Assessment of cell viability and permeation enhancement in presence of lipid-based self-emulsifying drug delivery systems using Caco-2 cell model: Polysorbate 80 as the surfactant

Accepted Manuscript Assessment of cell viability and permeation enhancement in presence of lipid-based self-emulsifying drug delivery systems using Caco-2 cell model: Polysorbate 80 as the surfactant

Pengli Bu, Yue Ji, Silpa Narayanan, Damon Dalrymple, Xingguo Cheng, Abu T.M. Serajuddin PII: DOI: Reference:

S0928-0987(16)30558-9 doi: 10.1016/j.ejps.2016.12.018 PHASCI 3835

To appear in:

European Journal of Pharmaceutical Sciences

Received date: Revised date: Accepted date:

2 August 2016 27 November 2016 19 December 2016

Please cite this article as: Pengli Bu, Yue Ji, Silpa Narayanan, Damon Dalrymple, Xingguo Cheng, Abu T.M. Serajuddin , Assessment of cell viability and permeation enhancement in presence of lipid-based self-emulsifying drug delivery systems using Caco-2 cell model: Polysorbate 80 as the surfactant. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Phasci(2016), doi: 10.1016/ j.ejps.2016.12.018

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

ACCEPTED MANUSCRIPT Assessment of cell viability and permeation enhancement in presence of lipid-based self-emulsifying drug delivery systems using Caco-2 cell model: Polysorbate 80 as the surfactant

IP

T

Pengli Bu1,2, Yue Ji2, Silpa Narayanan1, Damon Dalrymple3, Xingguo Cheng1, and Abu

CR

T.M. Serajuddin1*

Department of Pharmaceutical Sciences, College of Pharmacy & Health Sciences,

2

Department of Biological Sciences, College of Liberal Arts and Sciences, St. John’s

US

1

ABITEC Corporation, 501W 1st Avenue, Columbus, Ohio 43215

ED

M

3

AN

University, 8000 Utopia Parkway, Queens, NY 11439

CE

PT

* Corresponding author: Dr. Abu T. M. Serajuddin, email: [email protected]

AC

Short Title: Assessment of cell viability and permeation enhancement by lipid-based drug delivery systems

Key words: lipid-based drug delivery system, medium-chain lipid, monoglyceride, triglyceride, propylene glycol monoester, surfactant, polysorbate 80, Caco-2 cells, cytotoxicity, permeation enhancement

1

ACCEPTED MANUSCRIPT ABBREVIATIONS LBDDS (lipid-based drug delivery system); LBSEDDS (lipid-based self-emulsifying drug delivery system); SEDDS (self-emulsifying drug delivery system); SMEDDS (selfmicroemulsifying drug delivery system); MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-

T

diphenyltetrazolium bromide); PG (propylene glycol); HBSS (Hank’s Balanced Salt

IP

Solution); DMEM (Dulbecco’s Modification of Eagle’s Medium); FBS (fetal bovine

AC

CE

PT

ED

M

AN

US

CR

serum); DMSO (dimethyl sulfoxide), TEER (transepithelial electrical resistance)

2

ACCEPTED MANUSCRIPT ABSTRACT Purpose – Lipid-based self-emulsifying drug delivery systems (SEDDS) are commonly used for solubilizing and enhancing oral bioavailability of poorly water-soluble drugs. However, their effects on viability of intestine epithelial cells and influence on membrane

T

permeation are poorly understood. The present study was undertaken for safety

IP

assessment of lipid-based formulations containing medium-chain fatty acid esters as

CR

lipids and polysorbate 80 as the surfactant using the Caco-2 in vitro model. Any possible paracellular permeation enhancement through Caco-2 monolayers by the nontoxic

US

formulations was also investigated.

AN

Methods –Mixtures of monoglyceride (Capmul MCM EP or 708G) or propylene glycol monoester (Capmul PG-8 NF) of medium chain fatty acids with polysorbate 80, with and

M

without the incorporation of a medium-chain triglyceride (Captex 355), were prepared.

ED

After suitable dilution with aqueous culture medium, the formulations were incubated with a series of Caco-2 cultures of different maturity. Cell viability and membrane

PT

integrity were assessed. Any effects of nontoxic formulations on the transport of the

CE

fluorescent dye, Lucifer yellow, through Caco-2 monolayers were also determined. Results –Formulations containing 1:1 ratios of monoglyceride or propylene glycol

AC

monoester to triglyceride (30% polysorbate 80, 35% monoglyceride or monoester and 35% triglyceride) were best tolerated by Caco-2 cells. Increased maturity obtained through longer culture durations rendered Caco-2 cells greater tolerance towards lipidbased formulations, and maximum tolerance to lipid-based formulations was observed with Caco-2 monolayers after being cultured for 21-23 days. Furthermore, extent of cell membrane rupture caused by lipid-surfactant mixtures correlated positively with levels of

3

ACCEPTED MANUSCRIPT cytotoxicity, suggesting a potential underlying mechanism. Permeation studies using Caco-2 monolayer model revealed that certain formulations significantly enhanced paracellular transport activities. Conclusions – Lipid-based SEDDS containing mixtures of monoglyceride (or

T

monoester) and triglyceride of medium chain fatty acids formed fine microemulsions and

IP

were significantly less toxic than other formulations. Fully differentiated Caco-2

CR

monolayer was more resistant to lipid-surfactant mixtures than less mature cultures.

AC

CE

PT

ED

M

AN

US

Certain formulations were also capable of enhancing paracellular permeation.

4

ACCEPTED MANUSCRIPT INTRODUCTION Oral route is the most preferred way of administering therapeutic agents to the body. However, the majority of existing drugs or newly developed drug candidates do not dissolve adequately in the gastrointestinal fluid to enable rapid and complete oral

T

absorption, which necessitates the use of bioavailability enhancing formulations,

IP

including lipid-based drug delivery systems [1]. In a lipid-based drug delivery system

CR

(LBDDS), different types of lipids, surfactants and/or co-surfactants are used to solubilize hydrophobic drug [2-7]. Various types of lipids, such as (1) digestible

US

vegetable oil containing predominantly unsaturated long-chain fatty acids, (2) glycerides

AN

and partial glycerides of medium-chain fatty acids, (3) polyacyl esters of medium-chain fatty acids (e.g., propylene glycol, PG, esters), (4) polyoxylglycerides consisting of

M

polyethylene glycol (PEG) esters of fatty acids (either long-chain unsaturated or medium-

ED

chain), etc., are commonly used in LBDDS. The surfactants that are included in the lipidbased formulations are essentially fatty acid esters. For instance, one commonly used

PT

surfactant Cremophor EL is polyoxyethylene glycol ester of ricinoleic acid, and another

CE

surfactant polysorbate 80, which is also known as Tween 80, is polyoxyethylene sorbitan monooleate [1]. In these surfactants, the polyoxyethylene chains increase hydrophilicity

AC

of the esters and improve hydrophilic-lipophilic balance for oil-in-water emulsification. For a drug to be maximally absorbed in the GI tract, it is essential that the formulation containing it is highly dispersible forming small globules upon dilution with gastrointestinal fluids. For this reason, formulations that lead to the formation of oil-inwater microemulsions upon dilution with aqueous media is most preferred for solubilization and bioavailability enhancement of hydrophobic drugs. It is generally

5

ACCEPTED MANUSCRIPT accepted that the average particle or globule size of microemulsion is <250 nm, while lipid-based formulations with particle size between 250 and 1000 nm are considered fine emulsions, and those having globule size >1000 nm are coarse emulsions [8, 9]. A key difference among LBDDS with different particle sizes is that microemulsion is

IP

thermodynamically unstable and may phase separate with time [1].

T

thermodynamically stable and optically isotropic, whereas the others are

CR

As mentioned above, LBDDS has extensively been used for its effects on drug solubilization. In addition to solubilizing drugs and presenting them to intestinal fluids as

US

extremely fine globules (microemulsions), LBDDS may also influence permeation of

AN

drugs through intestinal membranes [10, 11]. Thus, LBDDS may have duel effects of drug solubilization and permeation enhancement. Such effects of LBDDS may make it

M

exceptionally useful for the Biopharmaceutical Classification System (BCS) Class IV

ED

drugs, which have poor bioavailability due to their low solubility and low permeability. The rate-limiting barrier for absorption of orally administered drugs is the

PT

epithelial layer lining the GI tract. A cell line-based system, the Caco-2 cell model, has

CE

been widely used for predicting intestinal permeability of drugs and to screen oral dosage forms and drug delivery systems for permeation enhancement [12-14]. Caco-2 cells are

AC

immortal cells that were originally isolated from a human colon tumor [15]. Caco-2 cells can differentiate under experimental conditions into monolayer structures that are morphologically and functionally similar to the intestine epithelial layers. In particular, Caco-2 cells were shown to establish tight junctions between neighbor cells, which are characteristic of the in vivo epithelial layer and constitute the rate-limiting barrier for intestinal absorption. Because of the above reasons, the Caco-2 monolayer model has

6

ACCEPTED MANUSCRIPT been used extensively in permeation studies to investigate paracellular transport of drugs and accepted by the pharmaceutical industry and regulatory agencies as a standard in vitro model for testing of oral formulations [12, 14, 16]. The tight junction in Caco-2 monolayer has been shown to be modulated by different components of LBDDS. Certain

T

lipids and surfactants or their mixtures have been reported to disrupt tight junction in

IP

Caco-2 monolayer transiently and reversibly and thus enhanced drug permeation [11, 17-

CR

19]. However, in some cases, the decreased viability of Caco-2 cells caused by lipids or excipients was shown to be associated with increased drug permeation [20, 21]. Such an

US

observation indicates that cytotoxicity of lipids, surfactants, or mixtures of both, which

AN

are common ingredients of the lipid-based drug delivery systems, can damage the integrity of Caco-2 monolayer. Consequently, permanent disruptions of tight junctions in

M

Caco-2 monolayers would lead to an increase in drug permeation, which can be

ED

designated as undesired toxicity-mediated permeation enhancement. Despite the importance of establishing tolerance (absence of cytotoxicity) of

PT

excipients on Caco-2 cell monolayer prior to conducting any drug permeation studies,

CE

there are not many reports in the literature assessing safety of lipid-based drug delivery systems. We recently reported the relative tolerance and possible cytotoxic effects of

AC

several lipid-based systems containing Cremophor EL as the surfactant using the Caco-2 model [22]. As a continuation of our studies, it was of interest to determine the tolerance on Caco-2 cells of microemulsions and emulsions containing polysorbate 80 as the surfactant. Polysorbate 80 is the most commonly used surfactant in pharmaceutical dosage forms and the Food and Drug Administration (FDA) assigned it the GRAS (generally recognized as safe) status [23]. We have also established in our laboratory that

7

ACCEPTED MANUSCRIPT polysorbate 80 produces microemulsions when used with lipids containing both mediumchain and long-chain fatty acids (A. Serajuddin, manuscript under preparation). For the present study, we prepared lipid-based self-emulsifying formulations containing polysorbate 80 and selected lipids consisting of mono- and triglycerides of medium-chain

T

fatty acids. Cytotoxicity of the formulations was assessed on differentially matured Caco-

IP

2 cells, and links between microemulsion formation, culture maturity and sensitivity

CR

towards lipid-based formulations were established. In addition, formulations showing good tolerance were selected and their effects on enhancing the permeation of lucifer

Lipids and surfactant

M

ED

MATERIALS AND METHODS

AN

US

yellow, a marker for paracellular transport, were determined.

PT

The surfactant and medium-chain lipids (mono- and triglycerides; PG monoesters) used

CE

in the present study, along with their trade names, manufacturers, chemical structures and compositions, are listed in Table 1. An aqueous buffer, Hank's Balanced Salt Solution

AC

(HBSS) (Hyclone, cat. # SH30588.02), was used to prepare the preconcentrate of the surfactant alone formulation (30% surfactant and 70% buffer). It was also used as the diluent for all formulations in particle size measurement.

Preparation of lipid-based formulations

8

ACCEPTED MANUSCRIPT Test lipids were mixed with the surfactant polysorbate 80 at the following ratios (v/v): (1) 30% polysorbate 80 and 70% monoglyceride; (2) 30% polysorbate 80 and 70% PG monoester; (3) 30% polysorbate 80, 52.5% monoglyceride and 17.5% triglyceride; (4) 30% polysorbate 80, 52.5% PG monoester and 17.5% triglyceride; (5) 30% polysorbate

T

80, 35% monoglyceride and 35% triglyceride; and (6) 30% polysorbate 80, 35% PG

IP

monoester and 35% triglyceride. To maximize the amount of lipid in the formulation,

CR

only 30% surfactant was used. When mixed lipids were used in formulations, the maximum triglyceride to monoglyceride (or PG monoester) ratio used was 1:1, since the

US

previous studies showed that a higher ratio of triglyceride led to phase separation of lipid

AN

upon dilution with aqueous buffer [24]. As listed in Table 1, two monoglycerides (Capmul MCM EP and Capmul 708G), each with different chemical composition, were

M

used. Capmul PG-8 NF was used as the PG monoester and Captex 355 as the triglyceride.

ED

The preconcentrates of all formulations were prepared as previously described [22].

PT

Formulation concentration selection and particle size analysis

CE

The lipid based formulations were diluted with aqueous HBSS solution prior to incubation of cell cultures and monolayers for evaluation of toxicity and membrane

AC

permeation. The dilutions were made based on the consideration that the volume of the preconcentrate, i.e., the formulation prior to the addition of water, that may be filled in an oral capsule could range from 0.25 mL to 1.25mL and, upon oral administration, it would be diluted in about 250 mL aqueous medium present in the human gastrointestinal (GI) tract [6, 8]. Since this will lead to concentrations of 0.1% to 0.5%, we selected concentrations 0.1%, 0.25% and 0.5% for test formulations.

9

ACCEPTED MANUSCRIPT The particle size of lipid-surfactant mixtures at 0.1%, 0.2%, and 0.5% (v/v) in aqueous HBSS solution was measured using a DelsaNano C particle size analyzer (Beckman Coulter Inc., Brea, CA) at room temperature (24-25°C). Three replicate solutions for each concentration were prepared following procedures described

IP

T

previously [22].

CR

Cell culture

Initiation and subculture of Caco-2 cells (ATCC, cat. # HTB-37) were described in

US

details in a previous study [22]. Briefly, for 1-day and 5-day cultures, Caco-2 cells were

AN

seeded into regular 96-well tissue culture plates at 2x104 cells per well in 200 μL of growth medium, and, for 21-23 day monolayer cultures, the cells were seeded at the same

Cell viability assessment

ED

M

density into 96-well transwell plates.

PT

After being cultured for indicated periods of time, Caco-2 cells were treated with

CE

individual test formulations in a final volume of 200 μL per well in 96-well plates at specified concentrations diluted with culture medium for two hours. Immediately after

AC

treatment, cytotoxicity was evaluated using MTT cell viability assay as previously described [22].

Membrane integrity assessment Caco-2 cells were seeded into 96-well tissue culture grade black plates with clear bottom (Corning, COSTAR, cat. # 3603) at 2x104 cells per well in a volume of 200 μL growth

10

ACCEPTED MANUSCRIPT medium on day 0 and cultured for 5 days with medium changed on days 2, 4 and 5. Treatment with lipid-surfactant mixtures was conducted on day 5. Immediately after treatment, cells were processed for MultiTox-Fluor Cytotoxicity Assay (Promega, cat. #

T

G9200) according to manufacturer’s instructions as described previously [22].

CR

IP

Permeation study

Caco-2 monolayer culture

US

Caco-2 cells were seeded into 24-well transwell plates (Falcon HTS Multiwell Insert

AN

System, cat. #351180, membrane pore size 1.0 μm) at 2.3x104 in a volume of 0.5 mL growth medium containing antibiotic (penicillin and streptomycin, Gibco, cat. # 15140-

M

122) per well and cultured for 21-23 days with medium changed every other day. From

ED

day 18 onwards, TEER (transepithelial electrical resistance) of Caco-2 monolayers was measured with an Epithelial Voltohmmeter (EVOM2, World Precision Instruments,

CE

PT

Sarasota, FL) every other day till the day of permeation study [12].

Standard curve

AC

Lucifer yellow is generally considered as neither transported via transcellular route nor permeable to the tight junction, and thus was routinely used as a marker for paracellular transport [20, 25]. Included with every permeation study, a standard curve of Lucifer yellow (Lucifer Yellow CH, Dilithium Salt, MW 457.2; Cat. # 155267, MP Biomedicals LLC, Solon, OH) was established using eight Lucifer yellow concentrations (25 μg/mL, 12.5 μg/mL, 6.25 μg/mL, 5 μg/mL, 2.5 μg/mL, 1 μg/mL, 0.5 μg/mL, and 0.25 μg/mL)

11

ACCEPTED MANUSCRIPT and the fluorescence intensity was quantified with a fluorescence microplate reader (Synergy H1, Biotek, excitation: 450 nm, emission 520 nm, gain 50).

Selection of formulations for permeation study

T

The test formulations that were well tolerated by Caco-2 monolayers in toxicity

IP

assessment in the present study were selected for studying their effects on the permeation

CR

of Lucifer yellow. These formulations contained polysorbate 80 as the surfactant. In a previous study [22], we also conducted safety assessment of lipid-based formulations

US

containing Cremophor EL (also known as Koliphor EL) as the surfactant; however, any

AN

effects of those formulations on the permeability of Caco-2 monolayers were not studied. Therefore, we also included in the present investigation formulations containing

M

Cremophor EL that were well tolerated by Caco-2 monolayers for a comprehensive

ED

evaluation of lipid-based formulations containing both polysorbate 80 and Cremophor EL for their effects on membrane permeability. All formulations were diluted to 0.2%

PT

concentration using the aqueous HBSS buffer, except for the mixture of polysorbate 80

CE

30% and Capmul MCM 70%, where both 0.2% and 0.5% concentrations were used. The polysorbate 80-Capmul MCM mixture was nontoxic at 0.2% and toxic when the

AC

concentration was increased to 0.5%.

Procedures of permeation study Fresh medium was added to Caco-2 monolayers 12-16 h before the permeation study. Prior to the beginning of the study, transwells containing Caco-2 monolayers were gently rinsed with pre-warmed transport buffer (HBSS supplemented with 25mM HEPES,

12

ACCEPTED MANUSCRIPT pH7.4) and incubated with transport buffer for 30 min at 37°C. Individual test formulations were diluted with pre-warmed transport buffer containing 50 μg/mL Lucifer yellow, thoroughly mixed, and dispensed into the apical compartment of 4-6 replicate transwells (400 μL per one apical compartment). The transwell inserts were placed into a

T

24-well regular tissue culture plate, where each well served as the basal compartment and

IP

contained 1.2 mL pre-warmed transport buffer. An aliquot (100 μL) from each basal

CR

compartment was collected at 30, 60 and 120 min after the initiation of the incubation; the basal compartment was then replenished with 100 μL of fresh transport buffer. The

US

aliquot from each basal compartment was mixed with 100 μL transport buffer in a black

AN

96-well plate with clear-bottom (Corning, cat. #3904) and fluorescence intensity was

ED

nm, emission 520 nm, gain 50).

M

quantified using a fluorescence microplate reader (Synergy H1, Biotek, excitation: 450

Calculation of apparent permeation coefficient (Papp)

PT

The apparent permeability coefficient (Papp, unit, cm/s) of Lucifer yellow (LY) was

CE

calculated according to the following equation [17, 26, 27]: Papp = (ΔQ/Δt)/(C0 x A)

AC

where ΔQ/Δt is the linear appearance rate of LY mass in the basal compartment (μmol/s); C0 is the initial concentration of LY in the apical compartment (μM); A is the surface area of each transwell (2.1 cm2 for BD Falcon HTS 24-multiwell Insert System).

Statistical analysis

13

ACCEPTED MANUSCRIPT Representative data of two or three independent experiments are presented as mean ± S.E.M., with 4-6 replicates for each treatment group in each experiment. Data were analyzed by one-way ANOVA, followed by Duncan’s post hoc test. Statistical

CR

IP

T

significance was set at p < 0.05.

US

RESULTS

AN

Preparation of lipid-based self-emulsifying drug delivery systems and particle size analysis

M

Components of lipid-based self-emulsifying drug delivery systems were listed in Table 1.

ED

The formulations were prepared as described in MATERIALS AND METHODS and in a previous study [22]. Particle size analysis was performed on all the formulations that

PT

were later tested for cytotoxicity with Caco-2 cells. The results, as presented in Table 2,

CE

indicate that the addition of the aqueous buffer Hank's Balanced Salt Solution (HBSS) to liquid-surfactant preconcentrates produced microemulsions (particle sizes less than 250

AC

nm) in most cases and fine emulsions (particle sizes between 250 and 1000 nm) in some cases. In general, larger globules were formed when only a monoglyceride or the monoester of propylene glycol was used in the formulation. It is noteworthy that ternary mixtures of polysorbate 80 with 1:1 ratio of monoglyceride and triglyceride formed microemulsions under all three concentrations (0.1%, 0.2%, and 0.5%). The impact of time on the particle size of lipid-surfactant mixtures was also assessed during the

14

ACCEPTED MANUSCRIPT dispersion test. The particle size slightly grew in two cases of fine emulsions (30% polysorbate 80 + 70% of Capmul MCM EP and 30% polysorbate 80 + 70% of Capmul PG-8 NF) in the 2-h test. When microemulsions were formed, the particle size remained practically unchanged over the test period, confirming that microemulsions are

IP

T

thermodynamically equilibrium systems [28].

CR

Cell viability assessment of lipid-surfactant mixtures on Caco-2 cells from various culture stages

US

Lipid-based formulations containing the surfactant polysorbate 80, as described in Table

AN

2, were tested for their influences on viability of Caco-2 cells. Caco-2 cells cultured for various periods of time (1, 5, and 21-23 days) were used in this experiment. Two

M

treatment groups were included as controls, the surfactant alone group and the buffer

ED

(70%)-surfactant (30%) mixture (the buffer occupies the same percentage as the test lipid does in the lipid-surfactant mixtures). The control group received only the culture

PT

medium and no lipid-surfactant mixture, and it was considered as 100% viable.

CE

On Caco-2 cells cultured for 1 day, most of the formulations caused a significant reduction on cell survival (Fig. 1). The cells appeared to tolerate the surfactant alone and

AC

the surfactant-buffer mixture (7:3, v/v) better than the lipid-surfactant mixtures, although the toxicity was still significant (Fig. 1A). Surprisingly, one particular formulation with a composition of 30% surfactant, 52.5% monoglyceride Capmul 708G and 17.5% Captex (where the ratio of monoglyceride to triglyceride was 2:1) showed excellent tolerance by Caco-2 cells and minimal toxicity at the concentration of 0.1% (Fig. 1C, middle panel). At this concentration, this particular formulation exists as a microemulsion with a particle

15

ACCEPTED MANUSCRIPT size around 50 nm (Table 2). With the same test lipid (Capmul 708G), when the ratio of monoglyceride to triglyceride became 1:1, a reduced toxicity (60% cell survival) was still seen at the concentration of 0.1% compared to the toxicity of the formulation containing surfactant and only Capmul 708G (30% cell survival) (Fig. 1C, first blue bar and first red

T

bar vs. first yellow bar). For test lipid PG monoester Capmul PG-8, the formulation

IP

containing 30% surfactant, 35% Capmul MCM and 35% Captex (where the ratio of

CR

propylene glycol monoester to triglyceride was 1:1) appeared to be least toxic at the concentration of 0.1% when compared to formulations containing only PG-8 or PG-8 and

US

Captex at a ratio of 2:1 (Fig. 1D, first red bar vs. first yellow bar and first blue bar).

AN

After growing for 5 days, the overall tolerance of the Caco-2 cells to lipidsurfactant mixtures improved markedly (Fig. 2). Although some cytotoxicity was still

M

obvious when the cells were exposed to the surfactant alone at the concentration of 0.5%,

ED

the buffer-surfactant mixture group showed no toxicity at all under all the concentrations tested (0.1%, 0.2% and 0.5%) (Fig. 2A). For formulations containing test lipids at 0.1%,

PT

cell viability was restored to control level (no lipid-surfactant mixtures) (Fig. 2B-D). At

CE

the highest concentration tested (0.5%), the test lipid monoglycerides Capmul MCM and Capmul 708G showed least toxicity when formulated with triglyceride Captex at 1:1 ratio

AC

(Fig. 2, B and C, third red bar). For the test lipid Capmul PG-8, the least toxicity was observed when it was formulated with the triglyceride Captex at 2:1 ratio (Fig. 2D, blue bars). When Caco-2 cells were growing on filter membrane for 21-23 days following the same procedure used in permeability studies [16, 17, 20] to form a monolayer structure, the cells tolerated lipid-surfactant mixtures even better than the 5-day culture. No toxicity

16

ACCEPTED MANUSCRIPT was seen in the control treatment group containing either surfactant alone or a mixture of 70% buffer and 30% surfactant (Fig. 3A). Consistent with what was observed in the 5day culture, the inclusion of a triglyceride into the formulation conferred an even greater tolerance, especially for the highest concentration (0.5% dilution) treatment (Fig. 3B-D).

T

For all three test lipids, the least toxicity was seen when the ratio of monoglyceride to

IP

triglyceride was 1:1 (Fig. 3B-D, red bars). For Capmul 708G and Capmul PG-8,

CR

significant toxicity was associated with formulations containing only the surfactant and the test lipid at both 0.2% and 0.5% (Fig. 3, C and D, second and third yellow bars).

US

Addition of the triglyceride at the ratio of either 2:1 or 1:1 (monoglyceride or propylene

AN

glycol monoester to triglyceride) markedly reduced toxicity, and the least toxicity was achieved with formulations containing a 1:1 ratio of monoglyceride to triglyceride (Fig.

ED

M

3B-D, red bars).

Assessment of membrane integrity in Caco-2 cells treated with lipid-surfactant

PT

mixtures

CE

We next investigated the underlying mechanism for toxicity induced by lipid-surfactant mixtures. Our working hypothesis was that lipids and lipid-surfactant mixtures damage

AC

cell membrane by interacting with the phospholipid component of the plasma membrane and thereby disrupt membrane integrity [29, 30]. The loss of membrane integrity subsequently leads to cellular injury and eventually cell death, which will be reflected as reduced metabolic activities at the cell population level. We chose the 5-day culture of Caco-2 cells for membrane integrity assessment for the following reasons: (1) the 5-day culture of Caco-2 cells showed an improved tolerance to lipid/surfactant mixtures, which

17

ACCEPTED MANUSCRIPT was much better than that of the 1-day culture and similar to that of the 21-day culture; and (2) the method that we used to assess membrane integrity is designed for cells cultured in regular tissue culture plates and not compatible with transwell plates that were used to grow the 21-day monolayers.

T

Cell membrane integrity was evaluated using the MultiTox-Fluor Cytotoxicity

IP

assay (Promega, Madison, WI), which measures the activity of a marker protease specific

CR

to cells with compromised plasma membrane. A fluorogenic and cell-impermeant peptide substrate is converted to fluorescent products only by the marker protease that leaks out

US

through damaged cell membrane. Consistent with cytotoxicity results obtained from

AN

Caco-2 cells with the same maturity (5-day culture), membrane damage was seen in a dose-dependent manner, i.e., the higher the concentrations of lipid-surfactant mixtures,

M

the more intense the fluorescent signals (indicating more severe cell membrane

ED

disruption) (Fig. 4). For example, the highest cytotoxicity was seen in treatments with 0.5% lipid-surfactant mixtures (Fig. 2B-D) and the same treatment regimen gave the

PT

highest fluorescent reading (Fig. 4B-D), indicating a positive correlation between cell

CE

membrane impairment assessed by the MultiTox-Fluor Cytotoxicity assay and

AC

cytotoxicity measured by the MTT cell viability assay.

Assessment of permeation enhancement of lipid-surfactant mixtures on Caco-2 monolayers One major objective of the present investigation was to evaluate the permeation enhancing effects of lipid-surfactant mixtures under nontoxic conditions. Based on results presented in Fig. 2 and in a previous study [22], the formulations that were well tolerated

18

ACCEPTED MANUSCRIPT during the cytotoxicity study were tested further for their effects on permeation of a marker molecule Lucifer yellow (LY) using Caco-2 monolayers. In general, it was observed that LBDDS containing polysorbate 80 as the surfactant had less effect on permeation enhancement than formulations containing Cremophor EL. Of eight nontoxic

T

formulations containing polysorbate 80 tested, only two of them, one containing 30%

IP

polysorbate 80 + 70% Capmul MCM and the other containing 30% polysorbate 80 +

CR

35% Capmul PG-8 + 35% Captex, showed significant effects in enhancing paracellular transport of LY. The other six formulations containing (i) 100% polysorbate 80, (ii) 30%

US

polysorbate 80 + 70% aqueous buffer, (iii) 30% polysorbate 80 +52.5% Capmul MCM +

AN

17.5% Captex, (iv) 30% polysorbate 80 + 35% Capmul MCM + 35% Captex, (v) 30% polysorbate 80 + 52.5% Capmul 708G + 17.5% Captex and (vi) 30% polysorbate 80 +

M

35% Capmul 708G + 35% Captex did not increase permeation of LY significantly.

ED

In contrast, all formulations containing Cremophor EL exhibited enhancement in LY permeation. Additionally, three toxic polysorbate 80-lipid formulations were tested to

PT

determine whether the cytotoxic effect could have any influence on membrane

permeation.

CE

permeation (Table 3 and Fig. 5). Indeed, all three toxic formulations exhibited membrane

AC

The apparent permeability coefficient (Papp) values of the formulations enhancing permeation are presented in Table 3 and Fig. 5. Caco-2 monolayers that were used in permeation studies were formed on filter membrane after 21-23 days of culture with consistent TEER reading above 1200 (Ωcm2) from days 18 onwards when measured under ambient temperature (24-25°C). Negative and positive controls for the permeation experiment included wells that had Caco-2 monolayer growing on top of the porous filter

19

ACCEPTED MANUSCRIPT membrane and wells that had only the porous filter membrane and no cells, respectively. The fluorescent dye Lucifer yellow was used as the marker for paracellular transport, and it was added to the apical chamber of all wells (including the negative control and positive control wells). The amounts of Lucifer yellow in the samples collected from the

T

basal compartments after 30, 60 and 120 min of incubation indicated levels of

CR

IP

paracellular transport.

Permeation enhancement of lipid-based formulations containing polysorbate 80

US

The Caco-2 cell membrane permeation enhancement of LY by formulations containing

AN

polysorbate 80 as a function of time are shown graphically in Fig. 5B and the apparent permeability coefficients (Papp) are tabulated in Table 3. All formulations were used at

M

0.2% concentration, except for the mixture of polysorbate 80 30% and Capmul MCM

ED

70%, which was tested at both 0.2 and 0.5% concentrations, the later being the toxic formulation. The amounts of LY in basal compartments from negative control (no lipid)

PT

and positive control (no cell) represented, respectively, the minimal and the maximal

CE

paracellular transport (Fig. 5A), and any permeation enhancement was evaluated and the statistical assessment conducted with respect to the negative control. As mentioned

AC

earlier, only two of the nontoxic formulations (Tw 30%+MCM 70% and Tw 30%+PG-8 35%+Captex 35%) demonstrated membrane permeation, that these formulations markedly enhanced paracellular transport of LY after 30 min of incubation as compared to the control (no lipid) (Fig. 5B, first panel, white bars). The three formulations that were toxic according to Fig. 3 (Tw 30%+PG 70%, Tw 30%+PG 52.5%+Cap17.5%, and Tw 30%+Capmul MCM 70%) enhanced paracellular transport to a greater extent

20

ACCEPTED MANUSCRIPT compared to the nontoxic formulations (Fig. 5B, first panel, orange bars; Papp in Table 3) after the initial 30 min of incubation, indicating, as expected, that the cytotoxic effect may also be responsible for the membrane permeation enhancement. After 60 and 120 min incubation, the nontoxic formulations achieved permeation enhancement similar to

T

that of toxic formulations (Fig. 5B, second and third panels). Throughout the 2h test

IP

period, both toxic and nontoxic formulations showed time-dependent permeation

CR

enhancement.

US

Permeation enhancement of lipid-based formulations containing Cremophor EL

AN

From the three selected formulations containing Cremophor EL, two (CEL 30%+MCM35%+Cap 35% and CEL 30%+708G 35%+Cap 35%) showed significant

M

permeation enhancements after 30 min of incubation in Caco-2 monolayers (Fig. 5C, first

ED

panel, white bars). The additional two formulations included were the “control formulations” as they either contained 100% of the surfactant (Cremophor EL) or 30%

PT

surfactant and 70% aqueous buffer (HBSS) (v/v) without any lipid. It was somewhat

CE

surprising that one control formulation, CEL 30%+buffer 70%, showed the most enhancement of paracellular transport at all three time points (Fig. 5C, the first white bar

AC

in all three panels). Two test formulations (CEL 30%+MCM35%+Cap 35% and CEL 30%+708G 35%+Cap 35%) showed continuous permeation enhancing effects from 30 min through 120 min (Fig. 5C, the second and third white bars in all three panels). The third test formulation (CEL 30%+PG-8 35%+Cap 35%) started to show significant permeation enhancement after 60 min of incubation and the effect was also sustained after 120 min (Fig. 5C, the fourth white bar in second and third panels). The apparent

21

ACCEPTED MANUSCRIPT permeability coefficients (Papp) in presence of all formulations as a function of time are given in Table 3, and the significance of permeation enhancement by formulations containing Cremophor EL relative to the control as well as relative to each other, where applicable, is also indicated in Fig. 5C.

T

Overall, the selected lipid-surfactant mixtures exhibited a time-dependent pattern

IP

of permeation enhancement, suggesting that lipid-based formulations and/or surfactants

CR

present in aqueous media are likely to interact with tight junction components to enable paracellular transport through Caco-2 monolayers. Further studies are necessary to

US

elucidate the mechanism of how surfactants and lipid-based formulations influence

AN

membrane permeation and why in some cases polysorbate 80 and Cremophor EL act

M

differently.

ED

DISCUSSION

PT

Extensive research during the past two decades demonstrated that lipid-based

CE

formulations, and especially the self-emulsifying drug delivery systems (SEDDS), are effective means of improving solubility and intestinal absorption of hydrophobic drugs

AC

[10]. The principle characteristic of SEDDS is its ability to form fine oil-in-water emulsions when diluted with aqueous medium such as the gastric fluid [1, 5, 31]. In the present investigation, we were able to obtain lipid-based drug delivery systems that selfemulsify into microemulsions or very fine emulsions (Table 2) without using any organic co-solvents. Further, we were able to reduce the surfactant component to 30% and increase the lipid component to 70% by still retaining the ability to form microemulsion.

22

ACCEPTED MANUSCRIPT It is known that lipids and excipients vary in their structural and physiochemical properties, and different compositions of lipid-based formulations may confer distinct biopharmaceutical attributes with respect to drug load, emulsification, drug distribution and absorption [32]. We hypothesized that different compositions of surfactant and lipids

T

may also be associated with distinct cytotoxicity. By varying the ratio of monoglyceride

IP

(or PG monoester) to triglyceride (Captex 355), we demonstrated in the present study that

CR

the formulation containing 30% surfactant (Tween 80), 35% monoglyceride (or PG monoester) and 35% triglyceride exhibited the least toxicity under all conditions

US

examined (Figs 1-3, B-D, red bars vs. yellow and blue bars).

AN

Although Caco-2 cells are used extensively as models for studying drug intestinal permeation and evaluation of different dosage forms, considerable discrepancies still

M

exist in reported cell culture protocols. Variations in key cell culture parameters, such as

ED

seeding density and culture duration, have direct influences on morphological and functional properties of the Caco-2 cell population [14, 33, 34]. The initial seeding

PT

density of Caco-2 cells reported in the literature varies considerably, ranging from

CE

3.5x103 to 5x105 cells per cm2 of growth area [20, 26, 33, 35, 36]. In the current study, an intermediate seeding density (6x104 cells per cm2 growth area) was selected because

AC

such a seeding density was shown to achieve consistent differentiation in Caco-2 monolayers after three weeks of culture [14]. There is also no general consensus about what should be the maturity of cells used in cytotoxicity and permeation tests, and the use of Caco-2 cell culture with relatively short duration of maturity is common. Gupta et al. [37, 38] used 3-day old Caco-2 cells for studying drug permeability through Caco-2 cell layers. Cell cultures with 4, 6 and 7 days of maturity have also been used [17, 39-41]. In a

23

ACCEPTED MANUSCRIPT systematic study, we established a series of Caco-2 culture with the same initial seeding density but varying culture durations. As expected, those Caco-2 cultures with different confluence and maturity exhibited distinct sensitivities toward the same lipid-surfactant mixtures (Figs 1-3). Specifically, the 1-day Caco-2 culture appeared to be most fragile

T

showing the highest sensitivity to most of the lipid-surfactant mixtures. Caco-2 cells after

IP

5 days of culture showed marked tolerance towards most of the formulations. These cells

CR

were able to withstand low concentration (0.1%) and tolerate medium concentration (0.2%) with moderate cellular injury (Fig. 2, B-D). Therefore, Caco-2 5-day cultures may

US

be used when a highly sensitive method is required to evaluate relative toxicities of

AN

different lipid-based excipients and formulations. Once monolayers were established after 21-23 days, the Caco-2 culture exhibited much better resistance towards all formulations

M

used, showing minimal cytotoxicity at the medium concentration used (0.2%) and, in

ED

some cases, mild cell injury at the high concentration (0.5%) (Fig. 3, B-D). The results clearly demonstrate that maturity of Caco-2 cell culture is a determinant factor for the

PT

cytotoxicity outcome. By comparing Caco-2 culture with different duration side by side,

CE

it is apparent that the monolayer culture was most tolerant to lipid-base formulations and should be considered as a more accurate model for cytotoxicity and drug permeability

AC

assessment. The cell viability results obtained in the present investigation are in general agreement with the results obtained previously when Cremophor EL was used as the surfactant [22]. Further studies are necessary to determine why lipid-based formulations containing the 1:1 ratio of monoglyceride (or PG monoester) and triglyceride showed the lowest cytotoxicity among different formulations used. It could be the presence of a

24

ACCEPTED MANUSCRIPT higher proportion of triglyceride, and there is also the possibility that it could be related to the particle size of the microemulsions formed. In the present study as well as in a previous study [21], it was observed that formulations containing the 1:1 ratio of monoand triglycerides along with polysorbate 80 or Cremophor EL provided much lower

T

particle size as compared to other formulations. The lower particle size may be the result

IP

of better efficiency in forming microemulsion. The micellar solutions of neat surfactants

CR

(polysorbate 80 and Cremophor EL) that did not show toxic effects in 21-day Caco-2 monolayers in the present study as well as in the previous study also had very low

US

particle sizes.

AN

The Caco-2 monolayer presents an excellent drug transport model by mimicking the in vivo epithelial lining of GI tract [14, 15] as a well-formed Caco-2 monolayer

M

constitutes the “barrier” for a drug or a marker molecule to cross predominantly via the

ED

paracellular route [27]. Appropriate controls of permeation studies include a monolayer with marker and without any lipid or excipient, i.e. the negative control that reflects the

PT

baseline leakage of the monolayer. A positive control would be a condition that allows

CE

the maximum possible passage of the drug or the marker, like the absence of monolayer barrier that served this purpose in the current study. In the present study, formulations

AC

and concentrations that were shown to be nontoxic by the cytotoxicity assessment were selected for permeation study. Several formulations that were shown to be toxic were also included. After the initial 30 min of incubation, more marker molecules (LY) passed through the monolayer in the presence of lipid-surfactant mixtures, although the absolute amounts of LY permeated were relative modest (Fig. 5, B & C). After longer incubation periods (60 and 120 min), more LY molecules crossed the monolayer barrier in all the

25

ACCEPTED MANUSCRIPT samples including negative and positive controls, indicating that the paracellular transport activity is a function of incubation time. When compared to negative control, the presence of lipid-based formulations significantly enhanced permeation of LY, particularly in the case of Tw 30%+PG 35%+Cap 35% and CEL 30%+MCM 35%+Cap

T

35%, the two formulations that were most effective on enhancing permeation. It is of

IP

interest to note that a similar surfactant-monoglyceride/PG monoester-triglyceride ratio

CR

(30:35:35) also consistently proved to be the safest formulations during the cytotoxicity assessment and formed nanoemulsions in aqueous medium in dispersion test.

US

We observed that polysorbate 80 by itself did not enhance permeation of LY

AN

through the Caco-2 monolayer to any significant extent, while Cremophor EL did. The effects of Cremophor EL are further illustrated in Fig. 5 and the apparent permeability

M

coefficients (Papp) are recorded in Table 3, where it was surprisingly observed that the

ED

permeation of LY was higher in presence of a lower concentration of Cremophor EL, indicating major influence the surfactant concentration on membrane permeation.

PT

Essentially, two formulations of Cremophor EL, one containing the neat surfactant (CEL

CE

100%) and the other containing 30% surfactant in an aqueous medium (CEL 30%+HBSS buffer 70%) were diluted to 0.2% concentration by using the buffer, where the only

AC

difference was that the later test solution had 70% less surfactant (0.2% vs. 0.06%). Further systematic studies are needed to determine the effect of Cremophor EL concentration on the Caco-2 cell membrane permeability. As mentioned earlier, the permeation of LY through Caco-2 cell monolayer in presence of toxic formulations was higher after the initial 30 min of incubation as compared to that with nontoxic formulations. However, at 60 and 120 min, both nontoxic

26

ACCEPTED MANUSCRIPT and toxic formulations displayed similar permeation enhancement. It is possible that the tight junction present in the Caco-2 monolayer was modulated by the nontoxic formulations in a gradual manner, i.e. the opening of tight junction occurred transiently and reversibly at the beginning of the incubation. In contrast, the toxic formulations

T

might have caused the opening of tight junction permanently and irreversibly at the first

IP

encounter and, therefore, a stimulatory effect was evident after a short incubation period.

CR

After longer incubation time, the modulation of tight junction by nontoxic formulations possibly reached saturation and, as a result, the opening became sustained, whereas the

US

“permanent and irreversible” opening caused by toxic formulations remained unchanged.

AN

Therefore, the difference in permeation enhancement by toxic formulations disappeared after prolonged incubation for 30 and 60 min. Future studies are needed to clarify

M

mechanisms underlying the modulation of tight junction by both toxic and nontoxic lipid-

PT

CONCLUSIONS

ED

based formulations.

CE

Lipid-based drug delivery systems (LBDDS) are widely used dosage forms for oral delivery of poorly water-soluble drugs. Toxicity assessment of lipid-based formulations

AC

and the components within is a prerequisite for the selection of suitable lipids and excipients. In the present study, we presented a systematic approach for cytotoxicity evaluation of common components of LBDDS using the Caco-2 cell model. We determined that fully differentiated Caco-2 monolayers were most suitable for assessing cytotoxicity as compared to cells with lesser maturity. Furthermore, we observed that well-tolerated microemulsions also effectively enhanced paracellular transport in the

27

ACCEPTED MANUSCRIPT Caco-2 monolayer model. Thus, the present investigation provides a systematic approach for identifying optimal lipid-based formulations based on their effects on cellular

CR

ACKNOWLEDGEMENTS & DISCLOSURES

IP

T

tolerance and membrane permeation.

This study was supported, in part, by a generous research grant from ABITEC

AC

CE

PT

ED

M

AN

US

Corporation, 501 W. 1st Avenue, Columbus, OH 43215.

28

ACCEPTED MANUSCRIPT REFERENCE: 1.

Rahman, M.A., et al., Oral lipid based drug delivery system (LBDDS): formulation, characterization and application: a review. Curr Drug Deliv, 2011. 8(4): p. 330-45.

2.

Shah, A.V. and A.T. Serajuddin, Development of solid self-emulsifying drug delivery

T

system (SEDDS) I: use of poloxamer 188 as both solidifying and emulsifying agent for

Lipinski, C.A., et al., Experimental and computational approaches to estimate solubility

CR

3.

IP

lipids. Pharm Res, 2012. 29(10): p. 2817-32.

and permeability in drug discovery and development settings. Adv Drug Deliv Rev,

4.

US

2001. 46(1-3): p. 3-26.

Porter, C.J., et al., Enhancing intestinal drug solubilisation using lipid-based delivery

5.

AN

systems. Adv Drug Deliv Rev, 2008. 60(6): p. 673-91.

Gursoy, R.N. and S. Benita, Self-emulsifying drug delivery systems (SEDDS) for

M

improved oral delivery of lipophilic drugs. Biomed Pharmacother, 2004. 58(3): p. 173-

6.

ED

82.

Prajapati, H.N., D.M. Dalrymple, and A.T. Serajuddin, A comparative evaluation of

PT

mono-, di- and triglyceride of medium chain fatty acids by lipid/surfactant/water phase

CE

diagram, solubility determination and dispersion testing for application in pharmaceutical dosage form development. Pharm Res, 2012. 29(1): p. 285-305. Charman, W.N., Lipids, lipophilic drugs, and oral drug delivery-some emerging

AC

7.

concepts. J Pharm Sci, 2000. 89(8): p. 967-78. 8.

Prajapati, H., D.M. Dalrymple, and A.T.M. Serajuddin, In Vitro Dispersion Test Can Serve as a Predictive Method for Assessing Performance of Lipid-based Drug Delivery Systems. The Journal of Excipients and Food Chemicals, 2013. 4(4): p. 111-125.

9.

Sintov, A.C. and L. Shapiro, New microemulsion vehicle facilitates percutaneous penetration in vitro and cutaneous drug bioavailability in vivo. J Control Release, 2004. 95(2): p. 173-83.

29

ACCEPTED MANUSCRIPT 10.

Singh, B., et al., Self-emulsifying drug delivery systems (SEDDS): formulation development, characterization, and applications. Crit Rev Ther Drug Carrier Syst, 2009. 26(5): p. 427-521.

11.

Lindmark, T., Y. Kimura, and P. Artursson, Absorption enhancement through

Shah, P., et al., Role of Caco-2 cell monolayers in prediction of intestinal drug

CR

12.

IP

Caco-2 cells. J Pharmacol Exp Ther, 1998. 284(1): p. 362-9.

T

intracellular regulation of tight junction permeability by medium chain fatty acids in

absorption. Biotechnol Prog, 2006. 22(1): p. 186-98.

Degim, Z., et al., The effect of various liposome formulations on insulin penetration

US

13.

across Caco-2 cell monolayer. Life Sci, 2004. 75(23): p. 2819-27. Sambuy, Y., et al., The Caco-2 cell line as a model of the intestinal barrier: influence of

AN

14.

Toxicol, 2005. 21(1): p. 1-26.

Artursson, P., K. Palm, and K. Luthman, Caco-2 monolayers in experimental and

ED

15.

M

cell and culture-related factors on Caco-2 cell functional characteristics. Cell Biol

theoretical predictions of drug transport. Adv Drug Deliv Rev, 2001. 46(1-3): p. 27-43. Tavelin, S., et al., Applications of epithelial cell culture in studies of drug transport.

PT

16.

17.

CE

Methods Mol Biol, 2002. 188: p. 233-72. Sha, X., et al., Effect of self-microemulsifying drug delivery systems containing Labrasol

18.

AC

on tight junctions in Caco-2 cells. Eur J Pharm Sci, 2005. 24(5): p. 477-86. Deli, M.A., Potential use of tight junction modulators to reversibly open membranous barriers and improve drug delivery. Biochim Biophys Acta, 2009. 1788(4): p. 892-910. 19.

Dimitrijevic, D., A.J. Shaw, and A.T. Florence, Effects of some non-ionic surfactants on transepithelial permeability in Caco-2 cells. J Pharm Pharmacol, 2000. 52(2): p. 157-62.

20.

Ujhelyi, Z., et al., Evaluation of cytotoxicity of surfactants used in self-micro emulsifying drug delivery systems and their effects on paracellular transport in Caco-2 cell monolayer. Eur J Pharm Sci, 2012. 47(3): p. 564-73.

30

ACCEPTED MANUSCRIPT 21.

El-Sayed, M., et al., Transepithelial transport of poly(amidoamine) dendrimers across Caco-2 cell monolayers. J Control Release, 2002. 81(3): p. 355-65.

22.

Bu, P., et al., Cytotoxicity assessment of lipid-based self-emulsifying drug delivery system with Caco-2 cell model: Cremophor EL as the surfactant. Eur J Pharm Sci, 2016. 91: p.

US Food and Drug Administration. Food additive status list.

IP

23.

T

162-71.

CR

http://www.fda.gov/Food/IngredientsPackagingLabeling/FoodAdditivesIng redients/ucm091048.htm.

Prajapati, H.N., et al., Effect of difference in fatty acid chain lengths of medium-chain

US

24.

lipids on lipid-surfactant-water phase diagrams and drug solubility. The Journal of

Zhang, W., et al., In vitro transport characteristics of EFdA, a novel nucleoside reverse

M

25.

AN

Excipients and Food Chemicals, 2011. 2(3): p. 73-88.

2014. 732: p. 86-95. 26.

ED

transcriptase inhibitor using Caco-2 and MDCKII cell monolayers. Eur J Pharmacol,

Buyukozturk, F., J.C. Benneyan, and R.L. Carrier, Impact of emulsion-based drug

PT

delivery systems on intestinal permeability and drug release kinetics. J Control Release,

27.

CE

2010. 142(1): p. 22-30.

Hubatsch, I., E.G. Ragnarsson, and P. Artursson, Determination of drug permeability and

28.

AC

prediction of drug absorption in Caco-2 monolayers. Nat Protoc, 2007. 2(9): p. 2111-9. Shah, D.O., Micelles, microemulsions, and monolayers : science and technology. 1998, New York: M. Dekker. xvii, 610 p. 29.

McCall, K., Genetic control of necrosis - another type of programmed cell death. Curr Opin Cell Biol, 2010. 22(6): p. 882-8.

30.

Moquin, D. and F.K. Chan, The molecular regulation of programmed necrotic cell injury. Trends Biochem Sci, 2010. 35(8): p. 434-41.

31

ACCEPTED MANUSCRIPT 31.

Sastry, S.V., J.R. Nyshadham, and J.A. Fix, Recent technological advances in oral drug delivery - a review. Pharm Sci Technolo Today, 2000. 3(4): p. 138-145.

32.

Mu, H., R. Holm, and A. Mullertz, Lipid-based formulations for oral administration of poorly water-soluble drugs. Int J Pharm, 2013. 453(1): p. 215-24. Natoli, M., et al., Cell growing density affects the structural and functional properties of

T

33.

Natoli, M., et al., Good Caco-2 cell culture practices. Toxicol In Vitro, 2012. 26(8): p.

CR

34.

IP

Caco-2 differentiated monolayer. J Cell Physiol, 2011. 226(6): p. 1531-43.

1243-6.

Anderle, P., et al., Messenger RNA expression of transporter and ion channel genes in

US

35.

undifferentiated and differentiated Caco-2 cells compared to human intestines. Pharm

36.

AN

Res, 2003. 20(1): p. 3-15.

Ranaldi, G., et al., Permeability characteristics of parental and clonal human intestinal

37.

ED

Vitro, 2003. 17(5-6): p. 761-7.

M

Caco-2 cell lines differentiated in serum-supplemented and serum-free media. Toxicol In

Gupta, V., N. Doshi, and S. Mitragotri, Permeation of insulin, calcitonin and exenatide

PT

across Caco-2 monolayers: measurement using a rapid, 3-day system. PLoS One, 2013.

38.

CE

8(2): p. e57136.

Gupta, V., et al., Delivery of Exenatide and Insulin Using Mucoadhesive Intestinal

39.

AC

Devices. Ann Biomed Eng, 2016. 44(6): p. 1993-2007. Konsoula, R. and F.A. Barile, Correlation of in vitro cytotoxicity with paracellular permeability in Caco-2 cells. Toxicol In Vitro, 2005. 19(5): p. 675-84. 40.

Liang, E., K. Chessic, and M. Yazdanian, Evaluation of an accelerated Caco-2 cell permeability model. J Pharm Sci, 2000. 89(3): p. 336-45.

41.

Sevin, E., et al., Accelerated Caco-2 cell permeability model for drug discovery. J Pharmacol Toxicol Methods, 2013. 68(3): p. 334-9.

32

ACCEPTED MANUSCRIPT FIGURE LEGENDS

Fig. 1 Cell viability assessment of lipid-based self-emulsifying drug delivery systems on 1-day Caco-2 cell culture. Caco-2 cells (1-day culture) were incubated with medium-

T

chain monoglyceryl lipids (Capmul MCM EP or Capmul 708G) or propylene glycol

IP

monoester (Capmul PG-8 NF) in mixture with a surfactant (polysorbate 80) with and

CR

without a medium-chain triglyceride (Captex 355) at indicated composition ratios and concentrations (diluted in culture medium) for 2 h. Immediately upon completion of

US

incubation, cytotoxicity was assessed with MTT viability assay. Data are presented as

AN

mean ± S.E.M., n=5-6 replicates per each treatment. Asterisks (*) indicate statistically significant differences (p<0.05) compared to control; a and b, statistically significant

M

differences compared to 0.1% and 0.2% of the same formulation, respectively; c and d,

ED

statistically significant differences compared to the same concentration of the first and second formulation within the same table, respectively. Key: Tw, Tween 80 (polysorbate

PT

80); MCM, Capmul MCM EP; 708G, Capmul 708G; PG8, Capmul PG-8 NF; Cap,

CE

Captex 355.

AC

Fig. 2 Cell viability assessment of lipid-based self-emulsifying drug delivery systems on 5-day Caco-2 cell culture. Caco-2 cells (5-day culture) were incubated with mediumchain monoglyceryl lipids (Capmul MCM EP or Capmul 708G) or propylene glycol monoester (Capmul PG-8 NF) in mixture with a surfactant (polysorbate 80) with and without a medium-chain triglyceride (Captex 355) at indicated composition ratio and concentrations (diluted in culture medium) for 2 h. Immediately upon completion of

33

ACCEPTED MANUSCRIPT incubation, cytotoxicity was assessed with MTT viability assay. Data are presented as mean ± S.E.M., n=5-6 replicates per each treatment. Asterisks (*) indicate statistically significant differences (p<0.05) compared to control; a and b, statistically significant differences compared to 0.1% and 0.2% of the same formulation, respectively; c and d,

T

statistically significant differences compared to the same concentration of the first and

IP

second formulation within the same table, respectively. Key: Tw, Tween 80 (polysorbate

CR

80); MCM, Capmul MCM EP; 708G, Capmul 708G; PG8, Capmul PG-8 NF; Cap,

US

Captex 355.

AN

Fig. 3 Cell viability assessment of lipid-based self-emulsifying drug delivery systems on Caco-2 monolayers. Monolayers of Caco-2 cells after 21 to 23 days of culture were

M

incubated with medium-chain monoglyceryl lipids (Capmul MCM EP or Capmul 708G)

ED

or propylene glycol monoester (Capmul PG-8 NF) in mixture with a surfactant (polysorbate 80) with and without a medium-chain triglyceride (Captex 355) at indicated

PT

composition ratio and concentrations (diluted in culture medium) for 2 h. Immediately

CE

upon completion of incubation, cytotoxicity was assessed with MTT viability assay (modified assay procedure, see details in MATERIALS AND METHODS). Data are

AC

presented as mean ± S.E.M., n=5-6 replicates per each treatment. Asterisks (*) indicate statistically significant differences (p<0.05) compared to control; a and b, statistically significant differences compared to 0.1% and 0.2% of the same formulation, respectively; c and d, statistically significant differences compared to the same concentration of the first and second formulation within the same table, respectively. Key: Tw, Tween 80

34

ACCEPTED MANUSCRIPT (polysorbate 80); MCM, Capmul MCM EP; 708G, Capmul 708G; PG8, Capmul PG-8 NF; Cap, Captex 355.

Fig. 4 Assessment of membrane integrity on 5-day Caco-2 culture treated with lipid-

T

based self-emulsifying drug delivery systems. Caco-2 cells (5-day culture) were

IP

incubated with surfactant (polysorbate 80) alone or lipid-surfactant mixtures for 2 h

CR

followed by assessment of cell membrane integrity using MultiTox-Fluor Assay. RFU, relative fluorescence unit; Neg. control: cells received no lipid-surfactant mixture; Pos.

US

control: cells were treated with 30 μg/mL digitonin for 15 minutes prior to membrane

AN

integrity assessment. Data are presented as mean ± S.E.M., n=5-6 replicates per each treatment. Asterisks (*) indicate statistically significant differences (p<0.05) compared to

M

negative control; #, statistically significant differences compared to positive control; a

ED

and b, statistically significant differences compared to 0.1% and 0.2% of the same formulation, respectively. Key: Tw, Tween 80 (polysorbate 80); MCM, Capmul MCM

CE

PT

EP; 708G, Capmul 708G; PG8, Capmul PG-8 NF; Cap, Captex 355.

Fig. 5 Assessment of permeation enhancing effects of lipid-based self-emulsifying

AC

drug delivery systems using Caco-2 monolayer model. Lucifer yellow was used as a maker for paracellular transport. (A) Baseline levels of lucifer yellow permeation in the presence of Caco-2 monolayers without lipid-surfactant mixture (no lipid) and maximum passive diffusion of lucifer yellow when there were no cells present (no cell). (B and C) Effects of lipid-surfactant mixtures on lucifer yellow permeation across Caco-2 monolayers. Formulations contain the surfactant Cremophor EL (B) or polysorbate 80

35

ACCEPTED MANUSCRIPT (C). The lipid/surfactant composition ratio for each mixture is specified under each bar and the time points of sample collection from the basal compartment are 30, 60, and 120 min. The concentrations for all formulations was 0.2% in culture medium, except for “Tw 30+MCM 70 (0.5)”, which had a concentration of 0.5%. The concentrations of lucifer

T

yellow were converted from fluorescence reading according to a standard curve with

IP

known lucifer yellow concentrations established in the same experiment. The orange-

CR

colored bars in (B) indicate these formulations were shown to be toxic according to the MTT cell viability assay. The green-colored bars in (C) indicate that these formulations

US

contain only surfactant (Cremophor EL). Asterisks (*) indicate statistically significant

AN

differences (p<0.05) compared to no-lipid control; #, statistically significant differences between the two specified groups. Key: Tw, Tween 80 (polysorbate 80); MCM, Capmul

AC

CE

PT

ED

M

MCM EP; 708G, Capmul 708G; PG8, Capmul PG-8 NF; Cap, Captex 355.

36

ACCEPTED MANUSCRIPT

AN

US

CR

IP

T

Table 1. Chemical name, trade name, structure and composition of surfactant and lipids

*

AC

CE

PT

ED

M

Lipids studied in the present study do not exist as pure species, but rather they are mixtures of glycerides (or PG esters) with differing degrees of esterification and different fatty acid compositions [2]. Therefore, only the chemical structure of the predominant component of each lipid is presented in the table.

37

*

ACCEPTED MANUSCRIPT Table 2: Mean particle size of lipid-surfactant mixtures in aqueous buffer (HBSS) at different concentrations (v/v %) and time intervals.

Polysorbate 80 30% Capmul MCM EP 52.5% Captex 17.5% Polysorbate 80 30% Capmul 708G 52.5% Captex 17.5%

-

-

0.2%

25 ±12

24±6

30±18

0.5%

15±2

15±3

12±2

0.1%

202±28

182±5

185±9

0.2%

190±3

190±5

0.5%

260±28

269±17

338±22

0.1%

140±7

136±8

134±11

0.2%

149±13

146±15

146±18

0.5%

309±13

318±11

309±14

0.1%

148±10

144±13

135±9

0.2%

184±8

213±12

243±28

0.5% 0.1%

256±45 187±16

341±54 183±14

372±26 186±17

234±16

231±14

223±24

264±10

251±9

259±18

52±5

51±5

50±2

0.2%

26±2

25±2

26±2

0.5%

223±17

316±15

362±24

0.1% 0.2%

140±1 142±5

140±5 140±4

141±10 138±5

0.5%

153±8

155±13

161±9

0.1%

117±9

119±12

118±10

0.2% 0.5% 0.1% 0.2% 0.5%

64±6 41±2 106±10 65±11 41±4

65±4 41±2 105±7 63±7 41±4

65±6 42±2 107±12 64±8 44±3

0.1% 0.2% 0.5%

169±9 172±4 145±7

168±10 175±5 148±9

168±12 175±7 150±8

0.2% 0.5% 0.1%

CE

PT

Polysorbate 80 30% Capmul PG-8 NF 52.5% Captex 17.5%

AC

Polysorbate 80 30% Capmul MCM EP 35% Captex 35% Polysorbate 80 30% Capmul 708G 35% Captex 35%

Polysorbate 80 30% Capmul PG-8 NF 35% Captex 35% (-) = undetectable Values are mean ± SD (n=3)

*

38

IP

T

-

CR

Polysorbate 80 30% Capmul PG-8 NF 70%

0.1%

US

Polysorbate 80 30% Capmul 708G 70%

0 min

AN

Polysorbate 80 30% Capmul MCM EP 70%

Dilution

M

Polysorbate 80 30% HBSS 70%

Particle size (nm)* 60 min 120 min

ED

Formulation

215±35

ACCEPTED MANUSCRIPT Apparent permeability coefficient (Papp) 30 min 60 min 120 min 2.00±0.60

2.99±0.60

3.13±0.89

102.26±5.42

62.82±37.68

43.65±1.65

Tween 80 30% Capmul MCM 70%

5.84±1.25

10.87±1.92

10.14±1.79

Tween 80 30% Capmul PG-8 35% Captex 35% Tween 80 30%* Capmul PG-8 70%

6.72±1.53

11.80±0.76

12.54±1.08

11.03±1.59

12.15±0.43

IP

Group Monolayer + LCY

Tween 80 30%* Capmul PG-8 52.5% Captex 17.5%

8.89±1.04

12.35±1.23

12.43±0.82

Tween 80 30%* Capmul MCM 70% (0.5%)**

9.24±1.26

12.17±0.28

10.84±0.70

CEL 100%

2.48±0.24

CEL 30% HBSS buffer 70%

8.62±0.70

CEL 30% Capmul MCM 35% Captex 35% CEL 30% Capmul 708G 35% Captex 35%

13.14±0.79

16.09±0.50

4.34±0.41

10.64±0.53

11.78±1.65

9.02±1.00

10.73±1.01

2.97±0.30

7.34±0.84

9.73±0.84

M

6.88±1.92

ED

AN

US

CR

11.00±0.48

5.73±0.46

PT

4.50±0.24

CE

CEL 30% Capmul PG-8 35% Captex 35%

T

No cells + LCY

AC

Table 3: Apparent permeability coefficient (Papp, cm s-1 x10-7) of lucifer yellow (LY) as a function of specific formulation, concentration (0.2%) and time (n=4-5).

CEL, Cremophor EL * Formulations that were toxic to Caco-2 monolayers. ** At 0.5% concentration.

39

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

Fig. 1 Cytotoxicity assessment of lipid-surfactant mixtures on 1-day Caco-2 culture

40

ACCEPTED MANUSCRIPT Fig. 2 Cytotoxicity assessment of lipid-surfactant mixtures on 5-day

AC

CE

PT

ED

M

AN

US

CR

IP

T

Caco-2 culture

41

ACCEPTED MANUSCRIPT Fig. 3 Cytotoxicity assessment of lipid-surfactant mixtures on 21-day

AC

CE

PT

ED

M

AN

US

CR

IP

T

Caco-2 culture

42

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

CR

IP

T

Graphical abstract

43