1 M, pH 7.0) at a final concentration of 70 mg/mL, which may indicate good stability according to Abdelwahed et al. (2006b). The physical properties of the freeze-dried SLNs with the three cryoprotectants at two concentrations (Table 2) were evaluated by DLS. Particle size measurement gives an indication of the diameter of the nanoparticles (NPs). In general, when comparing the SLNs without cryoprotectant with those freeze-dried with

Table 2 Mean values ± SEM of particle size (PS), polydispersity index (PI) and zeta potential (ZP) of Witepsol and Carnauba waxes SLNs, loaded with RA and freeze-dried with the different cryoprotectants (glucose, mannitol and trehalose) at different concentrations (5 or 10%). Cryoprotectant Witepsol RA-SLNs

Carnauba wax RA-SLNs

Content (%) Without cryoprotectant Glucose 5 10 Mannitol 5 10 Trehalose 5 10 Without cryoprotectant Glucose 5 10 Mannitol 5 10 Trehalose 5 10

PS (nm)

PI

1303 ± 85 5638 ± 395 1888 ± 87 2342 ± 142a 1722 ± 55a 2116 ± 62 817 ± 7 582 ± 227 1241 ± 59 2216 ± 48 1229 ± 82 1689 ± 170 2082 ± 124 1501 ± 65

0.300 0.379 0.337 0.165 0.284 0.336 0.342 0.310 0.101 0.403 0.005 0.026 0.130 0.174

± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.058 0.007 a 0.013 a 0.036 0.024 0.008 a 0.009 a 0.046 0.022 0.004 0.000a 0.001a 0.030a 0.035a

ZP (mV)

pH


38.7 ± 1.73
21.70 ± 0.42
15.19 ± 0.17
23.75 ± 0.47
16.58 ± 0.87
32.79 ± 0.29
19.91 ± 0.37
38.7 ± 0.46 2.96 ± 0.34
15.04 ± 1.66
23.41 ± 0.59
16.37 ± 1.38
24.71 ± 1.46
21.65 ± 0.92a

6.86

6.86

a The differences between the means in the same row labelled with same superscript are not statistically significant (P > 0.05). Analysis of variance was used to estimate the effects of each cryoprotectant percentage in SLNs. Bonferroni’s test was used as post test. Values are expressed as average ± standard error.

DSC (mW)

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10

5

Temperature (°C)

20

30

40

A

50

60

B

70

C

D

80

90

E

F

100

Fig. 3. Thermograms of Witepsol SLNs with different cryoprotectants (w/v). (A) 5% glucose, (B) 10% glucose, (C) 5% mannitol, (D) 10% mannitol, (E) 5% trehalose, (F) 10% trehalose.

Table 3 Mean values ± SEM of enthalpy changes (DH), melting temperature values and crystallinity indexes (CI%) of Witepsol and Carnauba waxes SLNs, loaded with RA and freezedried with the different cryoprotectants (glucose, mannitol and trehalose) at different concentrations (5 or 10%). Thermal properties

Witepsol H15 RA-SLNs

Carnauba wax RA-SLNs

Glucose

DH (J/g) Melting T. ( C) CI% DH (J/g) Melting T. ( C) CI%

Mannitol

Trehalose

5% (w/v)

10% (w/v)

5% (w/v)

10% (w/v)

5% (w/v)

10% (w/v)


1.86 ± 0.01 37.79 2.63
3.69 ± 0.10 86.14 2.18


10.02 ± 0.00 38.56 14.19
6.97 ± 0.40 86.40 4.11


1.90 ± 0.03 34.76 2.69
1.35 ± 0.05 86.95 0.80


1.09 ± 0.01 36.34 1.54
1.84 ± 0.03 85.99 1.09


1.96 ± 0.00 35.93 2.78
6.40 ± 0.12 87.74 3.78


3.16 ± 0.02 37.47 4.47
3.69 ± 0.30 87.04 2.18

bonding between OH groups of the cryoprotectant and the surface of SLNs. This behavior shows a rearrangement of the emulsifier at the SLNs surface, with consequent decrease on ZP values (de Chasteigner et al., 1996). Specifically, for Witepsol RA-SLNs the formulations with the lowest PS were those produced with 10% (w/v) cryoprotectant, and in particular, with trehalose; the formulation with the smallest PI value was the formulation with 10% (w/v) mannitol, revealing a free white cake. Moreover, for Carnauba RA-SLNs the formulations with lower PS were those produced with 5% (w/v) glucose and mannitol and 10% (w/v) mannitol and trehalose. The choice of the best cryoprotectant was based on the best physical properties registered, as well as the desired macroscopic features. The formulations with the lowest PI values were those produced with 5% (w/v) mannitol, but when analyzing at the best macroscopic features, the

DSC (mW)

cryoprotectants, it is clear that the range of PS increased after reconstitution of the SLNs, which may indicate NP aggregation. Nevertheless, the sizes were homogeneous as showed by the PI values. Zeta potential values indicate the general charge of the SLNs at the surface. In what concerns this parameter, a decrease in the stability of SLNs was observed, reflected in the occurrence of changes from moderate (
20 mV) to bad stability (
10 mV). The obtained results were somehow the expected, since the cryoprotectants added to the matrices would interact with the emulsifier (Tween 80), changing them, reducing repulsion and promoting aggregation. de Chasteigner et al. (1996) obtained the same behavior when adding 10% of sucrose as a cryoprotectant, i.e. a decrease of the negative surface charge from
40.9 to
20.4 mV was also observed. This phenomenon can be explained by the fact that the surface of the SLNs was masked as a result of hydrogen

Temperature (°C)

20

30

40

A

50

B

60

C

70

D

80

E

90

100

F

Fig. 4. Thermograms of Carnauba wax SLNs with different cryoprotectants (w/v). (A) 5% glucose, (B) 10% glucose, (C) 5% mannitol, (D) 10% mannitol, (E) 5% trehalose, (F) 10% trehalose.

0.031 0.034 0.00b 0.072 0.05b 0.057 0.038 0.030 ± ± ± ± ± ± ± ± 0.236 0.062 0.371 0.204 0.221 0.062 0.115 0.077 ± ± ± ± ± ± ± ± 0.308 0.287 0.319 0.291 0.250 0.283 0.200 0.342 ± ± ± ± ± ± ± ± 0.352 0.256 0.301 0.132 0.359 0.312 0.377 0.154 ± ± ± ± ± ± ± ± 1968 1843 2202 1921 1829 1247 1501 1254 ± ± ± ± ± ± ± ± 0.353 0.332 0.308 0.351 0.294 0.277 0.137 0.179 31 12 11 50 31 59 28a 13 ± ± ± ± ± ± ± ± 845 802 684 884 551 812 832 661 569 ± 1 575 ± 9 592 ± 24 1181 ± 16 485 ± 7 1024 ± 46 818 ± 15a 718 ± 40 Carnauba wax

Witepsol

Empty RA Sage Savory Empty RA Sage Savory

90 days

0.015 0.015b 0.017 0.011 0.019 0.030 0.029b 0.061

0.275 0.270 0.143 0.150 0.199 0.158 0.163 0.257

± ± ± ± ± ± ± ±

0.023 0.016b 0.052 0.072 0.036 0.060 0.068b 0.039

0 days 90 days

SLNs

ab The means in the same row labelled with same superscript are not statistically different (P > 0.05). Analysis of variance was used to estimate the effect of each test (particle size and polydispersion index) in different storage conditions in SLNs. Bonferroni’s test was used as post tes. Values are expressed as average ± standard error.

0 days

59 81 56 68 72 70 79 86 PS PI

0 days

PS

757 ± 76 921 ± 67 896 ± 78 794 ± 127 894 ± 177 1015 ± 129 772 ± 135 898 ± 130

PI B (Solid state) A (Liquid state)

365 days

0 days

0.032 0.012b 0.006 0.038b 0.022 0.011b 0.023 0.055

0.229 0.243 0.156 0.191 0.276 0.307 0.224 0.268

± ± ± ± ± ± ± ±

0.016 0.044b 0.076 0.056b 0.046 0.014b 0.036 0.014

1841 1586 1718 1707 1203 1708 1607 2278

± ± ± ± ± ± ± ±

100 158 119a 185a 101a 399 104 155

807 ± 47 1071 ± 95 1500 ± 317a 1619 ± 185a 1215 ± 343a 1206 ± 296 806 ± 132 1080 ± 237

PI PS

0 days 365 days

C (Solid state)

365 days

0 days

0.046 0.022 0.01b 0.037 0.04b 0.019 0.055 0.026

365 days

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10 Table 4 Mean values ± SEM of particles size (PS) and polydispersity index (PI) of Witepsol and Carnauba waxes SLNs, loaded with RA, extract of sage and extract of savory over 90 days for experimental design A and over 365 days for experimental designs B and C.

6

powder obtained by this formulation was slightly stickier (Fig. 2, b), than the formulation with 10% (w/v) mannitol, resulting in a cake of free white powder (Fig. 2, c). For these specific reasons, the cryoprotectant chosen for the follow-up study was mannitol at a concentration of 10%. Other authors also selected mannitol at 10% as cryoprotector, obtaining stable ibuprofen-SLNs (Zhang et al., 2008). Soares et al. (2013) have described the use of different cryoprotectans for the freeze-drying of insulin-SLNs and the results showed no difference between the freeze-drying with and without cryoprotectants, demonstrating the maintenance of physical characteristics (PS, PI and ZP). 3.2. Effect of cryoprotectants on the SLN thermal properties The thermodynamic profile study was performed using DSC. Both RA and tween 80 (polysorbate 80) didn’t present any peaks for the tested temperatures as in agreement with Campos et al. (2014). On the other hand, the SLNs, Witepsol and Carnauba wax lipids, presented endothermic peaks (41.15  C;
70.63 J/g and 92.39  C;
169.51 J/g) (Campos et al., 2014; Madureira et al., 2015), which indicates the occurrence of polymorphisms at the lipid structure by energy absorption. The thermal behavior of both lipidic matrices was evaluated for the different cryoprotectants at the different concentrations (Figs. 3 and 4). The study on Witepsol SLNs using different cryoprotectants presented temperature values ranging between 34.7 and 38.6  C (Table 3). The values were slightly lower than the melting point of control samples ca. 41.1  C without cryoprotectants (Campos et al., 2014). The same behavior was observed for Carnauba wax SLNs; the melting temperatures were similar for all the formulations, between 86.0 and 87.4  C (Table 3). Again, these values were lower than those for Carnauba wax alone (92.4  C) (Madureira et al., 2015). The reduction in the melting temperature value is mainly related to the nanocrystalline size of lipids in SLN systems as described elsewhere (Jee et al., 2006; Westesen and Bunjes, 1995). Thus, the reduction of these values is due to the presence of NPs and not directly connected to the presence of cryoprotectants. On the other hand, the enthalpy values of the Witepsol SLNs containing cryoprotectants decreases; the melting enthalpy values ranged between
10.0 and
1.09 J/g against that of the standard formulation
18.0 J/g indicating that the energy necessary to melt the SLNs was lower, as can be seen in Fig. 3. The same behavior was detected for Carnauba wax SLNs, with melting enthalpies values between
6.97 and
1.35 J/g against
21 J/g for the formulation without cryoprotectant (Fig. 4). It seems that the cryoprotectants created a kind of a cap at the SLNs surface with no interaction with the lipid structure. This was shown by the decrease in enthalpy values (set of all compounds at the sample), but differences were not visualized for the melting temperatures (Rosa, 2011). The percentage of crystallinity (CI%) can be calculated and allows understanding of the thermal behavior of the materials (equation (1)). High values of CI% lead to a faster compound release, yet this means that more energy is required to melt the crystal lipids. Thus, an ideal value of CI% should be near 50%, sufficiently high to make sure that most of the SLNs are stable when forming new particles, but sufficiently low to ensure the release of RA. The values of CI% are listed in Table 3. A decrease of CI% values was observed when compared to the standard formulations (49.5% Witepsol SLNs; 58.6% Carnauba SLNs), demonstrating a lower stability. It was an expected result, since these CI% are directly affected by the enthalpy values. The formation of hydrogen bonds between the cryoprotectant (mannitol) and the polar groups of the polysorbate 80 at the surface of the SLNs, preserves its native structure by serving as water substitutes during the freeze-drying process

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10

7

Table 5 Mean values ± SEM obtained with the ABTS assay, Folin-Ciocalteau method and HPLC method for the Witepsol and Carnauba waxes SLNs, loaded with RA, extract of sage and extract of savory, expressed in percentage (100% corresponds to the total RA added). Study on the storage of SLNs in study A (liquid state; plastic containers, absence of light, presence of oxygen, at 4  C), during 90 days. SLNs

Storage time 0 days

Witepsol

Carnauba wax

Empty RA Sage Savory Empty RA Sage Savory

90 days

ABTS

Folin-Ciocalteau

HPLC

ABTS

0.000 ± 0.000 29.935 ± 0.840 6.008 ± 1.780 7.477 ± 0.297 0.000 ± 0.000 22.247 ± 5.771 6.859 ± 1.984 12.313 ± 1.893

0.000 ± 0.000 21.168 ± 0.865 2.820 ± 0.216 1.198 ± 0.530 0.000 ± 0.000 24.788 ± 2.085 2.509 ± 0.265 3.819 ± 0.375

0.000 ± 0.000 24.939 ± 0.148 9.883 ± 0.149 7.834 ± 0.007 0.000 ± 0.000 24.156 ± 0.148 8.808 ± 0.078 8.039 ± 0.006

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

Folin-Ciocalteau ± ± ± ± ± ± ± ±

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

± ± ± ± ± ± ± ±

0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

HPLC 0.000 6.638 6.774 8.075 0.000 6.162 6.631 0.000

± ± ± ± ± ± ± ±

0.000 0.095 0.047 0.061 0.000 0.071 0.032 0.000

The differences between the means in the same row are statistically significant (P < 0.05). Analysis of variance was used to estimate the effect of each test (ABTS assay, FolinCiocalteau method and HPLC method) in beginning and ending of SLNs storage experiment. Bonferroni’s test was used as post-test. Values are expressed as average ± standard error.

(Abdelwahed et al., 2006a). However, weaker interactions require less energy to break the connections and therefore a lower enthalpy value and CI% are reported. 3.3. Effect of storage conditions on the stability of SLNs and selection of the best conditions The aim of this study was to evaluate the SLNs shelf-life. For the cases tested, storage was ended when microbiological contaminations were registered. Study A was performed in SLNs at liquid state, stored in plastic bottles, protected from light and refrigerated at 4  C. At 180 days, contamination with fungi colonies was found thus placing the shelf-life limit at 90 days. In Table 4 are described the values of PS and PI, for 0 and 90 days. Statistical significant differences were found fro PS of SLNs at 0 and 90 days of storage (P < 0.05), although defined trend was observed since in some cases the PS increased while in others decreased. On the other hand, the PI values followed a different behavior, the initial stable values (0.3) of polidispersity decreased after 90 days, increasing the monodispersity of the

samples. Significant statistical differences were found for all samples along storage time, with exception of Witepsol SLNs loaded with RA and Carnauba SLNs loaded with sage. In the specific case of Carnauba SLNs loaded with sage, this formulation maintained the PS along the 90 days and size homogeneity (P > 0.05). These observations were in agreement with Müller et al. (2000) who that described the same behavior; low PI values indicate higher stability and therefore maintenance of the initial and good values, showing higher stability through storage time. The results obtained for ABTS (antioxidant activity), FolinCiocalteau (total phenolic compounds) and HPLC method (RA release) for all tested formulations are shown at Table 5. Witepsol RA-SLNs at 0 days showed a RA release of ca. 25% and Carnauba RASLNs of ca. 24%. The SLNs containing herbal extracts (sage and savory) showed release of ca. 10% of RA, which may probably be the RA that was not encapsulated during the SLN production. After 15 days, RA almost disappeared (data not shown), and this result was maintained until the end of study (90 days, Table 5). At the beginning of the experiment (0 days), the total phenolic

Table 6 Mean values ± SEM obtained with ABTS assay, Folin-Ciocalteau method and HPLC method for the Witepsol and Carnauba waxes SLNs, loaded with RA, extract of sage and extract of savory, expressed in percentage (100% corresponds to the total RA added). Study on the storage of SLNs in study B (solid state; plastic containers, presence of light and oxygen, at room temperature), during 365 days. SLNs

Storage time 0 days

Witepsol

Carnauba wax

180 days

365 days

ABTS

Folin-Ciocalteau

HPLC

ABTS

Folin-Ciocalteau

HPLC

ABTS

FolinCiocalteau

HPLC

Empty

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

RA

15.786 ± 0.140a

21.169 ± 0.499a

56.939 ± 0.101

13.964 ± 0.554a

23.249 ± 0.042a

60.563 ± 0.221

Sage

16.905 ± 0.700

2.821 ± 0.125

13.549 ± 0.212

9.808 ± 0.367

9.769 ± 0.220

50.222 ± 0.158

Savory

19.144 ± 0.979

0.574 ± 0.330

11.569 ± 0.159

5.790 ± 0.720

3.736 ± 0.440

67.851 ± 0.143

Empty

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

RA

19.004 ± 1.145a

7.189 ± 0.572

77.270 ± 0.803

14.934 ± 0.416

0.000 ± 0.000 68.516 ± 0.989 27.394 ± 0.374 26.404 ± 2.194 0.000 ± 0.000 78.125 ± 5.054 29.513 ± 0.141 21.176 ± 0.245a

0.000 ± 0.000 90.400 ± 0.363 28.200 ± 0.220 29.656 ± 0.374 0.000 ± 0.000 96.877 ± 3.123 30.946 ± 0.166 20.503 ± 0.166

0.000 ± 0.000 89.554 ± 0.472 86.394 ± 0.251 73.895 ± 0.628 0.000 ± 0.000 97.137 ± 1.021 80.050 ± 0.456 78.209 ± 0.832

a

a

24.913 ± 1.191

88.247 ± 0.652

a

20.237 ± 0.667

Sage

16.625 ± 0.140

3.320 ± 0.818

16.615 ± 0.387

5.235 ± 0.277

4.901 ± 0.220

Savory

18.024 ± 0.840a

3.819 ± 0.216a

10.681 ± 0.321a

6.067 ± 1.736

2.488 ± 0.110a

11.942 ± 0.239a

a The means in the same row labelled with same superscript are not statistically different (P > 0.05). Analysis of variance was used to estimate the effect of each test (ABTS assay, Folin-Ciocalteau method and HPLC method) in beginning and ending of SLNs storage experiment. Bonferroni’s test was used as post test. Values are expressed as average ± standard error.

8

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10

Table 7 Mean values ± SEM obtained with ABTS assay, Folin-Ciocalteau method and HPLC method for the Witepsol and Carnauba waxes SLNs, loaded with RA, extract of sage and extract of savory, expressed in percentage (100% corresponds to the total RA added). Study on the storage of SLNs in study C (solid state; glass containers, absence of light and oxygen, at room temperature), during 365 days. SLNs

Storage time 0 days

Witepsol

365 days

Folin-Ciocalteau

HPLC

ABTS

Folin-Ciocalteau

HPLC

ABTS

FolinCiocalteau

HPLC

Empty

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

RA

15.786 ± 0.140a

21.169 ± 0.499a

52.056 ± 0.382

14.380 ± 0.843a

23.166 ± 0.125a

2.810 ± 0.135

7.952 ± 0.142

17.248 ± 0.085

9.946 ± 0.240

9.686 ± 0.250

0.388 ± 0.011

2.828 ± 0.377

0.000 ± 0.000 9.561 ± 0.900 0.699 ± 0.250a 0.000 ± 0.000a 0.000 ± 0.000 0.000 ± 0.000 5.941 ± 0.499a 0.000 ± 0.000b

0.000 0.000 1.255 0.071 0.202 0.010 0.000 0.000 0.000 0.000 0.176 0.000 0.000 0.000 0.000 0.000

Sage

Carnauba wax

180 days

ABTS

16.905 ± 0.700

2.821 ± 0.125

a

a

4.502 ± 0.153

5.097 ± 0.367

0.000 ± 0.000

0.000 ± 0.000

7.189 ± 0.572

29.551 ± 0.461

Savory

19.144 ± 0.979

0.574 ± 0.330

Empty

0.000 ± 0.000

RA

19.004 ± 1.145a

a

3.570 ± 0.499

0.000 ± 0.000

2.117 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

0.000 ± 0.000

14.657 ± 0.604a

24.788 ± 1.204

0.262 ± 0.026

9.660 ± 0.620

a

0.000 ± 0.000

10.514 ± 0.142

0.000 ± 0.000

2.117 ± 0.247

a

13.494 ± 0.522

4.404 ± 0.554

4.818 ± 0.250

10.360 ± 0.281

7.175 ± 1.447

2.446 ± 0.125ab

Sage

16.625 ± 0.140

3.320 ± 0.818

Savory

18.024 ± 0.840

3.819 ± 0.216a

a

± ± ± ± ± ± ± ±

ab The means in the same row labelled with same superscript are not statistically different (P > 0.05). Analysis of variance was used to estimate the effect of each test (ABTS assay, Folin-Ciocalteau method and HPLC method) in beginning and ending of SLNs storage experiment. Bonferroni’s test was used as post-test. Values are expressed as average ± standard error.

compounds and antioxidant capacity presented the same range of values however upon 90 days (end of experiment) no release of phenolic compounds or antioxidant capacity was detected (P < 0.05). These observations confirm that the initial RA values detected referred to the non-encapsulated form. The maintenance of the study demonstrated the typical degradation of these sensitive compounds, when not encapsulated. The stability of the SLNs systems tested was maintained for 90 days in liquid state, since RA concentration did not increase during this period. In contrast with our study, Howard et al. (2012) reported a high drug loss in SLNs loaded with dexamethasone palmitate during aqueous suspension storage. Experiments B and C were performed using SLNs in solid state, stored at room temperature, using different types of packages and atmospheres. The PS and PI results are depicted in Table 4, for 0, 180 and 365 days. The results of ABTS, Folin-Ciocalteau and HPLC methods are shown in Tables 6 and 7. Microbiological analyses were performed every three months up to one year. In both cases, no microbiological contaminations were detected until the end of the experiment. The water activity of the solid state samples was measured, and all samples showed values in the range of aw ¼ 0.6 which were maintained until the end of experiment, meaning that the SLNs did not absorb water from the air. As established in order to occur microbiological proliferation, the samples should present

values higher than aw ¼ 0.8. The color of the freeze-dried SLNs was also measured throughout storage time. The CIELab detected differences in SLNs color evolution and the results are displayed in Table 8. The parameters evaluated did not show differences between 0 and 365 days (values non-statistically different P > 0.05). All three parameters L, a* and b* were affected by the addition of the RA and/or the herbal extracts. Indeed, the addition of RA showed an increase in the SLNs lightness, but the herbal extracts decreased the lightness of the lipidic matrices, and a shift in color towards green and yellow, i.e. negative and smaller a* values and positive and slightly higher b* values, respectively. Negatives controls were performed with empty SLNs of Witepsol and Carnauba wax. The total difference in color, DE, were completely different between samples, showing values with statistical differences (P < 0.05). In study B, the solid state SLNs were stored in transparent plastic containers (flask used for urine analysis), in presence of light and oxygen. In all tested formulations, the PS decreased over storage time (P < 0.05), and the PI presented values lower than 0.3 showing high stability. In general, PI values were slightly higher than 0.3 at 0 days but decreased along storage time (P < 0.05), showing monodispersity of SLNs sizes. The initial concentration of RA found was ca. 57% for Witepsol RA-SLNs and ca. 77% for Carnauba wax RA-SLNs (Table 6). In what

Table 8 Mean values ± SEM of CIELab coordinates (Witepsol and Carnauba waxes SLNs, loaded with RA, extract of sage and extract of savory). SLNs

Samples color Parameters (0 days) L

Witepsol

Carnauba wax

Empty RA Sage Savory Empty RA Sage Savory

67.91 92.06 63.18 66.46 95.83 93.70 63.19 72.43

a* ± ± ± ± ± ± ± ±

0.84 0.29 0.34 0.27 0.23 0.33 0.47 0.17


0.04
0.29
0.49
1.00
0.63
0.04
0.17
1.21

DE

b* ± ± ± ± ± ± ± ±

0.00 0.02 0.01 0.02 0.01 0.10 0.01 0.00

0.24 5.40 8.82 7.56 3.02 5.17 8.05 6.87

± ± ± ± ± ± ± ±

0.01 0.13 0.01 0.01 0.13 0.17 0.04 0.00

e 24.70 ± 0.58 9.91 ± 0.60 7.55 ± 0.11 e 3.09 ± 0.32 33.03 ± 0.58 23.72 ± 0.19

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10

concerns, the herbal extracts loaded Witepsol and Carnauba SLNs, 11% and 17% of RA was released, respectively. At 180 days, a further 10% of RA was released from both lipidic matrices (P < 0.05). The same behavior was observed for herbal extracts SLNs, Witepsol SLNs reached values of ca. 60% (P < 0.05), and whereas sage and savory Carnauba-SLNs presented RA values of ca. 20% and 12%, respectively. At this point it was possible to predict, that the SLNs systems were not stable, probably due to the SLNs degradation that may have occurred. At the end of experiment, a high percentage of RA was detected. The Witepsol RA-SLNs released ca. 90% and the Carnauba RA-SLNs showed values ca. 97% (P < 0.05), which means, that after one year of storage the SLNs system became completely unstable. The herbal extracts SLNs showed release of ca. 80% for both lipidic matrices. The Folin-Ciocalteau method and ABTS assay performed on all samples presented similar results, but lower percentages when compared with the HPLC quantification. Such observation is somehow expected given the fact that HPLC method is an analytical, more sensitive and precise method for detection of phenolic compounds than the colorimetric methods Folin-Ciocalteau. In study C, the freeze-dried SLNs were stored at room temperature, in glass flasks, protected from light (absence of light) and in a modified atmosphere (nitrogen atmosphere). In this study, PS values decreased throughout storage time as reported in study B, but the differences between both sampling points are not as evident as registered in study B, inclusively some formulations the differences are not statistically significant. In general, all the PI initial mean values were 0.3 showing good dispersion of formulations and a decrease up to 365 days was observed. Statistical analysis showed that most samples were statistically different (P < 0.05), with exception of Witepsol SLNs loaded with sage extract and Carnauba empty-SLNs where no statistically differences were found over storage (P > 0.05), being these the most stable formulations. In what concerns the release of RA (Table 7), at 0 days Witepsol RA-SLNs demonstrated a release of ca. 52% and Carnauba RA-SLNs showed ca. 30%. Herbal extract SLNs showed lower RA percentage release between 5 and 17%. After 180 days, the RA concentrations decreased for all samples. Witepsol RA-SLNs reduced to ca. 2.8%, Carnauba wax RA-SLNs and herbal extracts SLNs did not present any RA. The percentage release of RA was maintained between 180 and 365 days, which demonstrates the SLNs stability during the storage period (P > 0.05). Once again the quantification results by the Folin-Ciocalteau method were different when compared with the results obtained by HPLC. At time 0 days, Witepsol RA-SLNs presented ca. 21% and Carnauba wax RA-SLNs ca. 7%. The herbal extracts for both lipids showed values between 1 and 4%. At 180 days of storage the RASLNs reached ca. 25% and SLNs of herbal extracts showed values between 3 and 10%. The ABTS values were in accordance with those of the Folin-Ciocalteau method. Differences were found between the colorimetric method and the HPLC results for the measurements taken at 180 days and this could be a consequence of free RA degradation. When RA suffers degradation it produces smaller metabolites, such as caffeic acid, that have antioxidant capacity. Caffeic acid was not detected by the HPLC method, since the method had been optimized to detect RA (Campos et al., 2014). Nevertheless, the Folin-Ciocalteau method detected the total phenolic compounds and the ABTS assay, the antioxidant capacity. For the samples containing herbal extracts the same behavior was observed. These extracts contain RA, but also clorogenic acid, vanillic acid, ferrulic acid, rutin and others that can be detected by the total polyphenolic content determination and associated antioxidant capacity (Campos et al., 2015). Studies B and C were statistically compared in order to choose

9

the best storage method. The analysis of variance was used to estimate the effect of each storage condition at all times assayed (0, 180 and 365 days). Between 0 and 180 days the mean values were not statistically different (P > 0.05), which means that SLNs showed similar behavior at both tested storage conditions. But when compared for 365 days, all samples were statistically different (P < 0.05), which indicates different behaviors beyond 180 d storage period, distancing from each other as time passes by. The study C showed lower percentages release of RA for 365 days storage, indicating higher stability through the storage time. The storage materials and modified atmosphere used were different between studies B and C, and were the responsible factors for the SLNs stability. The presence of oxygen could be directly correlated to the degradation of the lipidic matrices and RA release since the lipid crystal is affected by oxidation. Olbrich et al. (2002) has already studied the effect of storage conditions on SLNs using Dynasan 116 and reported that the storage time also affects SLNs crystallinity and consequently increases degradation velocity. A comparative analysis of the three studies (A, B and C) reveals that the less stable storage system was that assayed in study A. These matrices in liquid state show a higher degradability rate and consequently less stability. Other authors (Olbrich et al., 2002) have shown higher degradation of the lipid crystal when stored in liquid state at 4  C. No differences were found in both lipidic matrices tested in study C (the most stable storage package) and submitted to longterm storage for the last sampling point (365 days) (P > 0.05). Although the lipidic matrices presented different compositions this was an expected result since the CI% of the tested NPs revealed similar values after addiction of cryoprotectants and application of freeze-drying process. 4. Conclusion A study on the freeze-drying process of SLNs loaded with RA and herbal extracts of sage and savory was performed in order to improve drying and handling of the nanocarriers formulations, considering a future use industrial scale-up. Different cryoprotectants at different concentrations were tested. Mannitol at 10% (w/v) was defined as the most suitable cryoprotectant. A study on storage shelf-life and storage stability was also performed over a 365 days. Storage stability studies showed that SLNs in liquid state maintained all features until 90 days of storage. When studying the stability of solid systems, two types of packaging systems were evaluated, but only one maintained SLNs stable throughout storage, i.e. glass bottles with N2 atmosphere. In conclusion, SLNs loaded with RA and herbal extracts proved to be a stable system at the encapsulation of bioactive compounds, as well as an active and stable delivery system with long-term stability. Acknowledgments ~o para a Cie ^ncia e a The authors acknowledge FCT (Fundaça Tecnologia) for funding: project Nanodairy (PTDC/AGR-ALI/117808/ 2010), UID/Multi/50016/2013, and PTDC/AGR-TEC/2227/2012. This work was also financed by the European Regional Development Fund (ERDF) through the Programa Operacional Factores de Competitividade e COMPETE, by Portuguese funds through FCT, in the framework of the project PEst-C/SAU/LA0002/2013, and cofinanced by the North Portugal Regional Operational Program (ON.2 e O Novo Norte) in the framework of project SAESCTNPIIC&DT/2011, under the National Strategic Reference Framework (NSRF). Ana Raquel Madureira acknowledges FCT for the postdoctoral scholarship SFRH/BPD/71391/2010.

10

D.A. Campos et al. / Journal of Food Engineering 196 (2017) 1e10

References Abdelwahed, W., Degobert, G., Fessi, H., 2006a. Investigation of nanocapsules stabilization by amorphous excipients during freeze-drying and storage. Eur. J. Pharm. Biopharm. 63 (2), 87e94. Abdelwahed, W., Degobert, G., Fessi, H., 2006b. A pilot study of freeze drying of poly (epsilon-caprolactone) nanocapsules stabilized by poly (vinyl alcohol): formulation and process optimization. Int. J. Pharm. 309 (1), 178e188. Blasi, P., Giovagnoli, S., Schoubben, A., Ricci, M., Rossi, C., 2007. Solid lipid nanoparticles for targeted brain drug delivery. Adv. Drug Deliv. Rev. 59 (6), 454e477. Campos, D.A., Madureira, A.R., Gomes, A.M., Sarmento, B., Pintado, M.M., 2014. Optimization of the production of solid Witepsol nanoparticles loaded with rosmarinic acid. Coll. Surf. B Biointerfaces 115, 109e117. Campos, D.A., Madureira, A.R., Sarmento, B., Gomes, A.M., Pintado, M.M., 2015. Stability of bioactive solid lipid nanoparticles loaded with herbal extracts when exposed to simulated gastrointestinal tract conditions. Food Res. Int. 78, 131e140. Chacon, M., Molpeceres, J., Berges, L., Guzman, M., Aberturas, M., 1999. Stability and freeze-drying of cyclosporine loaded poly (D, L lactideeglycolide) carriers. Eur. J. Pharm. Sci. 8 (2), 99e107.
, G., Fessi, H., Devissaguet, J.P., Puisieux, F., 1996. Freezede Chasteigner, S., Cave drying of itraconazole-loaded nanosphere suspensions: a feasibility study. Drug Dev. Res. 38 (2), 116e124. Franks, F., 2013. Water a Comprehensive Treatise. In: Aqueous Solutions of Amphiphiles and Macromolecules, Vol. 4. Springer Science & Business Media. €rk, L., Trajkovski, V., 2000. Changes in antiGao, X., Ohlander, M., Jeppsson, N., Bjo oxidant effects and their relationship to phytonutrients in fruits of sea buckthorn (Hippophae rhamnoides L.) during maturation. J. Agric. Food Chem. 48 (5), 1485e1490. Howard, M.D., Lu, X., Jay, M., Dziubla, T.D., 2012. Optimization of the lyophilization process for long-term stability of solidelipid nanoparticles. Drug Dev. Ind. Pharm. 38 (10), 1270e1279. Jee, J.-P., Lim, S.-J., Park, J.-S., Kim, C.-K., 2006. Stabilization of all-trans retinol by loading lipophilic antioxidants in solid lipid nanoparticles. Eur. J. Pharm. Biopharm. 63 (2), 134e139. Kheradmandnia, S., Vasheghani-Farahani, E., Nosrati, M., Atyabi, F., 2010. Preparation and characterization of ketoprofen-loaded solid lipid nanoparticles made from beeswax and carnauba wax. Nanomedicine Nanotechnol. Biol. Med. 6 (6),

753e759. Madureira, A.R., Campos, D.A., Fonte, P., Nunes, S., Reis, F., Gomes, A.M., Sarmento, B., Pintado, M.M., 2015. Characterization of solid lipid nanoparticles produced with carnauba wax for rosmarinic acid oral delivery. RSC Adv. 5 (29), 22665e22673. Müller, R.H., M€ ader, K., Gohla, S., 2000. Solid lipid nanoparticles (SLN) for controlled drug delivery e a review of the state of the art. Eur. J. Pharm. Biopharm. 50 (1), 161e177. Nunes, S., Madureira, R., Campos, D., Sarmento, B., Gomes, A.M., Pintado, M., Reis, F., 2015. Solid lipid nanoparticles as oral delivery systems of phenolic compounds: overcoming pharmacokinetic limitations for nutraceutical applications. Crit. Rev. Food Sci. Nutr. (accepted for publication), 00e00. Olbrich, C., Kayser, O., Müller, R.H., 2002. Lipase degradation of Dynasan 114 and 116 solid lipid nanoparticles (SLN)deffect of surfactants, storage time and crystallinity. Int. J. Pharm. 237 (1), 119e128.
, J.V., 2011. Situaç~ Paulino, J., Folegatti, M.V., Zolin, C.A., Rom
an, R.M.S., Jose ao da
rio 2006. Irriga agricultura irrigada no Brasil de acordo com o censo agropecua 16 (2), 163. Pikal, M.J., 1999. Mechanisms of protein stabilization during freeze-drying and storage: the relative importance of thermodynamic stabilization and glassy state relaxation dynamics. Freeze-Drying/Lyophilization Pharm. Biol. Prod. 96, 161e198. Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., Rice-Evans, C., 1999. Antioxidant activity applying an improved ABTS radical cation decolorization assay. Free Radic. Biol. Med. 26 (9), 1231e1237. Rosa, N.A.B., 2011. Development of Nanoparticles Loaded with Bioactive Compounds for Application as Nutraceuticals. Oporto University. Soares, S., Fonte, P., Costa, A., Andrade, J., Seabra, V., Ferreira, D., Reis, S., Sarmento, B., 2013. Effect of freeze-drying, cryoprotectants and storage conditions on the stability of secondary structure of insulin-loaded solid lipid nanoparticles. Int. J. Pharm. 456 (2), 370e381. Westesen, K., Bunjes, H., 1995. Do nanoparticles prepared from lipids solid at room temperature always possess a solid lipid matrix? Int. J. Pharm. 115 (1), 129e131. Wissing, S., Kayser, O., Müller, R., 2004. Solid lipid nanoparticles for parenteral drug delivery. Adv. Drug Deliv. Rev. 56 (9), 1257e1272. Zhang, L., Liu, L., Qian, Y., Chen, Y., 2008. The effects of cryoprotectants on the freeze-drying of ibuprofen-loaded solid lipid microparticles (SLM). Eur. J. Pharm. Biopharm. 69 (2), 750e759.