Improving Relative Bioavailability through Drug-in-Adhesive Transdermal Patch of Duloxetine:MeβCD

Aim of the study was to develop optimised drug-in-adhesive (DIA) transdermal patch of duloxetine HCl. It is known that acrylic polymers having different functional groups play signicant role in enhancing drug permeation. Among various permeation enhancers (PEs), Transcutol P exhibited most enhanced permeation (ER ≈ 1.99) in terms of ux and Q 24 compared to control group having no PE. Hence, a transdermal DIA patch having DURO-TAK 87-2287 as DIA polymer and TP as PE loaded with 40% DLX previously complexed with MeβCD and duly characterised (FTIR, DSC and SEM) was developed for in vivo study. Mean of maximum plasma concentration (C max ) and area under time-concentration curve (AUC 0-72 ) in Wistar rats (n=6) for transdermal patch (10 mg/kg) was found to be 70.31±11.2 ng/ml and 2997.29±387.4 ng/ml*h respectively and these values were considerably higher than oral dose of DLX (20 mg/kg and 10 mg/kg). Albeit, T 1/2 was higher in case of transdermal delivery but this was due to sustained behaviour of delivery system. These ndings highlight the signicance of both inclusion complexation and transdermal delivery of DLX using DIA patch for ecient drug absorption and thus reducing the dose of drug and related side effects.


Introduction
Cyclodextrins (CDs) are cyclic oligomers of 6, 7 or 8 D -glucopyranosidic units, large molecular weight (1000 to > 2000 Da) and are poorly absorbed from biological membrane (French 1957). It represents the class of excipients having lipophilic internal cavity and hydrophilic outer surface comprised of (α-1, 4)linked α-D-glucopyranose units. Chair formation of these glucopyranose units renders CDs as cones shape having secondary and primary hydroxy groups extending from wider edge and narrower edge respectively. Such arrangement gives CD the desired shape with hydrophilic outer shape and internal hydrophilic cavity. Natural CDs are 6 membered (αCD), 7 membered (βCD), 8 membered (γCD) glucopyranose units. Substituted CDs include hydroxypropyl, dimethyl, randomly methylated, sulfobutylether-βCD. Human pancreatic amylase is reported to hydrolyse parent βCD, although get fermented in intestine and are non-toxic at low to moderate doses. Intrinsic reactivity of these enzymes and/or a nity of CDs to these enzymes get diminished on substitution of hydroxyl groups (Marshall &Miwa 1981).
DLX is an effective antidepressant with poor solubility pro le. Several attempts have been made to overcome this drawback of low oral bioavailability. Micro-emulsion loaded with DLX was prepared.
Permeation of DLX from micro-emulsion was 1.5 times more compared to DLX suspension. PK/PD study outcomes were also in complete agreement of permeation study data (Sindhu et al. 2018). Avoidance of rst pass metabolism along with predictable controlled transport is added advantage of TDDS. Albeit barrier property of skin is a deterrent which can be overcome by using Pes to appreciable extent and require lower initial concentration of drug for effective pharmacology (Hillery &Park 2016). For this purpose, we developed drug-in-adhesive (DIA) patches. DIA patches are comprised of active drug incorporated in pressure sensitive adhesive, a backing lm and release liner. It has advantages of smaller size, less thickness, better exibility and drug loaded adhesive layer remains in contact with skin surface after its application. DLX-SBEβCD spray dried inclusion complexes equivalent to 22.7 mg of DLX optimized on the basis of various studies were incorporated in place of pure DLX for enhanced permeation and reduction in dose.
Pressure sensitive adhesives (PSAs) are very critical in designing TDDS. Although, their main function is to adhere the patch to skin but they function as matrix for drug constituent and for most of excipients too. PSAs also affect the ux of drug, its release and physicochemical stability. In DIA systems, drug is mixed in polymeric mixture and skin itself act as rate-controlling membrane.

Characterization of fabricated patches
Patch thickness Thickness of the suitably cut DIA patches were measured at three different places with the help of a calibrated micrometer and the results were expressed as mean±SD.

Drug content uniformity
For content uniformity, each patch was rst separated from backing layer and release liner. The DIA lm was then cut into four equal quadrants. The rest of the procedure was same as discussed above.
Surface pH, Moisture content (MC) and Moisture uptake (MU) For determining surface pH, each lm was rst moistened and allowed to swell with one drop of distilled water and then electrode of pH meter was brought in contact with the even surface. For moisture content, active silica contained in desiccators was loaded with the DIA lms cut with area of 4.9 cm 2 for 3 days. During these days, each lm was weighed again and again until it shows constant weight indicating no residual moisture present. Finally, MC was computed taking into consideration the weight difference with nal weight. For moisture uptake study, fresh DIA lm was placed in saturated solution of KBr for 24 h.

Ex vivo permeation study of DIA patch
For this purpose, abdominal rat skin was used. Wistar rats of average weight 225 g were sacri ced and their abdominal skin was excised and to remove the hairs and underlying fat, skin was treated with 0.3 N ammonium hydroxide solution in normal saline. After removing underlying fat, skin was washed with normal saline to wash out residual amount of ammonium hydroxide. The skin was then mounted in between the donor and receptor chambers of Franz diffusion cell with the dermis facing the receptor compartment, previously lled with 30 ml of PBS (pH 6.8).

Pharmacokinetics of DLX from in blood
For carrying out pharmacokinetic studies, we used three groups of Wistar rats having 6 animals each.
Hairs from dorsal side of these animals were carefully removed without causing any damage to skin using commercial depilatory. Also, skin was washed with alcohol swab to remove any physical residue present on skin which could hinder the adhesion of formulation. A single medicated transdermal patch after removing the release liner was applied to clean and dry hairless skin for 72 h. DLX was administered via oral route in 20 mg/kg dose in suspension form, pellet of DLX:MeβCD spray dried complex equivalent to 10 mg/kg DLX and 10 mg/kg transdermal DIA patch applied on dorsal side of abdominal skin. After regular intervals, blood samples were withdrawn and collected in pre-heparinised Eppendorf vials upto 72 h post application of medication. Plasma samples were obtained by centrifugation (4500 rpm for 10 min) of blood samples and were stored at -20 ºC until analysed using validated HPLC method.
Pharmacokinetic parameters like C max , T max , T 1/2 MRT and AUC were calculated using PKSolver software performing non-compartmental analysis for oral doses and one compartmental modelling for transdermal dose. After analysis, DLX pharmacokinetics of transdermal dose was compared with oral doses of DLX and the data is presented as mean±SD (Su et al. 2020).

Histopathology
Histopathological examination on organs and tissues of all animal groups were conducted. After autopsy, the organs and tissues were removed and harvested to prepare histopathological slides according to the standard protocol. Organs and tissues were xed in 10% formalin solution. The organs observed were lungs, kidney, liver and spleen of animals of all groups.

Statistical analysis
Results expressed are mean±SD of three experiments. Difference of statistical signi cance among various groups was determined using GraphPad Prism 7.0 software. One-way ANOVA followed by posthoc test was used and comparisons were deemed signi cant when p-value was lower than 0.05 (p < 0.05).

Determination of stoichiometry and apparent stability constant
Initially a calibration curve of DLX was constructed at 289 nm using methanol as solvent (R 2 =0.996±0.02). Higuchi and Connors have classi ed the phase solubility diagram into two types i.e. "A" and "B" on the basis of type of graph obtained. Type "A" exhibits when substrate solubility linearly increases with increase in ligand concentration and the slope was less than unity over the concentration range. Such graph can be further classi ed as A L , A P or A N type corresponding to linear, positive and negative curvature respectively. As we can easily observe in the graph ( gure 1) to exhibit A P type and is suggestive of formation of inclusion complex in 1:1 stoichiometric ratio. It is evident from the thermograms that all samples exhibited endothermic peaks as presented by heat ows. Also, as our previous publications shows ( Kumar et al. 2020), DLX has sharp endothermic peak at 169 ºC with T onset = 168.43 ºC with ∆H = 28.36 J/g which corresponds to pure DLX compound indicating a typical crystalline state. MeβCD has shown a broad peak at 62 ºC corresponding to loss/volatilisation of water molecule present/trapped in the inner cavity of MeβCD. Another peak at 171.59 ºC presented the thermal decomposition of host molecule. While analysing the DSC of physical mixture, endothermic peak of MeβCD remained unchanged but an additional peak at 170.54 ºC could be visible and this may be attributed to presence of DLX. It indicates no interaction/complexation between drug and MeβCD occurs in this system. So far as DSC thermogram of spray dried inclusion complex of DLX:MeβCD is concerned, dehydration signal of MeβCD was seen to be smaller as activation energy required for its dehydration gets changed after interaction with guest molecule. Also, endothermic peak of DLX gets shifted due to its partial complexation. But no volatilisation peak of DLX was observed indicating formation of inclusion complexation in solid state (Santos et al. 2017). This type of DSC curve of spray dried complex is indicative of molecular encapsulation of drug in the hydrophobic cavity of MeβCD resulting in formation of an amorphous complex.

Selection of adhesive for formulating DIA patch
Adhesives can be multifunctional in TDDS. PSAs are the adhesives forming a bond with substrate on application of minor pressure. These adhesives leave no residue on the substrate after their removal. Viscoelastic materials have pressure sensitivity characteristics. PSAs form no chemical bond to adhere to substrate but form intermolecular and can be used without using solvents without Acrylic acid based PSAs were used for preparing the DIA patches. PSAs having different functional groups were used as shown in table 2.

Effect of adhesive type on the permeation characteristics of DLX
Selection of PSA is critical to the transdermal delivery of DLX. We considered three types of acrylic acid based PSAs for our study having different functional groups (see table 2). Therefore, every PSA was mixed with equal amount of DLX i.e. 10% and the release behaviour of the prepared DIA patches was observed by keeping all factors constant. Thickness of the patches was also kept uniform and equal for all three PSAs. Impact of adding different PSAs on the release behaviour of DLX is shown in gure 5 and different parameters like ux and Q 72 derived from the release pro les are tabulated in table 3 respectively. *Therefore, on the basis of In vitro release study of DLX, DURO-TAK 87-2287 was selected for carrying out further studies.

Effect of drug content on DLX permeation
After screening the optimum PSA for incorporating the drug to prepare DIA patch, DURO-TAK 87-2287 adhesive was loaded with different DLX content i.e. 5, 10, 20, 30 and 40% w/w of DLX. The effect of drug content on the in vitro permeability was evaluated as shown in gure 6 and various permeability parameters have been calculated and tabulated in table 4 to optimise the desired concentration of DLX for loading in DIA patch development.
It is evident from the gure 6 and table 4 that DLX ux from the skin increased with increase in its content in the DIA patch. Up to 40% DLX could be loaded in the DIA patches prepared with DURO-TAK 78-2287 as PSA but above this concentration, drug crystallisation starts as evident in section of drug crystallisation studies observed with the help of optical microscope. Such observation may be attributed to the maximum solubility of DLX in the PSA. Therefore, on the basis of these results, 40% DLX was selected for loading the drug into optimized PSA i.e. DURO-TAK 78-2287.

Selection of penetration enhancer
For selecting the optimum penetration enhancer, different PEs were mixed in xed 40% w/w concentration of DLX (table 5) and DIA patch were prepared.
One formulation was having no PE and treated as control group. Various PEs tested include oleic acid (OA), Brij 98, Transcutol P (TP) and isopropyl myristate (IPM). Different PEs were added in the same concentration. The results of ex vivo permeation study are shown in gure 7 and various permeation parameters are tabulated in table 6 and indicate that TP, Brij 98 and IPM were almost showing same enhancement in the permeation of DLX across the skin. Among these PEs TP and Brij 98 were almost equally capable of enhancing the permeation demonstrated by the enhancement ratio (ER) having values of 1.986 and 1.988 respectively. Oleic acid is a fatty acid added in formulation DIA-2, can also act as crystallization inhibitor which can be the reason for increased ux of DIA-2.
*Therefore, on the basis of ex vivo permeation studies, TP was selected as suitable permeation enhancer (PE) for developing nal DIA patch.

Characterization of DIA lms
Spray dried complexes of DLX loaded DIA patches were characterised for various physicochemical properties and the results are shown in table 7. All lm batches exhibited fair organoleptic properties like colour and transparency. The per cent drug loading or drug content was found to be in the range of 95.4±2.8 (for DIA-3)-98.7±4.2 (for DIA-4), although other formulation groups also exhibited fair drug loading of more than 96%. The uniformity of drug content and low SD values indicate the robustness of formulation development methodology.
Low value of moisture content of all lms is also indicative of its stability during storage as well as its non-brittleness as the complete dryness is not attained during storage. Low values ranging from 0.55±0.003 (for DIA-4) to 1.03±0.005 (for DIA-1) re ect formulated DIA lms to be enough stable.
Interestingly, moisture uptake for all lms was also very low ranging 0.54±0.01 (for DIA-5) to 1.1±0.012 (for DIA-1). Such low values of moisture uptake assure that the formulations will not attract microbial contamination and will certainly deter bulkiness. Surface pH of all lms ranged from 6.2±0.03 (for DIA-5)-6.6±0.06 (for DIA-3) which indicate closeness to pH 7 near to skin pH and thus lms are suitable for transdermal application and probably would not cause any local irritation during or after application (Su et al. 2020).
We conducted different studies to observe the physiochemical and skin permeability of DLX across the skin. Our ndings led to the conclusion that DIA formulations have better drug loading and better drug permeation across the same skin surface area used compared to matrix type of transdermal patches. Among DIA formulations, the outcomes of ex vivo permeation studies conducted using skin samples of Wistar rats, it was evident that formulation DIA-4 containing TP as permeation enhancer was having signi cant permeation enhancement compared to control group having no PE. Therefore, formulation DIA-4 was selected for conducting further studies. Here onwards, all studies are conducted using DIA-4 as optimized formulation.

In vivo pharmacokinetic study
Under optimum HPLC conditions, method developed exhibited fair selectivity and speci city for DLX. Good linear regression for DLX was observed with R value of 0.997 (over the linearity range of 0.5-20µg/ml with retention time of 3.13 min (Kumar et al. 2021). We conducted pharmacodynamics study where DLX could alleviate depression progression in rat CUMS models. Median lethal dose of DLX (LD 50 ) in female rats is 279 mg/kg, so we determined DLX pharmacokinetics at dose equivalent to 10 mg/kg (DLX-MeβCD pellet) and 20 mg/kg (pure DLX in suspension form) for oral route.
As we try to circumvent hepatic rst pass metabolism of DLX via transdermal route expecting 100% bioavailability of DLX through this route, the dose of the drug can be reduced to 10 mg/kg in place of 20 mg/kg. Therefore for administering DLX via transdermal route we administered dose equivalent to 10 mg/kg in laboratory animals (group 3). Plasma concentration-time pro les of DLX after oral and transdermal administration are reported in gure 8 and different parameters calculated are summarised and compared in table 8 respectively.
As shown in gure 8, after administration of oral dose at 10 mg (pellet of DLX complexed with MeβCD) and 20 mg/kg (suspension form), DLX was seen to be absorbed and excreted rapidly from the body. Mean of peak plasma conc. was 41.7±5.5 and 34.7±8.3 ng/ml respectively. AUC 0-72 was 646.27±66.3 and 1332.36±232.7 (ng/ml)*h and it increased non-linearly for doses extending from 10 mg/kg to 20 mg/kg. As it is seen that T 1/2 and T max values did not appreciably change for both doses, the change in So far as DLX in its complexed form with MeβCD is concerned, transdermal patch was applied and blood samples were analysed at different time intervals. C max (70.31±11.2 ng/ml) and AUC 0-72 (2997.29±387.4 ng/ml*h) was found to be improved via this route compared to oral administration of naïve as well as pellet dose of DLX. This increased AUC value clearly indicates signi cance of transdermal route in enhancing bioavailability of DLX. If we consider this bioavailability to be 100%, then relative bioavailability (dividing AUC of oral route by AUC of transdermal route x 100) through oral route for 20mg/kg dose of DLX would be merely 44.45%, which is near to reported bioavailability of DLX through oral route. This nding demonstrates signi cance of transdermal route in enhancing bioavailability of DLX and thus reducing the dose of DLX. T max for reaching C max was approximately 4±0.7 h for transdermal administration. C max obtained from transdermal patch was even higher than that obtained from oral administration of 20 mg/kg suspension which indicated the signi cance of this route in circumventing hepatic metabolism and signi cance of inclusion complexation process. But as the C max is higher, so the side effects related to high plasma drug concentration may not be reduced using transdermal delivery of DLX. Also, increased T 1/2 (35.04±5.2) for transdermal delivery compared to oral delivery of naïve DLX (22.2±3.7) and pellet of DLX-MeβCD spray dried inclusion complex (21.02±3.2) indicated a steady DLX plasma level to a certain extent and a sustained delivery of DLX to extended period of time in vivo. Also, increased MRT for transdermal delivery (53.43±5.1) compared to same dose of DLX (27.8±3.5) is indicative of practicability of sustained behaviour through transdermal delivery.
It is interesting to see that transdermal delivery ex vivo exhibits around 11025.8±343 µg/cm 2 (for a loading dose of 22.7 mg equivalent to 20 mg duloxetine) in 24 h while results of transdermal delivery in vivo indicated DLX to release for 72 h (for dose of 10 mg/kg DLX). This means that skin permeation in real animal body is lower than ex vivo models. This warrants some more study to establish better in vitroin vivo correlations.

Histopathological examination
As shown in gure 9, different tissues of different treated groups did not show any remarkable histopathological changes compared to control group. Liver of control group (a), at 100X showed normal hepatic architecture of liver along with lymphocytes except a mild portal tract while at 400X (b) and at low power (c), thickening by lymphocytes and possibly brosis is seen. There is a ne particulate matter in or over liver cells that may be an artefact. Low dose group (d) at 40X, looks normal at low magni cation, while (e) at 100X, mild fatty change in liver cells and (f) at 400X, further details are provided. At (g) 40X, liver of high dose group looks normal at this low magni cation, but at (h) 400X, marked pathological changes characterized by an increased cellularity of portal tract and extension towards the next portal region is seen. This may suggest chronic hepatitis.

Conclusion
Spray drying, loading of complexed drug in transdermal matrix followed by enhanced penetration and permeation advocate the potential application of inclusion complexation. The developed patch was stable, non sticky, and had a fair folding endurance. Ex vivo studies suggested DURO-TAK 87-2287 based patches to better permeate across the skin. With these interesting outcomes, it is evident that the said technique can work for better permeation and retention through drug amorphization and its concomitant inclusion. In vivo pharmacokinetic study in rats indicated a steady sustained absorption for transdermal delivery of DLX with enhanced AUC and T 1/2 suggesting obvious advantages of this route over oral administration. Therefore, it is assured that DIA patch transdermal delivery can help in reducing dose and Credit Authorship Contribution statement V.R.S. and A.S. designed the research study. R.K. performed majority of experimental protocols and was a major contributor in writing rst draft of manuscript. L.D. and T.S. contributed to reviewing and editing. All authors read and approved the nal manuscript.

Acknowledgement
Authors are thankful to Cyclolabs, Budapest Hungary for providing gratis samples of different cyclodextrin derivatives and department of SAIF/CIL, Panjab University Chandigarh for carrying out different analysis.

Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Ethical approval
All protocols involving laboratory animals are approved from institutional animal ethical committee, Panjab University, Chandigarh.

Consent for publication
Not applicable.

Consent to participate
Not applicable Figure 1 Phase solubility study of DLX in aqueous solution of MeβCD at 25±2 ºC and 37 ºC    In vitro hairless rat skin permeation pro les of DLX from patches containing 10% (w/w) DLX in PIB 87-900A, 87-235A and 87-2287 (n=3). Each point and vertical bar represents mean and standard deviation, respectively Figure 6 In vitro permeation pro le of DIA patches containing different DLX content, 5 (■), 10% (▲), 20% (x), 30% (*) and 40% (•). Each data point represents mean of three experiments and each vertical bar represents standard deviation Figure 7 Ex vivo permeation results of DLX containing various PEs. One batch (NoPE) was a control batch having no PE. All values represent mean (n=3) and each vertical bar represents standard deviation Figure 8 Mean plasma concentration-time pro le of DLX after oral (10 mg/kg and 20 mg/kg) and transdermal (10 mg/kg) administration in a single dose study (mean±SD, n=6)