2018 Volume 66 Issue 4 Pages 382-390
Nanoparticles prepared using human serum albumin (HSA) have emerged as versatile carriers for improving the pharmacokinetic profile of drugs. The desolvation of HSA using ethanol followed by stabilization through crosslinking with glutaraldehyde is a common technique for preparing HSA nanoparticles, but our knowledge concerning the characteristics (or functions) of HSA nanoparticles and their efficiency when loaded with drugs is limited. To address this issue in more detail, we prepared anthracycline-loaded HSA nanoparticles. Doxorubicin-loaded HSA nanoparticles with a size similar to doxorubicin-unloaded particles could be prepared by desolvating at a higher pH (8–9), and the size (100–150 nm) was optimum for delivery to tumor tissues. Using this procedure, HSA nanoparticles were loaded with other anthracycline derivatives, and all showed cytotoxicity in cancer cells. However, the efficiency of drug loading and dissolution rate were different among them possibly due to the differences in the type of association of the drugs on nanoparticles (doxorubicin and daunorubicin; covalently bound to nanoparticles, pirarubicin; both covalently bound to and adsorbed on nanoparticles, aclarubicin; adsorbed on nanoparticles). Since the formulation of such drug-loaded HSA nanoparticles should be modified for efficient delivery to tumors, the findings reported herein provide the useful information for optimizing the formulation and the production process for the HSA nanoparticles using a desolvation technique.
Human serum albumin (HSA), a major protein component of blood plasma,1,2) plays an important role in the regulation of colloidal osmotic pressure, the antioxidant capacity of human plasma, and the transport of numerous endogenous compounds including fatty acids, hormones, toxic metabolites (e.g. bilirubin), bile acids, amino acids, and metals.1,2) HSA has recently emerged as a versatile carrier for improving the pharmacokinetic profile of drugs or for drug targeting,3,4) since it shows a long half-life (about 19 d) and accumulation in tumor tissue. Furthermore, biodegradability, and a lack of toxicity and immunogenicity make HSA an ideal candidate for drug delivery.
HSA has been applied to extending the half-life of drugs by means of genetic or chemical linking between HSA and drugs. Albiglutide was first marketed albumin fused peptide drug (Eperzan®).5) This molecule consists of two molecules of a modified human glucagon-like peptide 1 (GLP-1) that is genetically fused to HSA and shows longer half-life allowing once-weekly dosing for the treatment of type 2 diabetes. HSA can be also used for delivering drugs to targeted tissues. Macromolecules such as HSA and nanoparticles with a size below 200 nm are able to accumulate in tumors due to the enhanced permeation and retention (EPR) effect.6–8) Such a passive targeting to tumors is suggested in cases of an albumin-binding doxorubicin prodrug (Aldoxorubicin)9,10) and albumin-bound paclitaxel nanoparticles (nab®-paclitaxel; Abraxane®).11,12) In addition, the active transport of HSA-bound drugs by an albumin receptor such as the 60 kDa glycoprotein (gp60) located on the endothelial cell surface is also thought to contribute to the accumulation of these drugs in tumors.12)
HSA nanoparticles prepared by a desolvation technique have been studied as another type of carrier for delivering anti-cancer drugs, since these nanoparticles have sizes suitable for EPR and are readily prepared.13–15) Generally, ethanol (used as a desolvating agent) is added to an aqueous solution of HSA. During this process, nanoparticles are produced due to the limited solubility of HSA in ethanol. The nanoparticles are stabilized by treatment with glutaraldehyde, crosslinking agent, where lysine and/or arginine residues of HSA react with aldehyde-group of glutaraldehyde. Stabilized nanoparticles can be obtained through purification and freeze-dry processes. Langer et al. systematically optimized the preparation process for HSA nanoparticles that were not loaded with drugs (HSA NPs) with regards to the desolvating agent used, HSA concentration and the pH of the aqueous solution, the amount of glutaraldehyde and the reaction time used.13–15) Several researchers have used and modified this procedure to prepare albumin nanoparticles loaded with, not only anticancer drugs (e.g. doxorubicin or paclitaxel)16–18) but also other types of drugs (e.g. ganciclovir, atorvastatin, gabapentin, metformin, noscapine or piceatannol).19–24) In spite of these studies, process parameters that affect the characteristics of drug-loaded HSA nanoparticles (drug-HSA NPs) or the applicability of the procedures for a wide variety of drugs are still unclear. Furthermore, questions regarding how such nanoparticles could function effectively as carriers for delivery of drugs by EPR effect also arise.
In this study, we first investigated the effects of process parameters on the particle characteristics (particle size, polydispersity, zeta-potential) of doxorubicin (DXR)-loaded HSA nanoparticles (DXR-HSA NPs) and on production efficiency (drug loading or production yield). Furthermore, we confirmed the applicability of this procedure for preparing HSA nanoparticles loaded with other structurally different anthracycline derivatives, namely, daunorubicin (DNR), pirarubicin (THP) and aclarubicin (ACR) (DNR-, THP- and ACR-HSA NPs) (Fig. 1). Finally, we investigated the functions of these drug-HSA NPs in in vitro experiments (dissolution, cell viability and intracellular uptake studies), and also address the issue of whether the produced nanoparticles could function as carriers for delivery of anthracyclines via the EPR effect.
DXR, DNR, THP and ACR are docorubicin, daunorubicin, pirarubicin and aclarubicin, respectively.
Recombinant human serum albumin (rHSA) was donated by Nipro Corporation (Shiga, Japan) and defatted using a charcoal treatment as described by Chen.25) After dialysis against distilled water, the protein was freeze-dried and stored at −20°C until used. DXR (as hydrochloride) and ACR were purchased from Toronto Research Chemicals Inc. (Toronto, Ontario, Canada). THP was obtained from LKT Laboratories, Inc. (St. Paul, MN, U.S.A.). DNR (as hydrochloride), proteinase K, RPMI-1640 solvents (methanol, ethanol, 2-propanol and acetonitrile), fluoresceinisothiocyanate isomer-I (FITC-I) and the cell counting kit-8 were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Glutaraldehyde, D-mannitol, tris (hydroxymethyl) aminomethane and sodium dodecyl sulfate were obtained from Nacalai Tesque, Inc. (Kyoto, Japan). Human plasma was obtained from Interstate Blood Bank, Inc. (Memphis, TN, U.S.A.). All other chemicals were obtained from commercial sources were of the highest grade.
Preparation of NanoparticlesNanoparticles were prepared using a previously described desolvation technique with minor modification.13–15) Briefly, HSA was dissolved in 1 mL of ultrapure water at concentration of 10–80 mg/mL, and the pH was adjusted to 4–9 using 0.5 M HCl and NaOH. Particles were generated by the addition of 3 mL of ethanol, methanol or 2-propanol as desovating agents to HSA solution with a syringe pump (Flow rate: 1 mL/min) under magnetic stirring (1500 rpm) at room temperature. The resulting particles were stabilized by addition of 8% glutaraldehyde as a crosslinking agent (1.175 µL/mg HSA), followed by stirring for 24 h (500 rpm). The suspension was centrifuged (20000×g, 20 min) at 24°C, and the collected nanoparticles were re-dispersed in 4 mL of ultrapure water. This procedure (i.e., centrifugation and re-dispersion) was repeated 3 times, and after the third centrifugation, sufficient ultrapure water was added to produce an HSA concentration of 10 mg/mL. In addition, D-mannitol, an aggregation protectant, was dissolved in this sample at concentration of 3% (w/v) and the sample was then freeze-dried. Drugs-HSA NPs were also prepared under the following conditions, (1) drugs (0.5 mg/mL); DXR, DNR, THP and ACR. (2) HSA concentration; 20 mg/mL, (3) pH; 8.5, (4) desolvating agent; ethanol.
Entrapment Efficacy, Loading Capacity and Production YieldThe supernatant after centrifugation during the washing of the nanoparticles was collected and the amount of drug in the solution was determined by HPLC. The amount of drug entrapped in nanoparticles was then determined by subtracting the amount in the supernatant (free drug) that was not entrapped in the nanoparticles. Entrapment efficiency, loading capacity and production yield were calculated as follows:
 | (1) |
 | (2) |
 | (3) |
The particle size, polydispersity index (PDI) and zeta potential of the prepared nanoparticles were determined by light scattering (ELSZ1000 series, Otsuka Electronics Co., Ltd., Osaka, Japan). The samples were measured in phosphate buffered saline (PBS) (pH 7.4) at 25°C. Measurements were conducted in triplicate (n=3). Transmission electron microscopy (TEM) images were obtained on a JEOL model JEM-1400Plus microscope (JEOL, Ltd., Tokyo, Japan). Twenty microliters of a suspension of DXR-HSA NPs was dropped onto the carbon side of formvar/carbon on a 400 mesh copper grid (Veco Specimen Grids, Electron Microscopy Sciences, Inc., Hatfield, PA, U.S.A.) and allowed to stand for 5 min. After drying the grid by absorbing the excess solution with filter paper, one drop of a saturated aqueous solution of uranium acetate (4–5% (w/v)) was added dropwise on the surface. The excess solution was discarded with the filter paper and the naturally dried grid was examined by TEM.
Dissolution TestA 10 mg portion of nanoparticles were horizontally shaken at 37°C and 150 rpm in a capped bottle containing a 10 mL of dissolution medium (0.01 M Tris–HCl buffer, pH 7.4 containing 0.5% sodium dodecyl sulfate (SDS) with 50 µg/mL proteinase K or human plasma). At specific time intervals up to a maximum of 100 h, 0.4 mL samples were withdrawn from the bottle, subjects to ultracentrifugation at 100000×g, 25°C for 10 min (Optima L-80 XP Ultracentrifuge, Beckman Coulter Inc., CA, U.S.A.). An equal volume of 2 M HCl was added to the supernatant and the resulting solution incubated for 90 min at 50°C. The incubated sample was deproteinized by treatment with acetonitrile and centrifuged at 10000×g, 25°C for 5 min. The supernatant was diluted with acetate buffer. Drug concentrations in the samples were determined by HPLC. Three dissolution tests were conducted for each powdered sample (n=3).
HPLC ConditionsThe HPLC system used in this study consisted of a Hitachi model D-2000 Elite HPLC system (Hitachi Co., Tokyo, Japan) and Jasco Intelligent Fluorescence Detector, FP-2025 Plus (Jasco Corporation, Tokyo, Japan). CAPCELL PAK C18 MGII (5 µm particle size, 4.6 mm i.d.×250 mm, Shiseido Co., Ltd., Tokyo, Japan) was used as the stationary phase and was maintained at 40°C. Two solvents, solvent A (acetonitrile) and solvent B (0.1 M acetate buffer, pH 5) were used as the mobile phases. The following linear gradient elution of the solvents was programmed for the quantitation of DXR, DNR and THP: 0–5 min (10–60% A), 5–11 min (60% A), 11–13 min (60–10% A), 13–15 min (10% A), and for ACR: 0–5 min (10–70% A), 5–11 min (70% A), 11–13 min (70–10% A), 13–15 min (10% A). The flow rate of the mobile phase was maintained constant at 1.0 mL/min. Excitation and emission wavelengths, 490 and 590 nm were used for the detection of DXR, DNR and THP, and 440 nm and 584 nm for ACR.
Covalent Binding Assay between Drug and HSAIn order to confirm whether each drug covalently binds to HSA via glutaraldehyde, 1 mL of ethanol was added to 1 mL of an HSA solution at pH 8.5 containing 0.5 mg of each drug, and glutaraldehyde was then added. After a 2 h reaction, each solution was deproteinized with acetonitrile and centrifuged at 10000×g, 25°C for 5 min. The concentration of the drug in the supernatant was measured by HPLC. Reactivity was calculated as follows:
 | (4) |
The cytotoxicity of the drug-HSA NPs was evaluated using a human breast cancer cell line, MCF-7 and a human hepatocellular carcinoma cell line, HepG2. Cells were maintained in RPMI-1640 (Wako Pure Chemical Industries, Ltd.) supplemented with 10% fetal bovine serum (FBS) containing 1% penicillin–streptomycin using a 5% CO2/95% relative humidity incubator at 37°C. To evaluate the cytotoxic effects of the free drug and drug-HSA NPs, cells were seeded on a 96-well plate at a concentration of 1×104 cells per well and pre-incubated for 24 h. After the incubation, pre-determined amounts of drug (DXR, DNR, THP, and ACR) or drug-HSA NPs (DXR-HSA NPs, DNR-HSA NPs, THP-HSA NPs, and ACR-HSA NPs) were added to give final drug concentrations of 10–10000 nM. In vitro cytotoxicity was determined using a cell counting kit-8 after 72 h incubation and the IC50 was calculated.
Intracellular UptakeIn order to prepare FITC-DXR-HSA NPs, HSA was labeled with FITC-I at pH 8.5. For cellular uptake examination, MCF-7 cells were seeded in 9.6 cm2 poly-lysine coated glass bottom culture dishes (Matsunami Glass Ind., Ltd.) at a cell density of 2×105 cells in 2 mL of RPMI 1640 Medium, with no Phenol Red. After 24 h, cells were incubated with DXR, FITC-DXR-HSA NPs at a final DXR concentration of 8 µg/mL at 37°C. The cells were observed by confocal laser fluorescence microscopy, with an excitation wavelength at 488 nm and an emission wavelength at 500–530 nm and 570–640 nm. Furthermore, to confirm the nuclear accumulation of DXR, DXR fluorescence in nucleus of randomly selected 10 cells were monitored. Relative fluorescence intensity to that after 4 h was calculated each time.
Statistical Data AnalysisAll data are expressed as the mean±standard deviation (S.D.) Statistical analyses for multiple comparisons in the study were determined by two-way ANOVA followed by the Bonferroni analysis. A probability value of p<0.05 was considered to be significant.
In a previous study, Langer and colleagues reported on the preparation of DXR-HSA NPs with small and uniform size using a desolvation method under the following conditions (standard condition), HSA concentration; 20 mg/mL, pH; 8.2, DXR concentration; 0.5 mg/mL, glutaraldehyde concentration; 1.175 µL/mg HSA and desolvating agent; ethanol.16) In initial experiments, we prepared DXR-HSA NPs under almost the same condition except for the pH (a pH of 8.5 was used in our experiments). Entrapment efficacy, loading capacity and production yield under these conditions were 95.7, 3.9 and 57.6%, respectively (Table 1). The prepared nanoparticles were characterized by light scattering and the findings indicated that the nanoparticles possesses a negative charge and had a uniform size (average size; 108 nm, PDI; 0.08) (Fig. 2A). The size of the DXR-HSA NPs was the same as that of HSA NPs (Fig. 2A, a and Table 1). The size and its distribution of DXR-HSA NPs was confirmed by TEM analysis (Fig. 2B).
| Nanoparticles | Characteristics of nanoparticles | Production efficiency | ||||
|---|---|---|---|---|---|---|
| Mean size (nm) | PDI | Zeta potential (mV) | Entrapment efficacy (%) | Loading capacity (%) | Production yield (%) | |
| HSA NPs | 105±5 | 0.06±0.03 | −40±8 | — | — | 42.5±5.2 |
| HSA-DXR NPs | 108±9 | 0.08±0.01 | −35±6 | 95.7±0.1 | 3.9±0.6 | 57.6±5.2 |
All values represent the mean±S.D. (n=3).
(A) is particle size histogram of HSA NPs which does not involve DXR. TEM images of DXR-HSA NPs (B) observed at 6000 times magnification, scale bars=1 µm (left figure) and at 50000 times magnification, scale bars=100 nm (right figure) were shown.
The pH of the HSA or HSA-DXR solutions in the desolvation process affected the appearance of the particles after desolvation by ethanol (Fig. 3). Particle formation was observed at pH 5–9 but not at pH 4. At pH 6, sedimentation was observed nearly immediately after desolvation in the HSA-DXR solution, but not in the HSA solution. After the crosslinking process, nanoparticles with size ranging from 100 to 200 nm were produced at pH 7–9 (Table 2). The size of the nanoparticles was decreased with increasing pH from 7 to 8 or 8.5, and was constant in the pH range 8.5–9. While an increase in entrapment efficacy was observed at pH 9, the production yield was decreased at this pH.
| Nanoparticles | pH | Characteristics of nanoparticles | Production efficiency | ||||
|---|---|---|---|---|---|---|---|
| Mean size (nm) | PDI | Zeta potential (mV) | Entrapment efficacy (%) | Loading capacity (%) | Production yield (%) | ||
| HSA NPs | 5 | 5077±4097 | 0.34±0.20 | −20±9 | — | — | 71.8±16.6 |
| 6 | 222±25 | 0.07±0.04 | −28±5 | — | — | 72.3±10.3 | |
| 7 | 179±27 | 0.04±0.03 | −42±16 | — | — | 71.5±3.0 | |
| 8 | 114±13 | 0.05±0.01 | −38±10 | — | — | 42.0±8.0 | |
| 8.5 | 105±5 | 0.06±0.03 | −40±8 | — | — | 42.5±5.2 | |
| 9 | 109±11 | 0.06±0.01 | −41±16 | — | — | 22.8±4.9 | |
| HSA-DXR NPs | 5 | 5532±3403 | 0.67±0.35 | −21±16 | 16.6±2.9 | 0.6±0.1 | 64.4±9.5 |
| 6 | — | — | — | — | — | — | |
| 7 | 212±19 | 0.06±0.02 | −32±6 | 95.5±1.1 | 2.9±0.1 | 77.7±3.2 | |
| 8 | 141±14 | 0.05±0.01 | −36±6 | 95.6±0.6 | 3.1±0.3 | 72.4±8.4 | |
| 8.5 | 108±9 | 0.08±0.01 | −35±6 | 95.7±0.1 | 3.9±0.6 | 57.6±5.2 | |
| 9 | 111±20 | 0.09±0.01 | −38±9 | 95.2±0.3 | 4.8±0.6 | 46.5±8.7 | |
Conditions except for pH were same as standard condition. Data of pH 8.5 are from Table 1. All values represent the mean±S.D. (n=3).
We prepared DXR-HSA NPs and HSA NPs using different concentrations of HSA (Table 3). The size of nanoparticles tended to be increased both in DXR-HSA NPs and HSA NPs, with increasing HSA concentration while the size of all particles were uniform. Production yields of HSA-DXR NPs were higher than HSA-NPs especially in the lower concentration of HSA.
| Nanoparticles | HSA Concentration (mg/mL) | Characteristics of nanoparticles | Production efficiency | ||||
|---|---|---|---|---|---|---|---|
| Mean size (nm) | PDI | Zeta potential (mV) | Entrapment efficacy (%) | Loading capacity (%) | Production yield (%) | ||
| HSA NPs | 10 | 104±22 | 0.13±0.01 | −23±7 | — | — | 16.7±10.1 |
| 20 | 105±5 | 0.06±0.03 | −40±8 | — | — | 42.5±5.2 | |
| 40 | 113±4 | 0.08±0.02 | −43±9 | — | — | 52.4±15.6 | |
| 80 | 119±3 | 0.07±0.01 | −49±13 | — | — | 66.9±4.2 | |
| HSA-DXR NPs | 10 | 103±7 | 0.09±0.02 | −28±4 | 73.4±1.1 | 6.7±1.1 | 53.3±9.2 |
| 20 | 108±9 | 0.08±0.01 | −35±6 | 95.7±0.1 | 3.9±0.6 | 57.6±5.2 | |
| 40 | 124±3 | 0.03±0.00 | −39±4 | 96.0±0.4 | 1.6±0.1 | 69.6±4.3 | |
| 80 | 113±5 | 0.06±0.01 | −39±7 | 95.6±0.5 | 0.6±0.1 | 102.2±15.2 | |
Conditions except for HSA concentration were same as standard condition. Data of HSA concentration 20 mg/mL are from Table 1. All values represent the mean±S.D. (n=3).
Methanol and 2-propanol were both tested for use as solvents for desolvation instead of ethanol (Table 4). The use of methanol produced a very low amount of particles, a significant reduction in production yield. Size of DXR-HSA NPs were similar produced when ethanol and 2-propanol were used.
| Nanoparticles | Desolvating agent | Characteristics of nanoparticles | Production efficiency | ||||
|---|---|---|---|---|---|---|---|
| Mean size (nm) | PDI | Zeta potential (mV) | Entrapment efficacy (%) | Loading capacity (%) | Production yield (%) | ||
| HSA NPs | Methanol | — | — | — | — | — | — |
| Ethanol | 105±5 | 0.06±0.03 | −40±8 | — | — | 42.5±5.2 | |
| 2-Propanol | 90±4 | 0.07±0.00 | −39±8 | — | — | 38.0±3.1 | |
| HSA-DXR NPs | Methanol | 153±40 | 0.16±0.09 | −24±6 | — | — | 5.7±7.6 |
| Ethanol | 108±9 | 0.08±0.01 | −35±6 | 95.7±0.1 | 3.9±0.6 | 57.6±5.2 | |
| 2-Propanol | 110±8 | 0.05±0.01 | −42±8 | 95.1±1.7 | 3.0±0.1 | 74.0±3.7 | |
Conditions except for desolvating agents were same as standard condition. Data using ethanol are from Table 1. All values represent the mean±S.D. (n=3).
Nanoparticles prepared with DNR, THP and ACR, which possess different structures, were compared to particles prepared with DXR (Table 5). Entrapment efficacy and loading capacity were higher in the case of the DNR-HSA NPs than in DXR-HSA NPs. These parameters for the THP- and ACR-HSA NPs were lower than in the DXR-HSA NPs and the values for the ACR-HSA NPs were lowest.
| Nanoparticles | Drugs | Characteristics of nanoparticles | Production efficiency | ||||
|---|---|---|---|---|---|---|---|
| Mean size (nm) | PDI | Zeta potential (mV) | Entrapment efficacy (%) | Loading capacity (%) | Production yield (%) | ||
| HSA NPs | — | 105±5 | 0.06±0.03 | −40±8 | — | — | 42.5±5.2 |
| Drug HSA-NPs | DXR | 108±9 | 0.08±0.01 | −35±6 | 95.7±0.1 | 3.9±0.6 | 57.6±5.2 |
| DNR | 109±23 | 0.06±0.01 | −33±10 | 97.1±0.1 | 5.9±0.7 | 46.3±5.8 | |
| THP | 121±2 | 0.05±0.02 | −43±3 | 87.8±1.5 | 2.6±0.1 | 82.6±2.2 | |
| ACR | 119±3 | 0.04±0.01 | −45±2 | 29.7±1.6 | 1.1±0.2 | 68.3±10.3 | |
Standard condition was used for preparations. Data of HSA NPs and DXR-HSA NPs are from Table 1. All values represent the mean±S.D. (n=3).
To speculate on the present state of drug in HSA NPs, we investigated whether the drug is covalently bound to HSA through glutaraldehyde. In an absence of glutaraldehyde, no binding to HSA was detected for any drug (data not shown). The covalent binding of drugs to has, expressed as reactivity, appeared to be in the order DXR, DNR and THP (order of reaction: DXR=DNR>THP), whereas no reaction was observed in the case of ACR (Fig. 4).
Values are expressed as the means±S.D. (n=3). * p<0.01.
As one of the functions of nanoparticles, the dissolution of drugs from prepared anthracyclines-HSA NPs were characterized in several mediums. In dissolution tests using a buffer with a protease, proteinase K (Fig. 5A), only an initial fast dissolution was observed for the ACR-HSA NPs, whereas sustained dissolution was also observed for the DXR-, DNR- and THP-HSA NPs. Dissolution rates at 0.5 h for the DXR-, DNR-, THP- and DNR-HSA NPs were 14.7, 16.4, 52.0 and 94.3%, respectively. Thus, the initial rapid dissolution from ACR- and THP-HSA NPs were more marked than those from DXR- and DNR-HSA NPs, and that from ACR-HSA NPs was the fastest. In the plasma (Fig. 5B), only an initial rapid dissolution was observed for all of the nanoparticles, and the dissolution rates at 0.5 h for DXR-, DNR-, THP- and DNR-HSA NPs were 4.9, 5.8, 34.1 and 57.1%, respectively. The patterns for initial dissolution were similar to those in the buffer with proteinase K.
Values are expressed as the means±S.D. (n=3).
As another function of nanoparticles, the cytotoxicity of nanoparticles was evaluated in MCF-7 and HepG2 (Table 6). The IC50 values using anthracycline-HSA NPs were increased when compared to those using free anthracyclines in both cell lines. No cytotoxic activity was observed in the case of the HSA NPs (data not shown).
| Cell lines | DXR | DNR | THP | ACR | ||||
|---|---|---|---|---|---|---|---|---|
| Free | NPs | Free | NPs | Free | NPs | Free | NPs | |
| MCF-7 | 537±74 | 2197±385 | 624±238 | 2825±290 | 334±119 | 845±151 | 110±96 | 454±142 |
| HepG2 | 325±130 | 1066±331 | 112±33 | 989±182 | 169±44 | 289±27 | 47±14 | 177±27 |
All values represent the mean±S.D. (n=3).
The intracellular uptake of free DXR and DXR-HSA NPs in MCF-7 was also examined using fluorescent confocal laser microscopy (Fig. 6). Signals corresponding to DXR was detected inside the cell for free DXR and DXR-HSA NPs (Fig. 6A). For the free DXR, DXR accumulated in the cell nucleus. For DXR-HSA NPs, both signals of DXR and nanoparticles labelled by FITC were co-localized in non-nucleus regions. Furthermore, the DXR fluorescence intensity was significantly increased at cell nucleus (Fig. 6B) and decreased at nanoparticles (Fig. 6C), suggesting that the DXR was released from the nanoparticles and accumulated in the nucleus.
After incubating 4 h, cells were visualized by using confocal laser scanning microscopy under 5% CO2 at 37°C. Images were standardized to the same setting. The red and the green indicates fluorescence of DXR and FITC labeled DXR-HSA NPs, respectively. Scale bars are all 50 µm. Values are expressed as the means±S.D. (n=3). * p<0.01, ** p<0.001.
The desolvation of HSA using ethanol followed by stabilization using glutaraldehyde is a simple technique for preparing HSA nanoparticles, while the characteristics of nanoparticles (e.g. size or size distribution) are affected several process parameters.13–15) In addition, the present findings showed that, when drugs are loaded in HSA nanoparticles, the preparation is not simple, since the production efficiency and characteristics of the nanoparticles may be dependent on the drug structures. In using such nanoparticles for the treatment of cancer, not only size but drug loading and functional aspects are important. In this study, especially the latter was shown to be different from one drug to another, possibly due to the differences in the drug structures and therefore the present state of drugs in nanoparticles.
Based on the conditions used by Langer and colleagues (standard condition),16) DXR-HSA NPs with small size (108 nm) optimum for accumulation to tumors through the EPR effect could be prepared (Table 1 and Fig. 2). Our data also indicated that size of nanoparticles decreases with increasing pH of the HSA and DXR-HSA solutions that are used in the desolvation step (Table 2). The repulsion of anionic HSA at pH 7–9 which is higher than the isoelectric point of HSA (pI 5.3) may lead to the formation of smaller nanoparticles. Thus, pH is an important factor for controlling the size of nanoparticles. In the presence of a cationic drug, DXR, the solubility of the HSA molecules or the repulsion of nanoparticles at around pH 6 may further decrease by charge neutralization, resulting in aggregation and sedimentation, when desolvation with ethanol was used (Fig. 3). Although a higher pH is recommended for producing smaller and stable DXR-HSA NPs with a higher entrapment efficacy, the reduction of the production yield possibly due to the decreased formation of nanoparticles should be considered in optimizing formulations in this condition.
A decrease in the nanoparticle size and production yield were observed at lower HSA concentrations (Table 3), which can be attributed to a reduction in contact between HSA molecules. Interestingly, DXR increased the production yield without a significant change in particle size. This effect is not specific for DXR because similar effects were also observed for the other anthracyclines regardless of their structures. Therefore, such drugs may contribute to the formation of nanoparticles at the desolvation step, for example, through the decrease in the solubility of an HSA-drug complex. Production yields for hydrophobic drugs such as THP and ACR might be higher as compared with those for more hydrophilic drugs such as DXR and DNR. Thus, lower HSA concentrations would be more suitable for producing smaller particles and the addition of drugs in the production system may improve production yields.
Methanol was reported to be a desolvating agent and to produce smaller nanoparticles.15) However, methanol showed a slower dehydration effect due to its higher dielectric constant, resulting in the need for a higher volume for desolvation. Desolvation with a smaller volume as used in the present study leads to a decrease in the amount of nanoparticles produced at the desolvation step and thereby the production yields would be expected to be decreased (Table 4). The use of ethanol and isopropanol which show lower dielectric constants than methanol lead to the producion of relatively small DXR-HSA NPs with higher production yields, even when their volumes are small.
Using standard conditions for preparing the DXR-HSA NPs, other anthracyclines-HSA NPs with a size optimum for the EPR effect also could be produced. However, drug loading and drug recovery were low, especially in the production of ACR-HSA NPs (Table 5). Covalent binding assays between a drug and HSA using glutaraldehyde suggest that DXR, DNR and THP, but not ACR, could be covalently bound to HSA (Fig. 4). Comparing the structures of these drugs (Fig. 1), the primary amine of DXR, DNR and THP could react with lysine and/or arginine residues of HSA through glutaraldehyde, thereby showing a high drug loading and drug recovery for these drugs. In contrast, ACR which possesses tertiary amine structures could not be involved in the crosslinking reaction, resulting in a reduction of drug loading and recovery. Interestingly, a reduction of these parameters was also observed in THP-HSA NPs even though it possesses a primary amine. The tetrahydropyranyl group of THP may produce steric hindrance in the reaction. Thus, even when this technique is applied to drugs other than anthracyclines, the structure of the drugs, especially the presence of reactive amino groups may become an important factor in producing nanoparticles with higher drug loading properties or production yield.
Such differences in the present states of drugs in drug-HSA NPs also affected the dissolution profiles as one of their functions (Fig. 5). The significant initial burst of release from the ACR-HSA NPs can be assumed to be due to ACR not being involved in the crosslinking reaction. ACR adsorbed in nanoparticles or those embedded near the surface of nanoparticles may be released at the initial phase of dissolution. Drugs that are adsorbed could be released without cleavage of the crosslinking by an enzyme such as proteinase K, since similar initial dissolution patterns have been observed even in ethanol and dissolution buffer in the absence of the proteinase K (data not shown). Based on the dissolution up to 30 min, the proportion of drugs adsorbed on particles appear to be in the following order: ACR-HSA NPs>THP-HSA NPs >> DXR-HSA NPs=DNR-HSA NPs. These data also suggest that most of DXR and DNR and some of THP on the surface of nanoparticles are involved in the crosslinking. Since no further release was observed after the initial release in plasma, plasma may not contain an enzyme that can cleave crosslinking in the nanoparticles. Considering the continuous release after the initial release when proteinase K was used, drugs involved in crosslinking and/or that are embedded inside nanoparticles are gradually released with the cleavage of crosslinking by such an enzyme.
The cytotoxicity of drugs-HSA NPs in tumor cells was lower than that of free drugs for all anthracycline derivatives (Table 6). Thus, in spite of the differences in present states of drugs in nanoparticles, all anthracycline-HSA NPs can be considered to show similar cytotoxic activity in tumor cells. Although ACR-HSA NPs rapidly released ACR in dissolution medium and plasma, its cytotoxicity was also lower than free ACR. The reasons for such decreased cytotoxicity of ACR-HSA NPs as compared with free ACR are unclear, but drug release pattern in cell viability assay might be different from those in dissolution tests using dissolution medium and plasma. The intracellular uptake of drug-HSA NPs followed by the release of the drug and the accumulation of the drug in the nucleus as observed in the case of the DXR-HSA NPs may also be a behavior of DNR-, THP- and ACR-HSA NPs in tumor cells (Fig. 6). Further studies to clarify the behaviors of the nanoparticles will be needed for DNR-, THP- and ACR-HSA NPs. It is well known that several proteases such as serine proteases or cysteine proteases are produced by tumors.26–28) Like proteinase K, proteases in tumor cells such as cathepsin B or matrix metalloproteinase may function to cleave crosslinking and to release drugs from nanoparticles. Langer and colleagues previously reported an equivalent or increased cytotoxicity of DXR-HSA NPs in neuroblastoma cell lines, UKF-NB3 and IMR 32 compared to free DXR.16) However, all our findings using DXR-HSA NPs showed the reduction of the antitumor activity in MCF-7 and HepG2 compared to free DXR. The reason for the differences between these findings remains unclear, but the degradation or uptake mechanisms of nanoparticles may be different in each cell line. Thus, anthracycline-HSA NPs would show antitumor effects in vivo if the nanoparticles could be delivered to tumor cells without significant degradation or elimination during systemic circulation. However, since most of the ACR and some of the THP will be released from the ACR- and THP-HSA NPs in the plasma just after administration, these drugs would not be efficiently delivered to tumor cells by the EPR effect when our produced nanoparticles are used as carriers. Unlike the THP- and ACR-HSA NPs, the DXR- and DNR-HSA NPs could be efficiently delivered to tumors by the EPR effect without the non-enzymatic release of the drug into the plasma. However, the release rate of DXR or DNR from HSA NPs should be optimized to maximize their cytotoxicity in tumor tissue through the cell viability assay using anthracyclines-HSA NPs with different release rates and tumor cells in the presence of proteases.
So far, antitumor, tolerance and biodistribution profiles of DXR-loaded albumin nanoparticles produced by desolvation have been separately investigated in vivo.17,29–31) Using mice transplanted liver cancer cells, Miao et al. showed that the tumor-suppressing effect of DXR-albumin NPs (diameter: 170 nm) was superior to that of DXR alone.17) Using healthy rats, Pereverzeva et al. suggested that DXR-HSA NPs (diameter: 404 nm) reduces cardio- and testicular toxicities observed in free DXR.29) Polyak et al. showed that 73 and 74% of the particles with the diameters of 176 and 429 nm, respectively were located in the liver 5 min after their administration and more than 50% of nanoparticles remained in the body, even after 22 h.30) Furthermore, Farlander et al. suggested that modifying the surface of nanoparticles by polyethylene glycol (PEG) could contribute to the extension of plasma half-life of drug-unloaded HSA nanoparticles (diameter; 140 nm).31) Taking such previous in vivo data and our present data into consideration, the anthracycline-HSA NPs produced by the desolvation technique in present study would also be expected to show therapeutic effects against tumors. However, further modification such as pegylation of these nanoparticles, especially for DXR- and DNR-HSA NPs should be necessary to make the nanoparticles suitable for drug delivery using EPR effect and to improve their therapeutic effects.
In this study, we identified some of the process parameters that can affect the characteristics of DXR-HSA NPs prepared by a desolvation technique. The size of these nanoparticles can be controlled to the extent that they are able to accumulate in tumors through the EPR effect. Using the conditions needed for preparing DXR-HSA NPs in smaller sizes (120 nm), DNR-, THP- and ACR-HSA NPs could be also prepared in the same size. A significant reduction in entrapment efficacy was observed in THP- and ACR-HSA NPs, which can be attributed to the partial and complete lack of covalent binding of these drugs to HSA. Most of the ACR and some of THP appeared to be adsorbed on nanoparticles, thereby promptly releasing drugs without enzymatic cleavage of the crosslinking. Although all the anthracycline-HSA NPs showed cytotoxicity in tumor cell lines, THP- and ACR-HSA NPs failed to be delivered to tumors efficiently since they release drugs readily, even in the plasma. Since the formulation of DXR- and DNR-HSA NPs need to be further modified for efficient delivery to tumors, our findings reported herein provide useful information regarding the formulation and the production processes.
This work was supported in part by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS) (KAKENHI 17K08261).
The authors declare no conflict of interest.
