Optimization and Scale-up Methodology in Preparing Microsponge Loaded with 5-Fluorouracil (5-FU).

The present investigation aims at developing models by response surface methodology (FCCCD) followed by the scale-up method in preparing control release microsponge particles loaded with 5-fluorouracil, a drug used to treat actinic keratosis and colon cancer, and producing a new Dermal Delivery System. The polymer-based (ethyl cellulose and eudragit RS 100) microsponge particles were prepared by the w/o/w double emulsification method. The optimized product was formed with the combination of independent variables levels: polymer (600 mg), stirring speed (1198 rpm) and surfactant (2% w/v), yielding responses as yield (~63.6257%), the average size of particles (~151.563 µm), entrapment efficiency (~75.319 %) and drug release in 8hr (~75.75%), with desirability value of 0.737. The products showed similar responses as obtained in scale-up work. FT-IR, DSC and SEM studies confirmed the drug's compatibility with polymers and porous morphology. Finally, gel embedded optimised product showed shear-thinning rheological property, ideal for drug release from the thixotropic gel.


Introduction
The Microsponge Drug Delivery system is unique among other control release microparticulate systems. It has high loading capacity in myriad microporous channels of its spongy structure, self-sterilising ability, and assuring thermal stability and chemical stability in a wide range of pH, making it flexible in developing improved product forms. Diffusion of active material from the porous structure of microsponge particles is triggered due to its solubility in an aqueous medium such as perspiration in topical drug and cosmetics products (antiseptics, deodorants and antiperspirants). Microsponge particles act as a storehouse of the drug and release drug molecules slowly at a controlled rate. The prolonged-release condition reduces the toxicity and allergic effects and improving patients' compliance.
The microsponge preparation method consists of emulsification, solvent evaporation and solidification. Types of emulsion (w/o, o/w, w/o/w) and the solvent are chosen depending on the characteristics of drug and polymer used. The type of emulsion is reported as o/w (aqueous external phase) in most of the earlier cases of microsponge preparation which makes the process economic provided drug loss during trial is minimum. Oil in oil type emulsion had been reported by several researchers [3,8,14] using Eudragit RS 100 in the preparation of microsponges. On the other hand, very few reports are based on w/o/w double emulsion, which is specially adapted to make the product more stable [16]. Eudragit RS 100 and ethyl cellulose are two structural components widely shown in earlier literature for porous structure formation. These are Food and Drug Administration (FDA) approved, safe, non-irritating, non-toxic and economic excipients, and widely used in the pharmaceutical industry.
Moreover, skin toxicity due to polymer use can be assessed by conducting cell line toxicity studies and in vivo skin irritation studies.
The use of Eudragit polymer had been reported in the preparation of microparticles by several investigators [3,7,8,14,[17][18][19][20][21]; use of ethylcellulose polymer was found in several earlier works [5, 6,12,16,[22][23][24][25][26]; eudragit RSPO by P.M. Barde et al. [27], as a single polymer use. A combination of polymers (ethyl cellulose +HPMC) was used by Yasser Shahzada et al. [10] and Jain S.K et al. (ethyl cellulose +eudragit RL 30D) [4]. The coupling of different polymers offers better control of drug release behaviour. The emulsification process often needs the addition of suspending agents like Na alginate [19] to facilitate the dispersion of polymer droplets in the emulsion. PVA as a stabilising agent in the external aqueous phase had been reported in most of earlier cases. Ahmed U. Ali et al. [28] used PEG 4000 solution (0.02% W/V) in water (outer phase). Besides polymers, plasticiser (triethyl citrate) is used to reduce fragility, and pore inducers/ porogens (hydrogen peroxide or sodium bicarbonate, gelatinised starch) are used for increasing number of pores to accommodate higher amount of drug. Other agent is surface active agent (tween 80) used to emulsify. Apart from microsponge composition, preparation techniques play an important role in regulating the performance of this delivery system. However, complexity arising due to physicochemical properties of ingredients and fluid dynamics in emulsification method and cost of preparation are the limitations in the production of microsponges. Preparation of multi particulates by 'emulsion solvent evaporation technique is a complex one as it involves fluid dynamics phenomenon and solidification of emulsion droplets (phase change) leading to loss of its emulsion character finally. Reproducibility is often questionable in repeating experiments on higher scales where it is difficult to maintain similar fluid dispersion dynamics. Therefore, controlling variables should be identified aptly by prior trials.
Optimisation by QBD for preparing microparticulate systems is nowadays practised by formulation scientists, and optimised operating variables aid in scale-up design. To enhance reliability and reproducibility of the method, several investigators started attempting simple factorial design [29], Central Composite Design -RSM method by design expert software [1,11,13,16,[28][29][30][31]. They presented linear as well as quadratic models.
The lack of knowledge /information related to the scaling-up of technology used for preparing polymeric microsponge may hamper/delay the launching of the product into the pharmaceutical market. Very few works have been reported on the scale-up method. In 2013, S. A. Galindo-Rodr´ıguez et al. [32] adopted scale-up technology in small batch size up to 1.5 L producing ibuprofen-loaded nanoparticles. They reported satisfactory results in scale-up with a slight difference in particle size of products. Effectiveness of the scale-up procedure had been reported by L. Sánchez-Silva et al. [33] at pilot plant scale using the optimal formulation of microcapsules polystyrene as found at lab-scale. However, there is a paucity of scale-up data in the literature. In 2012, K. Mitri et al. [34] attempted the scale-up of nanoemulsions (NEs) produced by emulsification and solvent diffusion process. They established two power-law relationships between droplet size and Reynolds number; and droplet size and shear stress; and compared nanoemulsion droplet diameter in laboratory and pilot scale.
The present study is emphasised on optimisation of the method of preparation of a new formulation of microsponges (MS) loaded with hydrophilic drug 5-fluorouracil (5-FU) by w/o/w emulsion solvent evaporation method to ensure reproducibility of the product and extending its potentiality to scale up. The results obtained during this optimisation made a starting point for the second stage of this study. To our knowledge, this is the first report on the scale-up approach in the preparation of 5-FU loaded microsponge delivery system. 5-FU is chosen as the model drug in the present work, one of the most potent chemotherapeutic drugs. 5-FU is a fluorinated pyrimidine antimetabolite structure similar to that of the pyrimidine molecules of DNA and RNA; 5-FU interferes with nucleoside metabolism and is converted within the cells into 5-fluorodeoxyuridine monophosphate, which constrains the synthesis of DNA, leading to cytotoxicity and cell death.
This study aims to develop and optimise the preparation method of 5-FU loaded microsponge particles by the design expert software, characterise products, and extend its possibility to scale up.

2.2.Method of preparation of Microsponges
Accurately weighed 50 mg of drug 5-FU was mixed with a specified amount of polymer mixture (ethylcellulose and eudragit RS 100, in the ratio 1:1) and then dispersed in 15 mL of a mixture of solvents (DCM: ethanol,1:1 v/v) (inner phase). The inner phase was sonicated for 30 minutes to make homogeneous dispersion. 100 mL of aqueous solution of sodium alginate (0.4% w/v) was prepared and then surfactant, tween 80 (0.5-2% w/w) was added to it (external phase). Next, 1 mL of 1% (v/v) of external phase was added to the inner phase and mixed by a cyclo mixer (Remi, CM 101) to prepare w/o emulsion. Then this primary emulsion was added dropwise to the external phase followed by stirring continuously in a mechanical stirrer (Remi motor RQT-124A) at a specified rpm for 4 hr. Upon complete evaporation of solvent during stirring, droplets get hardened, and solid microsponge particles were isolated by filtration (Whatman-150 mm filter paper). The product was dried in a hot air oven for 6 hr. It was stored in desiccators till further study.

2.3.Method of preparation of gel incorporated with 5-FU-microsponge particles
The gel was prepared with carbopol 934, which is a water-soluble polymer. First, accurately weighed carbopol 934 (0.25% w/v) was mixed with double distilled water (DDW) using a magnetic stirrer at 1200-1400 rpm for 45 min. Then a batch of 10 mg of microsponge particles was incorporated in 1 gm of prepared gel with slow stirring for equal distribution.
Then triethanolamine was added 1 to 2 drops to adjust pH 5.5-6. Microsponge particles embedded in carbopol gel were kept overnight, and then it was used for in-vitro release study.

2.4.Experimental Design
Experimental design is the process of planning a study to meet specified objectives. Planning an experiment properly is essential to get reproducible data. This could eliminate the timeconsuming phase, which could not be achieved with the conventional empirical method.
Among various designs, the Central Composite Design (CCD) is well suited for fitting a quadratic surface in process optimisation. Response surface methodology, a relation between factors and responses, was used for the experimental design and optimisation with minimum runs of the experiments [35]. This study used' Design Expert 13 version (Stat-Ease, USA) [36][37][38][39] where Y i is the measured response; b 0 is an intercept of the polynomial equation, representing the model's coefficient. b 1 -b 9 represent regression coefficients of main effects (X 1 , X 2 , and X 3 ), interacting effects (X 1 X 2 , X 1 X 3 , and X 2 X 3 ) and quadratic effects (X 1 2 , X 2 2 , and X The ranges or levels of independent variables were determined through preliminary trials and displayed in Table 1. The value of α was fixed at 1 for face-centred design. Each variable in the design was studied at three different coded levels (−1, 0, 1).  DSC thermogram depicts the profile of heat flow vs temperature.

Scanning Electron Microscopy (SEM)
The surface morphology, shape and size of microsponge particles can be analysed using Scanning Electron Microscopy (Carl-Zeiss, SEM, Tokyo, Japan). Particles were mounted on a metal stub with conductive tape. Particles were coated with a thin coating of platinum under reduced pressure.

Determination of yield (%)
Each batch of dried microsponges was weighed accurately, and yield was calculated as a percentage using the following equation: (%) *100 weight of microsponges Yield weight of polymer weight of drug   (2)

Determination of Drug Entrapment Efficiency
Drug entrapment efficiency was determined by adopting the solvent extraction method. First, the amount of the drug was estimated in a UV-VIS spectrophotometer (ANALAB UV -180).
Then, accurately weighed 10 mg of microsponge particles was dissolved in 5 mL of methanol 2.5.6. Particle size analysis The average particle size of microsponges for 50 particles of each batch (in run design) was measured by optical microscope (GOKO-Miamb, Japan). First, average particle size was determined.

Study of in vitro drug release (diffusion) for gel containing microsponge
In vitro, drug release studies were carried out using Franz's diffusion cell (Remco, India) at The aliquots were assayed in a UV spectrophotometer (ANALAB UV-180) to determine drug concentration at λ max 265 nm. The cumulative per cent release (CPR) was plotted against time.
Each experiment was repeated thrice.

Determination of kinetics of drug release from the gel
To understand the mechanism of drug release from gel loaded with microsponge formulations, the release data were computed to various mathematical models: zero-order ) and Hixon Crowell model ( ) to evaluate the drug release mechanisms [40]. Where, Q o = initial amount of drug release; Q t = amount of drug release at time t; k o , k 1 , k H , k P and k HC are release rate constants of each model equation; M T /M α = fraction of drug release at time t, and n is release % 100* Actual drug content of microsponges DEE Theoretical drug content of microsponges [40]. The following plots were constructed: Q t against t (zero order), [ln Q t -lnQ 0 ] against t (First-order kinetic model), Q t against t 1/2 (Higuchi model), log (M t /M α ) against log t (Korsmeyer-Peppas model), and cube root of drug amount remaining in dosage form against time (Hixson-Crowell model).

Micrometric properties of microsponge formulations
The specific quantity of particles was poured into a 5 mL graduated measuring cylinder, and the volume of initial packing was noted. The bulk density was determined by dividing the weight of the sample by the volume of initial packing. Tapped density of the particles is the ratio of the mass of the powder to the volume occupied by the particles after it has been tapped for a defined period. Tapping was continued until no further change in volume. It was determined by dividing the weight of the sample by the volume of packing after tapping.
Hausner's ratio (H r ) is a number that is correlated to the flowability of a particle. It was calculated by dividing tapped density and bulk density. Carr's index % (CI) is an indication of the flowability of particles through a hopper. The formula calculated it: (1-1/H r ).

Study of the rheology of gel :
The topical drug in gel form needs adequate consistency to maximise the contact period between the medication and the skin. This can be accomplished by modifying the nature of the vehicle. Therefore, studying the rheology of the product gel is necessary to know its consistency and rheological behavior [41]. Thus, the rheology of gel loaded with microsponge particles was studied by Rheometer (Anton Paar, Austria). Rheological characteristics plots such as strain against viscosity graph, strain against G` (storage modulus) and G`` (loss modulus) and angular frequency vs G` and G`` were generated using Rheoplus/32 Version 3.

3.1.Parametric Sensitivity and Optimisation
In the present study, the drug 5-FU was embedded in polymeric microsponge by quasi emulsion solvent evaporation method. Few controlling variables were chosen: polymer amount, stirring speed, surfactant concentration, and their effects on response variables were studied. A design of factors' combinations was generated by Design expert version 13 software. Accordingly, 20 batches of products were prepared, and products were characterized to obtain response variables (yield%, average particle size, drug entrapment efficiency% and drug release % in 8 hr). The statistical software analyses the results to generate ANOVA, fit summary and model equations, various graphs using the linear regression square root method to avoid the lack of fit analysis. The results for all the cases are displayed in Table 3.   and then increases at a higher level. In the case of particle size factors, SA and SS have more control over particle size.
In the case of EE (%), factors B and C have more control over the entrapment of the drug than that of polymer amount. Figure 1. c (2D contour plots and 3D response surface plots) illustrates the effects of two controlling variables (B against C) on the responses while keeping the third factor (A=400) constant at the centre level. EE% decreases with the increase of SA% up to its centre level then increases, and EE% decreases with the addition of stirring speed (B) at any specific value of the C factor.
In the case of Rel 8hr (%), factors A and C have more significant effects on the release of drugs from microsponge particles. From the plot (figure 1.d) (A against C at B=1000) and the model equation, it was observed that Rel 8hr (%) increases when SA(%) was increased up to its centre level. It decreases at a higher level of C, at any specific level of factor A factor. On the other hand, the increase of A factor decreases release% up to the middle level and increases at a higher level, at any fixed factor C factor.
On further development, Fit statistics represented in Table 3 suggests that the Predicted R 2 is in reasonable agreement with the Adjusted R 2 for the case of yield %, average particle size, EE% as the difference is less than 0.2. But in the case of drug release % in 8 hr, the difference is not in close range, which may indicate a significant block effect or possible complications with the block of data. But as the Adequate Precision (ratio of the signal to noise) is 28.035, which is greater than 4, it suggests an adequate indication and can be used for further process.
Coefficients in terms of coded factors, displayed in  and AICc (Akaike's Information Criterion).        Table 6. The desirability is 0.737 for all the cases represented in figure 2. A high value of desirability coefficient (0<y<1) indicates that the operating points can produce an acceptable formulation.
Based on input conditions of factors (Table 6), the design expert software generated a list of solutions.

3.2.Scale-up for more significant batch production of 5-FU microsponge gel
Reproducibility of a method can be checked for a larger batch size of the product by scale-up technique. To convert the formulation prepared in small scale to higher scale production, geometric similarities were maintained as much as possible with the power-law approach. The same stirring speed had been maintained on a larger scale. Specifications maintained in scaleup were displayed in Table 8. System geometry of scale-up was beakers diameter (6. showed similar characteristics as that of optimised batches, such as yield, particle size, entrapment efficiency and Rel8hr (Table 9). To verify the method's effectiveness, an approach was made to apply it to a higher scale. Table 9 shows the input data and response variables for three batches with increasing volume.
The optimised formulation was produced in 100 mL of the continuous phase, and the scale was increased to 100mL:200mL:800 mL to check the reproducibility of quality and yield % of the product. The same system geometry (S 1 , S 2 ) was maintained at a larger scale. Reynolds number is doubled on the larger scale. Similar rpm (1200 /min =20/sec) was kept on a higher scale as per power law (N 2 /N 1 = (D1/D2) n ) by putting n equal to zero. Table 9 gives the evidence that products formed in scale-up volumes were similar concerning yield%, average particle size, entrapment efficiency and cumulative release% in 8 hr. Like any other chemical industry ,pharmaceutical manufacturing units too require to adopt scale up from the laboratory to the pilot plant to total production. The transition from one scale to another may cause alterations in macroscopic and microscopic properties of formulation components and products at different production scales [42]. The right Power-law approach in conjunction with geometric similarity [43]  Thus it saves money and time in unproductive trial tests.

3.3.Drug release study of optimised and scaled up batches:
The study of drug diffusion is crucial to assure its release from microsponge products obtained from the optimised and scaled up batches. Release study was conducted in Franz diffusion cell containing receptor medium as aqueous phosphate buffer of pH 7.4 and pH 5.5 to mimic diffusion in the oral and dermal systems respectively. Drug load is ~625 µg / 10 gm microsponge per 1 gm of gel formulation. The content of the drug in each microsponge sample used for the release study is nearly the same. Therefore, the release pattern shown by the profiles were observed as identical (vide Figure 3.). Drug release in the buffer of pH 5.5 was observed ~4% higher than that of pH 7.4 as the drug is weakly basic.

Spectroscopy) and Differential scanning calorimetry (DSC)
The FTIR spectra of drug, polymer, drug-polymer physical mixture and optimised microsponge formulation are given in the figures below. Figure 4 shows the FTIR of 5-FU (drug), ethylcellulose, Eudragit RS 100, the physical mixture of drug-polymer (1:1) and microsponge formulation. Each spectrum was a plot as Wavenumber (cm -1 ) against %Transmittance.
The spectrum of pure 5-FU (Fig. 4a) showed characteristics peaks of 1648.5 cm -1 and 3067. Differential scanning calorimetry or DSC is a technique in which the difference in the amount of heat required to increase temperature (30-500ᵒC) of a sample and reference is measured as a function of temperature. Thus, both the sample and reference are maintained at nearly the same temperature throughout the experiment under a nitrogen purge of 25mL/min. Generally, the temperature program for a DSC analysis is designed such that the sample holder temperature increases linearly as a function of time.
The DSC thermogram of pure 5-FU is shown in Figure 5 Figure 6 depicts the morphology of microsponge particles with a rough and fine porous surface. The rough surface of microsponges is due to high polymer content, shearing effect and solidification of large-sized globules in the emulsion at a high solvent evaporation rate.

Surface Morphology of microsponges (SEM)
The particle's surface appears mor to be more porous after 1hr of drug release as drug elutes from porous structure; pores were non-uniform and irregular and more extensive in shape due to erosion of the drug.

Rheology of Gel
Rheology  Figure 7, the flow curve of the gel is shown. The viscosity of gel was high at a meagre shear rate (953 Pa.s at 0.0018 sec -1 ). Viscosity is dropped at higher rates of shear rate (9.93 Pa.s at 100 sec -1 ). This is the ideal phenomenon of gel (shear thinning), which becomes more extensive as the shear rate increases. It showed pseudoplastic properties because viscosity decreased after increasing shear rate, which causes better drug release. Figure 7 showed amplitude sweep analysis. It was performed to assess the linear viscoelastic range and viscoelastic properties of the polymer. The applied strain range within which G' and G" remain constant represents the linear viscoelastic range (LVE). The strength of the gel was so high that's why it's linear part is longer. At specific strain (50.1%), G' was declined. It was suggested that from this amount, strain% breakdown of the structure started. G" is more significant than G' which indicated that gel was highly structured with elastic characteristics.
Usually, the rheological properties of a visco-elastic material are independent of strain up.
Beyond this critical strain level, the material's behavior is non-linear, and the storage modulus declines. So, measuring the strain amplitude, the dependence of the storage and loss moduli (G', G") is an excellent first step taken in characterising viscoelastic behavior: A strain sweep will establish the extent of the material's linearity. In this graph, G" is more significant than G' indicating the gel becomes progressively more fluid-like and the module decline.
Frequency sweep analysis within the LVE range obtained from the amplitude sweep test indicates the structural integrity and mechanical strength of material more precisely and accurately [44] (Figure 7). The structural integrity of the sample was determined by the structural response to deformation at longer and shorter oscillatory stress (100-0.1 rad/sec).
Higher values of storage modulus (G') over the loss modulus indicate a solid elastic gel.
Higher yield stress due to the sample was unable to show any crossover point. Moreover, the absence of any crossover region stated a lack of gel to solid transformation. In a frequency sweep, measurements are made over a range of oscillation frequencies at a constant oscillation amplitude and temperature. Below the critical strain, the elastic modulus G' is often nearly independent of frequency, as would be expected from a structured or solid-like material.    The more frequency-dependent the elastic modulus is, the more fluid-like is the material. In Figure 7 (strain vs G'/G"), high strain amplitudes showed better fluid-like behaviour (G">G').
Rheological study indicates that the gel behavior of the present microsponge formulation of 5-FU in gel form is suitable for dermal drug delivery.

Conclusion:
In the microencapsulation approach, it is challenging to encapsulate 5-Fluorouracil (watersoluble drug) with a single emulsion process. This active ingredient is mainly used as an oral dosage form (tablet, capsule and injections) to treat colon cancer, oesophagal cancer, stomach cancer, pancreatic cancer, breast cancer, and cervical cancer. The main objective of the present study is to prepare a microsponge of (5-FU) by w/o/w double emulsion method.
Finally, its gel form is developed for dermal delivery of drugs used to treat actinic keratoses on the skin. The procedure is optimised by Response surface methodology. The study encompasses various preformulation studies to ensure drug compatibility with other ingredients used, optimisation of the method and scale-up. The response-surface optimisation was carried out to optimise levels of the independent factors (polymer ratio, stirring speed and surfactant concentration) to achieve the desired responses. The ANOVA results showed that builder polymers, mixing rate and amount of surfactant (tween 80) had the most substantial effects on the percentage yield, particle size, entrapment efficiency and release in 8 hr. The combination of independent variables levels polymer (600 mg), stirring speed (1198 rpm) and surfactant (2% w/v) were found to give a desirability value of 0.737 (by design generated statistical method), showing yield (63.6257%), average particle size (151.563 µm), entrapment efficiency (75.319 %) and release in 8hr (76.097 at pH 7.4, 74.460 at pH 5.5). The final formulation was a gel, so rheological characteristics were also studied, confirming its shear-thinning property facilitating dermal applicability. The highlight of this study is the scale-up approach using the composition of optimised formulation. It was done by the 'power law approach' coupled with fixed shape factors (system geometry). The properties of product microsponge obtained from higher scale are found similar to that of lower scale.
The development of scale-up technology of this type is complex under the laboratory facility as the cost of drug and set up is high for a more significant scale. However, there is still plenty of scope for up-gradation of this method to a larger scale by the pharmaceutical industry.

Declarations
Ethics approval and consent to participate: This article does not involve human participants, so it is not applicable. All the experimental procedures were performed according to the guidelines of the Jadavpur University, West Bengal, India. The article also follows the National Institute of Technology guidelines, Tiruchirappalli, Tamil Nadu, India.
Consent for publication: Not applicable. However, the authors declare that no known competing financial interest or personal relationships could have appeared to influence the work reported in this manuscript.
Availability of data and material: There are no available data and materials Competing interests: The authors declare that they have no competing interests.