1 Introduction

Liver cancer is the leading cause of death worldwide. Difficulty of treatment, recurrence, and survival of patients significant rely on the tumor size, vascular invasion, node status and hepatitis B history [1, 2]. Systemic chemotherapy is a conventional treatment where anticancer drugs are injected intravenously into the body in order to reach and destroy cancer cells [3]. Many drugs are currently used in clinic such as cis-platinum, interferon, 5-fluorouracil, combination therapy of cisplatin/interferon α-2b/doxorubicin/5-fluorouracil (PIAF), and doxorubicin (Dox). Among them, Dox is the frontline drug which has been used for cancer treatment for over 30 years. It is a highly active anthracycline antibiotic that is the most frequently used in liver cancer treatment [4, 5]. The systematic administration of Dox, however, cannot promise an effective accumulation in the tumor region, resulting in a low response rate despite the potent activity of the medication. Moreover, the circulation of Dox in the body results in tremendous harmful side effects such as cardiotoxicity, bone marrow suppression, hyperuricemia, leukopenia or skin flare at the injection site which limit the dose of the treatment. In fact, there are many physiological barriers for liver cancer chemotherapy. Transportation of drugs by blood vessels to reach liver cancer tissues must pass several compartments. First, drug molecules must distribute to blood vessel, move through the microvascular wall and are transported into the interstitium of tumors which involve both convection and diffusion [6, 7]. High interstitial pressure in the center of the tumor is the main challenging physiologic barrier for liver cancer therapy. Besides, drugs may be metabolized or degraded, consequently reducing its efficacy [8]. It was reported that systemic chemotherapy is ineffective against liver cancer cells and the survival of patients could not be prolonged using single drug or drug combination [1]. Drug delivery systems that can bypass these obstacles and directly deliver therapeutic agents to liver tumors will therefore become a great approach to achieve the therapeutic goal. To overcome these problems, Dox-loaded drug delivery systems have been developed in several forms such as micelles [9–13], liposome [14], dendrimer [15], nanoparticles [16, 17], etc. Local drug delivery systems such as polymer millirods [18, 19] and injectable polymeric depots [20–22] are another approach that can directly deliver drugs to cancer cells. Anticancer drugs are delivered directly to the problematic region to maximize the local drug concentration at the target site and minimizing the systemic toxicity. One of the local drug delivery system is an injectable polymeric depot. This system can spontaneously solidify as a spherical depot by injecting through a small needle [21]. High amount of therapeutic agents can be directly delivered at the target site and requires only small open wound [23, 24]. This system also shows ease of administration and less complication in fabrication process. Moreover, multiple injections throughout the target areas are possible [25]. Previously, our group successfully developed self-solidifying depots using tri-block copolymers of D,L-lactide (LA), ɛ-caprolactone, and PEG or PLECs to encapsulate trypan blue, a hydrophilic dye, which provided more than 90% encapsulation efficiency. In vitro release and degradation study were also carried out [21]. Recently, SN-38 was successfully loaded into these depots with high encapsulation efficiency (>98%). It was found that the highest amount of SN-38 release could be achieved using PLECs at 50 kDa with 22.5% LA and 29.7% of SN-38 loading in depots. We also found that depots protected almost 100% of encapsulated SN-38 from converting to an inactive form for more than 45 days. Additionally, SN-38-loaded depots exhibited the cytotoxic effect against human glioblastoma cell line (U-87MG) [20]. We evaluated the biocompatibility of glycofural (GF) and polymeric depots in rat brains. It was found that both GF and PLEC depots at 25 and 10 µL could be used safely with no systemic or neurologic toxicity in rat brains [26, 27]. For intratumoral drug delivery to be effective, the therapeutic agent should be able to distribute throughout the tumor and reach the therapeutic concentration. SN-38-loaded depots were injected into U-87MG xenograft tumor model in nude mice. It was found that intratumoral distribution and antitumor efficacy of SN-38 released from depots depended on dose and the number of depot injections [28]. Depots were also found to prevent the leakage of released SN-38 out of tumors and sustained higher level of SN-38 inside tumors [29]. Results from in vitro and in vivo studies definitely confirm the possibilities and potential of using these depots to deliver therapeutic agents to the brain parenchymal tissues. All of these results lead us to use this window of opportunity for other anticancer drugs. In the present study, we explored Dox-loaded polymeric depots as a local drug delivery system for liver cancer chemotherapy. To achieve this goal, effects of depot formulations on the self-solidifying process, encapsulation efficiency of sustained-release Dox (%), and in vitro release were examined. The degradation and morphology of these depots were also evaluated by Scanning Electron Microscope (SEM). In order to evaluate the efficiency of these depots, cytotoxicity studies of depots against human liver cancer cells, HepG2 cell lines, were carried out.

2 Materials and methods

2.1 Materials

Poly(ethylene glycol) (PEG) with molecular weight (Mw) of 1 kDa was purchased from Sigma-Aldrich. D,L-lactide (LA) (Sigma Aldrich) was recrystallized from ethyl acetate, vacuum-dried at room temperature and kept under argon. ε-Caprolactone (CL) (ACROS) was dried with calcium hydride (CaH2) and then distilled at reduced pressure. Toluene (Sigma Aldrich) was dried by refluxing over sodium and distilled under dry argon. Stannous(II)octoate (Sn(Oct)2) was obtained from Sigma Aldrich. Doxorubicin hydrochloride (Dox) was obtained from Guanyu bio-technology Co, LTD, Xian, China. Hydroxypropyl-beta-cyclodextrin (HP-β-CD) which has the average degree of substation of 0.5–1.3 unit per glucose unit was purchased from Trappsol, CTD Holdings Inc. Tetraglycol (Glycofurol or GF) and other reagents were purchased from Sigma Aldrich. Phosphate buffer saline (PBS) used in this study was comprised of the following components: 137 mM of NaCl, 2.7 mM of KCl, 10 mM of Na2HPO4, and 2 of mM KH2PO4.

3 Methods

3.1 Synthesis and characterization of PLECs

PLECs were synthesized following the protocol reported by previous publications [30–34]. All of the monomers must be dried to very low levels of moisture, therefore, they were dried under vacuum. Poly(ethylene glycol) (PEG) was weighted to 0.16 g and dried in the reactor at 65 °C under reduced pressure for 6 h. After that, 1.27 g of D,L-lactide (LA) and 6.8 mL of ε-caprolactone (CL) were the mixture in the reactor and dried under reduced pressure at room temperature for 6 h. Then, dehumidified argon was purged through the system to prevent oxygen from solution. Dried toluene was added into the reactor as a solvent. The mixture was dried by azeotropic distillation under reduced pressure to remove the trace of water. The reactor was immersed in an oil bath and heated at 140 °C. Stannous(II)octoate, a catalyst, was introduced into a polymerization reactor and the reaction was carried out for 48 h. The product was obtained by precipitating the toluene solution of the raw product by cold diethyl ether. The purity and molecular weight (Mw) and polydispersity index (PDI) of PLEC copolymers were characterized by Nuclear Magnetic Resonance (H1NMR) and gel permeation chromatography (GPC), respectively.

3.2 Preparation of Dox-loaded PLEC depots

PLEC with the MW of 48.41 kDa and LA content at 18.67% was dissolved in glycofurol (GF) after that various amounts of Dox were added into the mixed solution of polymer and GF. For the study of depot formation, Dox loading contents were varied from 0, 10, 15, 20 and 30% w/w and the polymer concentration was varied in the range of 10 to 45% w/v. For the in vitro release study, different formulations (Table 1) of polymeric solution (PLEC, glycofurol and Dox) were mixed by vortex mixing. Depots were prepared by injecting polymeric solution through 1 mL syringe needle into 10 mL of phosphate buffer saline (PBS, pH 7.4). Various formulations of polymeric depots that were used for release study were shown in Table 1. Percentage of sustained-release Dox (%) from PLEC depots was calculated by the following equation as reported by the previous study [35].

$$\begin{array}{l}\\ Percentage\,of\,sustained - release\,Dox\,\left( \% \right) = \\ \left( {Total\,drug - Drug\,released\,during\,depot\,formation} \right)\,\left( {mg} \right) \times \\ 100/Total\,drug\,\left( {mg} \right)\\ \end{array}$$
Table 1 Different formulations of polymeric depots for release study
Full size table

3.3 In vitro release study

Vials were placed in an incubator (Wisecube®) with the rotating speed of 90 rpm at 37 °C. At a certain time, the buffer in vials were periodically replaced with 10 mL of fresh PBS. Dox concentration was determined by UV–visible spectroscopy (Evolution 600 model, Thermo Scientific) at 480 nm. The release profiles were obtained by plotting the amount of drug release against time.

3.4 Degradation of polymeric depots

Depots were removed from PBS after 0, 14 and 30 days of incubation after that they were washed with distilled water in order to remove a residue salt on the surface of depots and were immediately freeze-dried using freeze dryer (EYELA-FDU-1200, Tokyo Rikakikai Co., Ltd) at −50 °C, 10 Pa for 4–5 h. These depots were coated with gold and place on the stub. Then, an interior and surface morphology of blank and Dox-loaded depots were examined by SEM (JSM-5410LV) at the electron beam voltage of 10 kV.

3.5 Cytotoxicity study

Cytotoxicity of both blank and Dox-loaded depots at 15% of Dox loading content and 25% of polymer concentration were evaluated using medium extraction method. Both polymer and drug were sterilized by UV light for 30 min, whereas the glycofurol was sterilized by filtration through a 0.2 μm syringe filter. In this test, the human hepatocellular carcinoma cell lines (HepG2) was purchased from JCRB Cell bank, Osaka, Japan and cultured in Dulbecco’s Modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/mL of penicillin and 100 μg/mL of streptomycin. Two days before adding the extracted medium, HepG2 cells at the cell density of 3000 cells per well were plated into a 96-well plate. Depots were formed in 10 mL of DMEM then the extracted DMEM was collected at different time periods including 3, 6, 18 and 24 h at 37 °C and 90 rpm of the incubator and added into each well (100 µL/well). After 3 days of incubation, the extracted medium was removed and cells were washed with PBS. The number of survival cells was measured by MTT assay. Briefly, MTT stock solution was prepared by dissolving 5 mg of MTT powder in 1 mL of sterilized PBS. Then, the MTT solution was added, together with a phenol red free medium, into each well and incubated at 37 °C for 4 h. After labeling cells with MTT, 85 µL of MTT solution was replaced by 100 μL of DMSO to dissolve the formazan precipitations. Finally, the mixture was mixed and measured by spectrophotometer (M965 MATE, Metertech Inc.) at the wavelength of 570 nm.

4 Results

4.1 Synthesis and characterization of PLECs

Copolymers of randomized D,L-lactide and ɛ-caprolactone at the two ends of poly(ethylene glycol) (PEG) named PLECs were synthesized and analyzed by 1H NMR spectroscopy before use. The peak of CL (–CH2–), PEG (–CH2)–) and LA (–CH–) were represented by the chemical shift at 4.0, 3.6 and 5.1 ppm, respectively. The integral intensity of NMR peaks was used to calculate the molar percentage of LA and CL as 18.67 and 81.33%, respectively. The molecular weight of PLEC was at 48.4 and 44.0 kDa as confirmed NMR and GPC, respectively. The polydispersity index of the PLEC was determined at 1.56 by GPC. 1H NMR spectrum and GPC chromatogram as shown in figure s1 and s2 in the Supplementary Data.

4.2 Preparation and characterization of Dox-loaded polymeric depots

Dox loading content (%) was investigated by fixing the depot weight in the range of 25–30 mg and polymer concentrations were varied in the range of 10–45% w/v. It should be noted that the polymer concentration at 10% w/v comprises of 10 mg of polymer in 100 µL of glycofurol. Depot formulations were categorized into four groups as shown in Table 2. The formation of polymeric depots was indicated by different signs including: (− −) depots could not be formed, (−) depots were extruded and formed as irregular shape, (+) depots were formed as oval shape and (++) depots were formed as spherical shape. In addition, the extruded polymer solution was divided into two classes; (VL) polymer solution had low viscosity and (VH) polymer solution had high viscosity. Results from Table 2 and Fig. 1 clearly show that the formation of depots mainly depended on polymer concentrations. At polymer concentration of 10% w/v, depots in all conditions could not be formed. When the polymer concentration was increased to 15% w/v, depots could be formed into an irregular shape. The higher polymer concentration between 25 to 45% w/v provided more stable solid depots as presented in Table 2. The instability of depots after loading Dox were observed at the polymer concentration of 25%. The shape of Dox-loaded depots was comparatively less stable than those of blank depots after 1 h (Fig. 2o–p).

Table 2 Properties of polymeric depots prepared by different formulations
Full size table
Fig. 1
figure 1

Photo images of blank depot formation at various polymer concentrations including a–f 10, 15, 20, 25, 30 and 35% w/v. The scale bar is 0.2 cm

Full size image
Fig. 2
figure 2

Photo images of blank polymeric depots and 20% Dox-loaded depots at different polymer concentrations (20, 25, and 30% w/v). Ti is the time when depots was injected through the needle into PBS; T0 hr is the time after depots were released from the needle; T1 hr is the time after placing depots in PBS for 1 h. The scale bar is 0.2 cm

Full size image

4.3 Percentage of sustained-release Dox (%)

Depot formation occurred after GF diffused out of polymer matrix to the surrounding aqueous environment and polymers aggregated into depots. During this process, the small amount of drugs released out of depot as showed in Fig. 2 (Ti). Therefore, the percentage of sustained-release Dox (%) was defined as the amount of Dox after the burst release and were shown in Table 3. It was found that percentages of sustained-release Dox in depots were above 90% when Dox loading content was in the range of 10 to 20%. We found that Dox loading content at 30% provided the lowest percentage of sustained-release Dox which was approximately 70% (Table 4).

Table 3 Dox loading content (% w/w) from different depot formulations
Full size table
Table 4 Actual amount of Dox loading contents and percentage of sustained-release Dox at different polymer concentrations
Full size table

4.4 In vitro release of Dox from depots

In vitro release profiles of Dox from depots were plotted as the relationship between the cumulative drug release (milligram and percentage) vs. time. The residual drugs inside depots were normalized and presented as cumulative release in percentage while the actual amount of drug release from depots at each time point was presented as cumulative drug release in milligram. Release profiles of Dox from depots with different PLEC concentrations were varied from 25 to 35% and Dox loading content was fixed at 20% which is the maximum loading capability of these depots. Dox released from depots at 25% polymer concentration showed the highest percent cumulative release as shown in Fig. 3b while the cumulative release of Dox decreased with increasing polymer concentrations. Another effect was drug loading content in depots. In this experiment, Dox loading content was varied from 10, 15 and 20% w/w where the polymer concentration was fixed at 25%. The cumulative release of Dox in milligram and percentage were shown in Fig. 3c and d. Depots with higher loading content showed faster release than those of lower loading content as shown in Fig. 3d. Various molar ratios of Dox and HP-β-CD were physically mixed with PLEC to form polymeric depots as shown in Table 1. Results showed that an increase in HP-β-CD content clearly slowed the release of Dox. For example, the burst release at 6 h of Dox from depots containing Dox:HP-β-CD at 1:1, 1:0.5, and 1:0.25 molar ratio at 25% polymer concentration was 5.1, 32.0, and 47.9%, respectively (Fig. 3e and f). The burst release of Dox was reduced by incorporating of HP-β-CD. The high burst release of Dox from PLEC depots at 25% polymer concentration (Fig. 3b) could be significantly reduced from 96.14 to 32.30% by adding HP-β-CD at 1:1 molar ratio.

Fig. 3
figure 3

Release profiles of Dox from Dox-loaded depots at different formulations. Different polymer concentrations: 25, 30 and 35% (a and b). Different Dox loading contents: 10, 15 and 20% (c and d). Different molar ratios of Dox and HP-β-CD: 1:0, 1:0.25, 1:0.5 and 1:1 (e and f) at the polymer concentration of 25%

Full size image

4.5 Degradation of depots

The degradation of polymeric depots such as the change in size and surface roughness was studied at different time periods. The surface and cross section of blank and Dox-loaded depots were also analyzed by scanning electron microscopy (SEM) with low (×150) and high magnification (×350) as showed in Fig. 4. Blank depots showed smooth surface and spherical shape within the first day (Fig. 4a1). The surface of these depots changed from smooth to rough at day 14 (Fig. 4b1 and b2). After 30 days, significant degradation inside both types of depots was observed (Fig. 4c1–5 and f1–5). Close observation by SEM images showed that the surface of both types of depots had small pores (10 µm) and the size of these pores were gradually increased as large as 50 µm at day 30. Polymer matrix inside blank depots was dense and had small shallow grave (Fig. 4a5). At day 60, both blank depots and Dox-loaded depots showed bulk erosion with hollow core which is the characteristic of autocatalytic effect generally occurring in the degradation of polyester (Fig. 5) [36, 37].

Fig. 4
figure 4

Degradation of blank depots (row a–c) and 15% Dox-loaded depots (row d–f) at 25% polymer concentration. Photo series of blank depots and Dox-loaded depots at day 0, 14, and 30 were presented in column 1 of row a–c and row d–f, respectively. SEM images of the surface of depots were in column 2–3 and the cross section of depots were in column 4–5

Full size image
Fig. 5
figure 5

Photo of blank depot (right) and Dox-loaded depot (left) at day 60

Full size image

4.6 Cytotoxicity study

Cytotoxicity study of Dox-loaded depots was evaluated in HepG2 (human liver cancer) cell by medium extraction method (MEM). Incubated in DMEM medium, depots released Dox together with GF which were collected in the culture media at various incubation times which were 3, 6, 18 and 24 h. Results showed no cytotoxic effect of extracted medium from blank depots where extracted medium from Dox-loaded depots demonstrated cytotoxic effect against HepG2 cells as shown in Fig. 6. This indicated that depot component such as GF and PLEC did not cause cytotoxic effects to HepG2 cells. A morphological observation of HepG2 cells that treated with extracted medium at various times from blank depots indicated no significant morphologic change in all conditions. On the other hand, Dox-loaded depots showed different degree of cytotoxicity. Due to Dox mechanism, cells could not adhere and grow after Dox penetrated into cytoplasm resulting in death and nonfunctional cells. Cell membranes and internal organelles were notably damaged.

Fig. 6
figure 6

Cell viability of HepG2 cells treated with 100% of extracted medium at various incubation times from blank and Dox-loaded depots (at 25% polymer concentration and 15% Dox loading)

Full size image

5 Discussion

Copolymers of randomized D,L-lactide and ɛ-caprolactone at the two ends of poly(ethylene glycol) (PEG) or PLECs were used as a material in this study due to its biocompatibility and biodegradation properties [26]. This polymer was dissolved in GF which is a biocompatible and pharmaceutical accepted solvent [26]. Both poly(ɛ-caprolactone) and poly(D,L-lactide) are polyester that were reported as biodegradable and biocompatible materials for in vivo applications [28]. Degradation products of PLA and PCL can be hydrolyzed into carboxylic acid and/or hydroxyl chain ends which may be further oxidized later [38–40]. Therefore, PLEC was selected as a material for the development of an injectable depot in this study. An improvement in degradation properties of PCL which is a highly hydrophobic semicrystalline polymer could be achieved by an introduction of PLA which is a less hydrophobic amorphous polymer. Accordingly, the degradation rate of PCL could be reduced from approximately 2 years to a couple months [41]. The ester linkages between each monomer can be hydrolyzed by water, leading to the degradation of PLECs. Typically, the degradation rate of PLEC could be controlled by adjusting the ratio of D,L-lactide to glycolide contents [21]. Molecular weight and LA content of PLECs are the important factors that effect on drug release profiles. From previous studies, depots prepared from PLECs approximately 50 kDa with 18–20% LA could delay the burst release of trypan blue, a hydrophilic dye [21] and sustained the release of SN-38 from depots over 60 days [20]. At this LA content, depots were more hydrophilic and could absorb more water which increased the ester bond cleavage and caused complete degradation of depots within 2 months. Moreover, SN-38-loaded PLEC depots that have molecular weight and %LA in this range also successfully injected into U-87MG xenografted tumor model in nude mice which showed antitumor efficacy and biocompatibility [42].

When the mixture of PLEC, GF, and Dox were injected into PBS, the solidification process began as a result of water influx while GF diffusing out of depots (Fig. 1e and f). Since this polymer was insoluble in PBS, it aggregated themselves to form semi-solid depots. This solidifying process could be observed as depots gradually turned from clear to opaque as showed in Fig. 2. Hydrophobicity and affinity between solvent (GF) and non-solvent (PBS) affected solidification rate of PLEC depots [21]. For low polymer concentration, there was less amount of hydrophobic affinity of solvent to non-solvent resulting in slow phase inversion. Depots could be formed as spherical shape at the polymer concentration of 20% w/v, however depots at this concentration could not maintain their shape after 1 h as shown in Fig. 2m and n. This might be the result of low density of polymer solution leading to loosen polymer matrix. Moreover, the solidification process of depots made from low viscosity polymer might be affected by the pressure of surrounding water that made these depots to be deformed to oval shape. We found that the shape of Dox-loaded depots was comparatively less stable than those blank depots after 1 h because GF quickly diffused of blank depots providing the rapid polymer aggregation. On the other hand, the introduction of Dox increased the hydrophilicity of depots and reduced the rate of solvent exchange during depot formation resulting in slower phase inversion. Moreover, the shape of Dox-loaded depots at 25% polymer concentration (Fig. 2p) is similar to trypan blue-loaded PLEC depots which was reported by previous work. It should be noted that trypan blue is a hydrophilic dye.

These depots exhibited very high percentage of sustained-release Dox (%) that was in agreement with our previous study [20]. However, we found that Dox loading content at 30% provided lower percentage of sustained-release Dox which can be explained by the hydrophobic interactions between drugs and hydrophobic segments of polymers. In this case, Dox is inherently hydrophobic, therefore, it was less likely to be entrapped in the hydrophobic segment of polymer matrix [12, 43]. Therefore, an actual Dox loading content in all polymer concentrations was approximately 20% compared to the theoretical loading at 30%. This indicated that the maximum drug loading content of depots was approximately 20%. Thus, 20% of Dox loading content was selected for the release study. It should be noted that drug loading content at 20% is comparatively higher than that of other drug delivery systems such as micelles [44–46], nanoparticles [47], liposomes and dendrimers [48].

For release study, it is important to consider release profiles in both cumulative percentage and milligram in order to compare factors affecting drug release rate. Therefore, various parameters affecting on the release profiles were evaluated. In this study, different depot formulations including polymer concentration and drug loading content were studied. Results showed that release profiles of depots prepared from all polymer concentrations showed burst release at the beginning which was commonly found in Dox-loaded drug delivery systems [49–51]. It was found that cumulative release of dox decreased when polymer concentration was increased due to a denser polymer matrix that caused Dox to be encapsulated inside polymer matrix rather than releasing out of depots. This can be explained by Fick’s First Law where denser polymer matrix lowers the diffusion coefficient. Accordingly, less drug molecules diffused out of depots.

Drug loading content in depot is the other factor that affect the release profile. The higher drug loading content, the faster release profile was observed. These results were also reported by Rom E. and Rosenberg W. D. R. [23, 52] where high loading content of highly hydrophilic agents such as protein and nicotine provided high release rate. Drug loading content also has the effect on the distribution of drug molecules inside depots as can be described by percolation theory [53]. Percolation threshold is used to describe the release of drugs from diffusion-controlled drug delivery systems consisting of drug and polymer matrix. Incomplete release occurs at drug loading content below this threshold where drugs molecules are isolated and surrounded by hydrophobic polymer matrix. On the other hand, the complete release was found when using drug loading content above this threshold. In this case, drugs formed interconnecting channels throughout matrix which is the path for drugs to release out of samples. For Dox-loaded depots, the percolation threshold was between 15–20% w/w. This is supported by SEM images where 15% w/w of Dox-loaded depots showed interconnecting channels after 2 days of drug release (Fig. 4d3–f3). An ideal drug release profile should have a burst release at an early period to produce drug concentration within therapeutic range followed by prolonged release of drugs in the therapeutic range [51]. Nevertheless, an uncontrollable burst release will cause side effects to patients.

In this experiment, the burst release of Dox from depots was controlled by introducing excipient molecules into Dox-loaded depots. Cyclodextrin (CD) is a cyclic oligosaccharides molecule that have a well-defined shape of truncated cone with the cavity used for chemical modification or conjugation to form inclusion complex with hydrophobic compounds [54, 55] There are three common types of CD including α, β, and γ-cyclodextrin. The β one especially hydroxypropyl-β-cyclodextrin (HP-β-CD) has good cavity size and architecture and also has relatively low cost [56–58]. Accordingly, HP-β-CD was selected to study its effects on the burst release of Dox from depots. Dox molecule could stay and fit into the cavity of the cyclic oligosaccharide of HP-β-CD [59] while the rest of Dox molecules was outside which can be defined as free Dox. Nevertheless, the inclusion complex between Dox and HP-β-CD is a reversible process and controlled by equilibrium constant, Kc [55]. Therefore, there were two forms of Dox including free Dox and Dox at the cavity of HP-β-CD. A number of hydrogen bonds between amino, hydroxyl or carbonyl groups of Dox and hydroxyl groups locating on the outer surface of HP-β-CD can be formed [60]. Without HP-β-CD, the hydrogen bond is only found between Dox and the ester bond of the polymers. Accordingly, hydrogen bonding is indeed one of a significant factor that entrap Dox molecules inside the polymeric matrix of depots. Moreover, HP-β-CD lowered the burst release due to the physical crosslinking in polymer matrix from hydrogen bonding between hydrogen and hydroxyl groups of HP-β-CD and an ester bond of polymers. [61]. It can be implied that this physical crosslinking is the other factor that entraps Dox within the polymer matrix, thus drugs cannot diffuse out of polymeric depots.

Degradation of depots was started by water absorption and hydrolysis of ester linkage in polyester. Small amount of GF immediately diffused out during depot formation causing very fine and homogenous pore size which was subsequently changed to big pores after 30 days of degradation (pore size = 30 µm) (Fig. 4c5). Instantly after injection, the surface of Dox-loaded depots was slightly rougher than that of blank depots (Fig. 4d1). This was due to the fact that Dox-loaded depots were more hydrophilic than blank depots because of Dox encapsulation. Therefore, Dox-loaded depots showed faster degradation rate which obviously increased the amount of pores and caused big pores with the size of 20 µm inside depots as shown in SEM images Fig. 4d3–f3. This produced interconnecting channels throughout depots which were the route for Dox to release out of depots.

In order to study cytoxicity of these depots, it should not be tested directly on HepG2 cell due to its size and weight which may cause a physical trauma to cells leading to cell dead [62]. Moreover, this system could not be dissolved in culture medium. Therefore, cytotoxicity study of Dox-loaded depots was evaluated in HepG2 (human liver cancer) cell by medium extraction method (MEM). This method used an extracted medium of depots to treat cells instead of direct incubation of depots with cells. As expected, no cytotoxic effect of extracted medium from blank depots where extracted medium from Dox-loaded depots demonstrated cytotoxic effect against HepG2 cells. Due to most human cancer cells have doubling time in the range of 12 to 48 h [63, 64], the concentration of Dox released from depots as shown in the release study section will reach the cytotoxic level quickly resulting in a certain prevention of tumor accretion or metastasis.

We expect that this study will provide several important insights for those who has developed an injectable drug delivery system for cancer treatment. However, additional studies are required to demonstrate the in vivo behaviors of these depots such as biodistribution, antitumor efficacy, etc.

6 Conclusion

PLEC was synthesized and used as a material for Dox-loaded depots. These depots were prepared with high percentage of sustained-release Dox (%). Results from this study reveal that polymer concentration and drug loading content have effects not only on solidification process of depots but also the release profile and degradation. Moreover, Dox-loaded depots showed the cytotoxic effect against HepG2 cell line over blank depots at various incubation time. Therefore, these results indicated that Dox-loaded depots could be promising drug carriers for local delivery of Dox to liver cancer cells and provides an important information for the study of these depots in vivo.