e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/ejps Thermosensitive polymer-conjugated albumin nanospheres as thermal targeting anti-cancer drug carrier Zheyu Shen a,b,c,1 , Wei Wei a,b,1 , Yongjiang Zhao a,b , Guanghui Ma a,∗ , Toshiaki Dobashi c , Yasuyuki Maki c , Zhiguo Su a , Jinpei Wan a,b a State Key Laboratory of Biochemical Engineering, Institute of Process Engineering, Chinese Academy of Sciences, P.O. Box 353, Beijing 100080, China b College of Chemistry and Chemical Engineering, Graduate University of Chinese Academy of Sciences, 19 Yuquan Road, Shijingshan District, Beijing 100049, China c Department of Biological and Chemical Engineering, Faculty of Engineering, Gunma University, 1-5-1 Tenjin-cho, Kiryu, Gunma 376-8515, Japan a r t i c l e i n f o a b s t r a c t Article history: Thermosensitive Poly(N-isopropylacrylamide-co-acrylamide-co-allylamine) (PNIPAM-AAm- Received 13 March 2008 AA)-conjugated albumin nanospheres (PAN) was developed as a new thermal targeting Received in revised form anti-cancer drug carrier by conjugating PNIPAM-AAm-AA on the surface of albumin 14 July 2008 nanospheres (AN). AN with diameter below 200 nm and narrow size distribution was suc- Accepted 20 July 2008 cessfully prepared in the ﬁrst step with desolvation technique. PNIPAM-AAm-AA with Published on line 31 July 2008 different molecular weight (Mw ) was synthesized in the second step by radical polymerization and conjugated onto the surface of AN. Anti-cancer drug adriamycin (ADR) was then Keywords: entrapped into the AN and PAN during the particle preparation. Compared with AN, the Poly(N-isopropylacrylamide-co- release rate of ADR from PAN in trypsin solution was slower, and decreased with increasing acrylamide-co-allylamine) the conjugation amounts (hairy density) or Mw of PNIPAM-AAm-AA (hairy length). Moreover, Albumin nanospheres the release of ADR from PAN above the cloud-point temperature (Tcp ) of PNIPAM-AAm-AA Adriamycin became faster due to shrinkage of hairy thermosensitive polymer. To testify the thermal Thermal targeting targetability in vivo, PAN was incubated with HepG2 cells. As expected, PAN can target can- Entrapment efﬁciency cer cells above the Tcp of PNIPAM-AAm-AA, whereas it cannot below the Tcp . These results Controlled release might reﬂect that PAN may selectively accumulate onto solid tumors that are maintained above physiological temperature due to local hyperthermia. © 2008 Elsevier B.V. All rights reserved. 1. Introduction The efﬁcacy of cancerous chemotherapy is often limited by serious side effect because of the toxicity of anti-cancer drugs to both tumor and normal cells. A strategy could be to associate anti-cancer drugs with intelligent drug carriers. Recently, stimuli-sensitive polymers have been used in thermosensitive polymeric micelles for the preparation of intelligent drug car- ∗ Corresponding author. Tel.: +86 1082627072; fax: +86 1082627072. E-mail address: [email protected] (G. Ma). 1 These authors contributed equally for this work. 0928-0987/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2008.07.006 riers (Soga et al., 2004; Bae et al., 2003; Taillefer et al., 2000; Schild, 1992; Pelton, 2000; Fujishige et al., 1989). The thermosensitive polymeric micelles, which could accumulate on the heated cancer tissues and cells, have been investigated in drug delivery systems with the aim to increase selectivity of drugs towards cancer cells and reduce their toxicity towards normal tissues (Rijcken et al., 2005; Kohori et al., 1999; Topp et al., 1997; Chung et al., 2000; Inoue et al., 1998; 272 e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 Kwon et al., 1995; Jeong et al., 2003). However, there are two disadvantages for thermosensitive polymeric micelles. Firstly, they were unstable because of being constructed by hydrophobic interaction, which is weak comparing with covalent bond (Xu et al., 2007). Secondly, thermosensitive polymeric micelles with a core–shell structure can only be used as a reservoir for water-insoluble drugs because of the hydrophobic core, but most of the anti-cancer drugs are hydrophilic (Neradovic et al., 2001; Soga et al., 2005). In order to overcome the disadvantages of thermosensitive polymeric micelles, we have designed thermosensitive polymer-conjugated albumin nanospheres (PAN) as a new thermal targeting anti-cancer drug carrier (Shen et al., 2008). This PAN was designed by following strategy. On intravenous administration, particles are normally rapidly coated by the adsorption of speciﬁc blood components known as opsonins and then recognized and taken up by the reticuloendothelial system (RES) (Siegel, 1998). The surface of albumin nanospheres (AN) with drug inside and thermosensitive copolymer chains on the surface is hydrophilic at and below 37 ◦ C due to the hydrophilicity of the copolymer if the cloud-point temperature (Tcp ) of the copolymer is about 42 ◦ C. This hydrophilic copolymer layer can dramatically affect the opsonization of particles by plasma components (Moghimi et al., 1993; Lin et al., 1999). Therefore, the uptake of PAN by the RES can be reduced and a signiﬁcantly longer circulation half-life of the particles in the blood stream will be obtained (Illum and Davis, 1984). The target principle of this carrier is presented in Scheme 1. PAN will circulate in vivo and not precipitate on the healthy tissues and cells because of the hydrophilic surface, but will precipitate on the heated cancer tissues and cells around 42 ◦ C due to local hyperthermia because the copolymer will become hydrophobic at higher temperature due to conformational transition of the chain. Then, the interspace between the polymer chains will increase after they shrunk due to the high temperature and the biodegradable AN will be attacked and degraded easily by proteinase. After that, the drug will release there to attack the tumor cells. Local hyperthermia is an approved method for treatment of solid tumors by selectively heating tumor tissue to 40–44 ◦ C (Schlemmer et al., 2004; Zintchenko et al., 2006). In a previous study, we have prepared biodegradable AN with Rose Bengal (RB, model drug) inside by ultrasonic emulsiﬁcation method, and successfully conjugated Poly(Nisopropylacrylamide-co-acrylamide-co-allylamine) (PNIPAMAAm-AA, thermosensitive polymers) onto AN (Shen et al., 2008). However, there still existed problems: (1) the hydrophilicity of AN was not perfect resulting poor dispersity of AN in water because the oil-soluble emulsiﬁer cannot be removed completely; (2) the effect of conjugation amounts (hairy density) and Mw of polymers (hairy length) on drug release were unable to be studied because the conjugating amounts of polymers on AN were lower due to the low amounts of amino groups at the end of polymers; (3) anticancer drug has not been entrapped into AN for ‘in vitro’ release studies and (4) the targetability of PAN in vivo has not been testiﬁed. In this paper, AN was prepared by desolvation technique method instead of ultrasonic emulsiﬁcation method in order to increase the hydrophilicity of AN, and the conjugating amounts of polymers on AN (hairy density) was raised. Moreover, adriamycin (ADR, anti-cancer drug) instead of RB was entrapped into AN for ‘in vitro’ release studies, and the effect of hairy density and hairy length on drug release was studied. Finally, the targetability of PAN in vivo was testiﬁed. 2. Materials and methods 2.1. Materials N-isopropylacrylamide (NIPAM) and 2,2 -azobis(isobutyronitrile) (AIBN) were purchased from Tokyo Chemical Industry Co. Ltd. (TCI, Japan). NIPAM was puriﬁed by recrystallization from n-hexane. Acrylamide (AAm) was from Sigma (USA). Allylamine (AA), 50% glutaraldehyde (GA) aqueous solution, trypsin (≥250NF U/mg) and 1-ethyl-3-(3dimethylaminopropyl) carbodiimide (EDAC) were ordered from Wako Pure Chemicals (Tokyo, Japan). 8% GA aqueous solution was prepared by 50% GA aqueous solution and Scheme 1 – Target principle scheme of new thermal targeting anti-cancer drug carrier. 273 e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 MilliQ water. Glycine (Gly) and bovine serum albumin (BSA), fraction V (pH 7.0), were purchased from Roche (Germany). ADR was from Beijing HuaFeng United Technology Co. Ltd. Rhodamine–phalloidin (RP) and hoechst 33258 were ordered from Molecular Probes (Eugene, OR). All other reagents were of analytical grade. 2.2. Synthesis of thermosensitive polymers PNIPAM-AAm-AA was synthesized following a procedure developed earlier in our laboratory (Shen et al., 2008). Brieﬂy, 15.55 g of NIPAM, 1.25 g of AAm, AIBN (from 0.05 to 0.40 g) and 100 mL of ethanol were added into a three-necked roundbottle ﬂask with a magnetic stirrer. The round-bottle ﬂask was moved in an oil bath and stirred at 60 ◦ C for 16 h under nitrogen atmosphere. After 16 h, 5.0 mL of AA was added into the mixture. The whole mixture was stirred at 60 ◦ C for 8.0 h under nitrogen atmosphere. The resultant polymers were separated and puriﬁed by reprecipitation into diethylether and then dried in vacuum. Gel permeation chromatography (GPC) measurements were carried out on PNIPAM-AAm-AA in THF at 35 ◦ C. The mean amount of amino groups at the end of polymers was quantiﬁed by conductometric titration. 80 mL of PNIPAM-AAm-AA (5.0 mg/mL) aqueous solutions were titrated with 2.0 mM HCl. The results are summarized in Table 1. 2.3. Preparation of blank and ADR-loaded albumin nanospheres 10 mM NaCl solutions with different pH values (from 9.3 to 11.3) were used to prepare BSA aqueous solutions with variable concentrations (from 20 to 70 mg/mL). 2.0 mL ethanol (desolvation agent) was added into 2.0 mL BSA aqueous solution and stirred for 10 min. After that, various volume of ethanol (from 1.0 to 5.0 mL) was continuously added with different ethanol addition speed (from 0.29 to 4.32 mL/min) under stirring (600 rpm) at room temperature. After ethanol addition, 60 ␮L of 8% GA (120 ␮g GA/mg BSA) in water was rapidly added to induce particle crosslinking. The crosslinking process was performed under stirring of the suspension over 24 h. Variable amount of Gly (from 0.20 to 1.75 mg Gly/mg BSA) was then added to cap free aldehyde groups. After a 2.0 h reaction time, the suspensions were ﬁrstly centrifuged (2000 × g, 20 min) at 15 ◦ C (Allegra 21R Centrifuge, Beckman Coulter) to remove the aggregations of nanoparticles. Then the supernatants were centrifuged (20,000 × g, 20 min) at 15 ◦ C to obtain the AN. The harvested AN was washed two times by pure water to eliminate non-desolvated BSA, the excess of the crosslinker GA and Gly. Between each washing, AN was centrifuged and supernatants were discarded. After that, AN was lyophilized for 48 h. The results are summarized in Table 2. If not speciﬁed, the BSA concentration, ethanol addition speed, volume ratio of ethanol to BSA solution, pH value of BSA solution and Gly/BSA ratio were ﬁxed at 20 mg/mL, 2.0 mL/min, 3.0, 10.8, 1.00, respectively, in the following experiments. ADR-loaded albumin nanospheres (ADR-AN) and corresponding blank albumin nanospheres (Blank-AN) were prepared by the same method with a variable amount of ADR (from 6.25 to 50.0 ␮g/mg BSA) dissolved in 20 mg/mL BSA aqueous solution and a variable concentration of GA (from 20 to 240 ␮g GA/mg BSA). The preparative conditions and the results of these samples are summarized in Table 3. If not speciﬁed, the amount of ADR and the concentration of GA were ﬁxed at 25.0 and 120 ␮g/mg BSA, respectively, in the following experiments. 2.4. Preparation of blank and ADR-loaded PNIPAM-AAm-AA-conjugated albumin nanospheres The harvested AN or ADR-AN sample was dispersed in 10 mL of NaCl solution (pH 4.0, 10 mM). 12.4 mg of EDAC was dissolved in 5.0 mL of ice-cold PNIPAM-AAm-AA aqueous solution (60 mg/mL), and then the solution was added into the AN or ADR-AN suspension and the mixture was stirred at room temperature for different conjugating time (from 2.0 to 14.0 h). After that, PNIPAM-AAm-AA-conjugated-AN (PAN) or PNIPAM-AAm-AA-conjugated-ADR-AN (ADR-PAN) was ultracentrifuged to remove unreacted PNIPAM-AAm-AA and EDAC. Then, the harvested samples were washed one time by pure water and lyophilized. They were stored at −20 ◦ C after their mass being weighted. The mass difference before and after conjugating polymers was used to calculate conjugation amounts. The preparative conditions and the results are summarized in Table 4. 2.5. Optical transmittance determination Optical transmittance (OT) of aqueous PNIPAM-AAm-AA solutions (5.00 mg/mL) was measured from lower to higher temperatures at 500 nm of wavelength with a visible optical spectrophotometer (HITACHI, U-2000 Spectrophotometer). A sample cell with a path length of 10 mm was used. Heating rate Table 1 – Polymerization conditions and results of PNIPAM-AAm-AA Sample PNIPAM-AAm-AA1 PNIPAM-AAm-AA2 PNIPAM-AAm-AA3 PNIPAM-AAm-AA4 a b c AIBN (g) 0.05 0.10 0.20 0.40 Diethylether-insoluble fraction. Determined by GPC. Determined by conductometric titration. Yields (%)a 89 71 95 72 Mw (×104 )b Amounts of –NH2 (mol/mol polymer)c Tcp (◦ C) 4.11 3.54 2.84 1.96 0.74 0.55 1.00 1.01 41.3 41.4 41.7 42.5 274 e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 Table 2 – Preparative conditions and results of AN Sample AN1 AN2 AN3 AN4 AN5 AN6 AN7 AN8 AN9 AN10 AN11 AN12 AN13 AN14 AN15 AN16 AN17 AN18 a CBSA (mg/mL) Ethanol addition speed (mL/min) 20 30 40 50 60 70 20 20 20 20 20 20 20 20 20 20 20 20 Volume ratio of ethanol to BSA solution 2.00 2.00 2.00 2.00 2.00 2.00 0.29 0.74 1.23 4.32 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 1.5 2.5 3.0 3.5 3.0 3.0 3.0 3.0 pH value of BSA solution Yielda (%) 9.8 9.8 9.8 9.8 9.8 9.8 9.8 9.8 9.8 9.8 9.8 9.8 9.8 9.8 9.3 10.8 11.1 11.3 61.1 60.8 57.4 62.7 60.5 59.6 55.8 59.3 58.2 57.1 57.5 59.0 72.8 58.3 73.8 66.3 50 44.3 Diameter (nm) 119.1 135.4 148.8 171.4 189.4 211.2 157.9 147.9 124.7 123.2 105.9 124.9 136.0 147.8 146.0 131.0 124.9 124.0 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.6 0.6 3.4 1.2 2.9 4.2 2.4 3.4 2.8 3.2 2.4 2.9 2.5 0.9 2.8 2.2 2.4 3.2 Calculated from the weight ratio of AN to feed weight of BSA. Table 3 – Preparative conditions and results of ADR-AN Sample ADR/BSA ratio (␮g/mg) GA/BSA ratio (␮g/mg) ADR-AN1 ADR-AN2 ADR-AN3 ADR-AN4 ADR-AN5 ADR-AN6 ADR-AN7 50.0 25.0 12.5 6.25 25.0 25.0 25.0 240 240 240 240 120 60 20 a b Yield (%)a 72.1 70.9 71.5 70.7 69.9 65.6 60.5 Drug loading ratio (␮g ADR/mg AN)b 17.26 15.37 8.92 4.11 15.68 13.68 9.11 Entrapment efﬁciency (%)b 34.3 64.2 66.5 68.8 61.9 56.8 35.5 Calculated from the weight ratio of AN to feed weight of BSA and drug. Determined by UV spectrophotometer and calculated from the standard calibration curve. was 0.1 ◦ C/min. The Tcp of a polymer solution was determined at a temperature showing an optical transmittance of 50%. 2.6. Dynamic light scattering measurement and conductometric titration Particle size and size distribution of particles were measured by dynamic light scattering (DLS) at the scattering angle of 90◦ in pure water at room temperatures using 90 Plus Particle Size Analyzer (Brookhaven Instruments Corporation). The particle solutions were optically cleaned through a 0.45 ␮m membrane ﬁlter just before measurements. The amount of carboxyl groups on the surface of one AN was also quantiﬁed by conductometric titration. 28 mg of AN was dissolved in 80 mL pure water and titrated with 10.0 mM NaOH. PAN was measured by DLS from lower to higher temperatures in pure water. The solutions were equilibrated at given temperatures for 10 min before each measurement. Diffusion coefﬁcient was obtained by cumulant method and the Table 4 – Preparative conditions and results of ADR-PAN Sample ADR-PAN1 ADR-PAN2 ADR-PAN3 ADR-PAN4 ADR-PAN5 ADR-PAN6 ADR-PAN7 a Conjugating time (h) 2.0 4.0 8.0 14.0 14.0 14.0 14.0 PNIPAM-AAm-AAa PNIPAM-AAm-AA4 PNIPAM-AAm-AA4 PNIPAM-AAm-AA4 PNIPAM-AAm-AA4 PNIPAM-AAm-AA1 PNIPAM-AAm-AA2 PNIPAM-AAm-AA3 Conjugating amounts (×104 mol/mol AN) 3.23 3.38 4.90 5.98 1.92 1.79 3.37 The kind of thermosensitive copolymers used for the conjugating reaction. Drug loading ratio (␮g ADR/mg AN) 10.37 9.11 8.13 7.45 9.16 9.62 8.78 Entrapment efﬁciency (%) 45.5 40.6 42.4 42.8 43.8 42.2 45.7 e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 275 Blank-AN, ADR-AN and ADR-PAN were observed by a JEM100cx (JEOL, Japan) transmission electron microscope (TEM). Approximately 2.0 ␮L of the diluted nanoparticles were mounted on copper grids, and then dried at room temperature. After that, they were observed by TEM (no coloration). dish. On days 3 and 4 of culture, PAN was added into the culture medium without fetal bovine serum at a concentration of 1.0 mg/mL. Cells were then incubated at 43 or 37 ◦ C for 0.5 h and the free nanospheres in the medium were removed by washing the cell layer three times with ice-cold phosphatebuffered saline. For the LSCM imaging, the cells were ﬁxed in paraformaldehyde for 0.5 h, and stained actin with RP and nuclear with hoechst 33258. The samples were excited at 364, 488 and 543 nm. Three ﬂuorescent images at different wavelengths (420–450, 510–540 and 570–600 nm) were taken by LSCM. 2.8. Determination of drug loading ratio and entrapment efﬁciency 3. Results and discussion 3.1. Polymerization results hydrodynamic diameter (Dh ) was estimated from the diffusion coefﬁcient by using the Stokes–Einstein equation. 2.7. Transmission electron microscopic observation of nanospheres The ADR content was measured by digesting 24 mg of ADR-AN (or ADR-PAN) and corresponding Blank-AN (or Blank-PAN) in 6 mL of 20 mg/mL trypsin solution in the dark for 4.0 h at 37 ◦ C, respectively. The digested solutions of Blank-AN and BlankPAN were used to make baseline of absorbance measurement by UV spectrophotometer (HITACHI, U-2000 Spectrophotometer). The absorbance (at 480 nm of wavelength) of the digested solutions of ADR-AN (or ADR-PAN) was converted into the concentration of ADR using a calibration curve constructed with standard ADR solutions containing 4.0 mg/mL BSA and 20 mg/mL trypsin. The drug loading ratio (DLR, ␮g ADR/mg AN) and entrapment efﬁciency (EE, %) were calculated from the following formula: DLR = total amount of ADR loaded (␮g) total amount of nanospheres harvested (mg) EE (%) = total amount of ADR loaded × 100 total amount of ADR added 2.9. ‘In vitro’ release studies (1) (2) 30 mg ADR-AN (or ADR-PAN) was dispersed in 20 mL of NaCl solution (pH 7.4, 10 mM) with or without 2.0 mg/mL of trypsin. The suspensions were stirred slowly and incubated in a water bath at 37 ± 1 ◦ C. At predetermined time intervals, these samples were ultracentrifuged at 20,000 × g for 20 min and 1.0 mL of these supernatants were taken and analyzed for concentration of the released ADR by spectrophotometer. The release of ADR from ADR-AN without and with 2.0 mg/mL of trypsin was determined at 37 ± 1 ◦ C. Moreover, with 2.0 mg/mL of trypsin, the release of ADR from ADR-PAN was determined at both 37 ± 1 and 43 ± 1 ◦ C. 2.10. Observation of the autoﬂuorescent property The suspension of AN in Petri dish (Mattek) was observed by LSCM. Sample was excited at 488 nm and two ﬂuorescent images at different wavelengths (510–540 and 570–600 nm) were taken. 2.11. Bioadhesion of PAN on HepG2 Cells The human hepatocellular carcinoma HepG2 cells of passages 30–33 were plated at a density of 2.0 × 105 cells/cm2 in Petri The results for polymerization and characterization of PNIPAM-AAm-AA are summarized in Table 1. The yields were determined as diethylether-insoluble fraction for PNIPAMAAm-AA and the impurities (such as NIPAM, AAm, AA, ethanol, AIBN and oligomers) were removed because they are soluble in diethylether. In order to overcome the problem in previous study that the conjugating amounts of polymers on AN were too low, in this study, the mean amount of amino groups at the end of polymers was enhanced by increasing the amounts of AA and reaction time. So that, the effect of conjugation amount (hairy density) on drug release could be studied. Although the mean amount of amino groups at the end of polymers was enough for conjugation of polymers on AN, it was still low as shown in Table 1. That was because side reactions have terminated propagating polymer chains in some extent before the addition of allylamine (e.g. chain transfer to ethanol or termination by disproportionation or coupling of two growing chains) and these undesirable side reactions led to mixture of amino functionalized polymers (successful addition of AA) and polymers without allylamine content. Moreover, free radicals that were not connected to polymer chains were also present due to relatively slow degradation of AIBN at 60 ◦ C. Therefore, allylamine might also was homopolymerized in the solution after the addition step. The homopolymerized allylamine could be removed from the polymers because it was soluble in diethylether due to lower molecular weight. As the amino content was concerned, the result in Table 1 was the mean value in the mixture of amino functionalized polymers and polymers without allylamine content. Polymers without amino functionality were most likely removed during the ultracentrifugation in conjugation step. Therefore, the conjugated polymers on AN carried higher amino amounts compared to the mean value. Furthermore, the Mw of PNIPAM-AAm-AA was determined by GPC in THF at 35 ◦ C. It was found that Mw of PNIPAMAAm-AA decreased with increasing amount of AIBN, whereas the Tcp was almost the same. Therefore, the effect of Mw of polymers (hairy length) on drug release also could be studied. Tcp was determined for PNIPAM-AAm-AA when the optical transmittance of the solutions crossed 50% as the temperature increased (see corresponding optical transmittance in Fig. S1). In a previous study, we have found that the reciprocals of Tcp of 276 e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 Fig. 1 – Diameter of BSA nanospheres prepared at different BSA concentrations (a), ethanol addition speed (b) and pH value (c) in the presence of 10 mM NaCl (mean ± S.D., n = 3). aqueous solution of PNIPAM-AAm increased linearly against volume fraction of NIPAM (Shen et al., 2006). Therefore, Tcp of PNIPAM-AAm-AA can be controlled by . Indeed, the biocompatibility should be considered. According to a toxicity test of the PNIPAM, which was performed in comparison with that of the NIPAM monomer, PNIPAM showed no toxicity in mice (Sheng et al., 2006). Therefore, although NIPAM, AAm and AA are toxic monomers, PNIPAM-AAm-AA is non-toxic and may be removed from the body due to relatively low molecular weight. 3.2. Optimization of nanosphere preparation by yield and average particle size In order to optimize the conditions of nanosphere preparation, AN1–18 were prepared at different preparative conditions. The results and conditions are summarized in Table 2. AN1–6 were prepared at different BSA concentrations, and the yields determined as weight ratio of AN to feed weight of BSA were 61.1, 60.8, 57.4, 62.7, 60.5 and 59.6%, respectively. The effect of BSA concentrations on particle diameter is shown in Fig. 1a. It is interesting to ﬁnd a linear relationship between the size of the particles and the BSA concentration. However, it is beyond of the scope of this study. The size distributions of AN1–6 are shown in Fig. 2. It indicates that there are no aggregations of the particles in aqueous system. The subsequent experiments were performed at a concentration of BSA aqueous solution of 20 mg/mL because of the smallest particle size and the narrowest size distribution. AN7–10 were prepared at different ethanol addition speed, and the yields were 55.8, 59.3, 58.2 and 57.1%, respectively. The inﬂuence of the ethanol addition speed on particle diameter of the resulting samples is shown in Fig. 1b. The mean particle diameter ﬁrst decreased rapidly with increase of the ethanol addition speed. However, it did not change apparently when the ethanol addition speed was higher than 1.23 mL/min. The subsequent experiments were performed at an ethanol addition speed of 2.00 mg/mL because of the smallest mean particle diameter. AN11–14 were prepared at various volume ratio of ethanol to BSA solution, and the yields were 57.5, 59.0, 72.8 and 58.3%, respectively. The mean particle diameter increased with increase of the volume ratio of ethanol to BSA solution (see the corresponding data in Table 2). Excessive desolvation agent (ethanol) might result in the aggregation of AN, which was removed from AN by centrifugation at 2000 × g. So that, higher volume ratio of ethanol to BSA solution (excessive ethanol) resulted in lower yield. Lacking in desolvation agent might result in non-desolvated BSA. Therefore, lower volume ratio of ethanol to BSA solution (lacking in ethanol) also resulted in lower yield. Because of the high yield (72.8%), the subsequent experiments were performed at a volume ratio of ethanol to BSA solution of 3.0. AN15–18 were prepared at different pH value of BSA solution, and the yields were 73.8, 66.3, 50.0 and 44.3%, respectively. The inﬂuence of the pH value on particle diameter of the resulting samples is shown in Fig. 1c. The mean particle diameter decreased with the increase of pH value of BSA solution. This was because the higher pH value of BSA solution, the stronger electrostatic repulsive force of BSA molecules (pI ≈ 4.7) which resulted in smaller nanoparticles. However, the hydrolysis of BSA at strong alkaline solution led to lower yield. In order to prepare AN with higher yield and smaller particle size, the subsequent experiments were performed at a pH 10.8 of BSA solution. Fig. 2 – Size distribution of BSA nanospheres according to 90 Plus Particle Size Analyzer. e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 277 Fig. 3 – TEM photos of Blank-AN2 observed at 33,000× magniﬁcation (a), Blank-AN5 observed at 33,000× magniﬁcation (b), Blank-AN6 observed at 61,000× magniﬁcation (c), Blank-AN7 observed at 33,000× magniﬁcation (d), ADR-AN5 observed at 49,000× magniﬁcation (e) and ADR-PAN1 observed at 81,000× magniﬁcation (f). Free aldehyde groups of AN16 were capped at different amount of Gly, and then the amount of carboxyl groups on the surface of one AN16 was quantiﬁed by conductometric titration with 10.0 mM NaOH. It was 41.1, 42.6, 53.0, 61.3 and 61.1 × 104 when the amount of Gly was 0, 0.20, 0.50, 1.00 and 1.75 mg Gly/mg BSA, respectively. The amount of carboxyl group on the surface of one particle ﬁrst increased with increasing the amount of Gly. However, little change was found when the amount of Gly was higher than 1.00 mg Gly/mg BSA, which implied that the free aldehyde groups have all been capped (see the inﬂuence of the Gly/BSA ratio on the amount of carboxyl group on the surface of one particle in Fig. S2). Therefore, the subsequent experiments were performed at 1.00 mg Gly/mg BSA. Considering the inﬂuence of these different parameters on particle size and yield, a standard protocol for the preparation of AN was established as follows: the nanospheres were prepared in NaCl solution (pH 10.8, 10 mM) at 20 mg/mL of BSA concentration, a 2.00 mg/mL of ethanol addition speed, 3.0 volume ratio of ethanol to BSA solution and 1.00 mg Gly/mg BSA. Together these conditions led to particles with average diameters below 200 nm at a particle yield of about 70%. 3.3. TEM observation of blank and ADR-loaded albumin nanospheres Fig. 3a–d is TEM photos of nanospheres showing the effect of crosslinker amount. Blank-AN2 and Blank-AN5, spherical in shape, were dispersed well in pure water implying good hydrophilicity of particles and the coalescence was hardly found. On the contrary, obvious aggregation and poor sphericity was revealed for Blank-AN6 and Blank-AN7 due to low particle crosslinking degree. Therefore, the concentration of GA should be 120 ␮g/mg BSA at least. From Fig. 3e, the morphology of ADR-AN5 was similar to that of Blank-AN5. 3.4. Optimization of nanosphere preparation by drug loading ratio and entrapment efﬁciency In a previous paper, the ability of AN to carry a model drug RB was evaluated (Shen et al., 2008). In this paper, desolvation technique method was used instead of ultrasonic emulsiﬁcation method, an anti-cancer drug ADR was entrapped and the inﬂuence of the ADR/BSA ratio and the concentration of glutaraldehyde on the drug loading ratio and the entrapment 278 e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 efﬁciency initially increased rapidly with increasing the glutaraldehyde concentration. However, they were signiﬁcantly close when the glutaraldehyde concentration was higher than 120 ␮g GA/mg BSA. Considering the inﬂuence of these parameters on the drug loading ratio and the entrapment efﬁciency, the optimal amount of ADR and concentration of GA were 25.0 and 120 ␮g/mg BSA, respectively. 3.5. Fig. 4 – Inﬂuence of ADR/BSA ratio (a) and concentration of glutaraldehyde (b) on the drug loading ratio (␮g ADR/mg AN) and the entrapment efﬁciency (expressed in %) in AN. efﬁciency were investigated in detail. Corresponding results are summarized in Table 3. Fig. 4a shows the capacity of AN for ADR loading as a function of ADR/BSA ratio. With increase of the ADR/BSA ratio, the drug loading ratio increased quickly at ﬁrst, but increased slowly at ADR/BSA ratios higher than 25.0 ␮g ADR/mg BSA. On the other hand, the entrapment efﬁciency decreased gently at ﬁrst, but decreased rapidly at higher than 25.0 ␮g ADR/mg BSA. Fig. 4b shows that the drug loading ratio and the entrapment ADR release from albumin nanospheres Fig. 5 shows ADR release from AN particle in pH 7.4 NaCl solution (10 mM) with or without 2.0 mg/mL trypsin at 37 ◦ C showing the effect of crosslinker amount. The release of ADR from ADR-AN2 and ADR-AN5 was very slow and their release behavior was similar. The release rate of ADR from ADR-AN6 and ADR-AN7 increased because of low particle crosslinking degree. In the presence of trypsin, the release of ADR from all of the particles was dramatically accelerated. For the purpose of using the AN as anti-cancer drug carrier for tumor targeting, the slower drug release rate was during circulation in the blood stream, the weaker side effect of anti-cancer drugs would be achieved. On the contrary, the release rate should increase at the target sites (cancer cells or cancer tissues) where there are more enzymes than the blood (Muller et al., 1996). The release of ADR from AN in vitro at 37 ◦ C in the absence of the proteinase trypsin was slow and dramatically increased in the presence of trypsin, which indicates that the AN can be used as an anti-cancer drug carrier. 3.6. Preparation result of PNIPAM-AAm-AA-conjugated albumin nanospheres PNIPAM-AAm-AA with amino group at the end of the polymer chains was conjugated on the surface of AN in the presence of EDAC. The results are summarized in Table 4. Using the mass difference before and after conjugating polymers, the amounts of PNIPAM-AAm-AA on the surface of AN were determined to be 28.96, 30.25, 43.93, 53.57, 17.21, 16.04 and 30.18 nmol/mg AN for ADR-PAN1–7, respectively. Fig. 5 – ADR release from ADR-AN2, ADR-AN5, ADR-AN6 and ADR-AN7 in NaCl solution (pH 7.4, 10 mM) with or without 2.0 mg/mL trypsin at 37 ◦ C. e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 279 Approximating 4 3 3 d̄ 2 4 d0 ∼ =n× 3 2 3 (3) where d̄ is the mean diameter of AN, d0 is the mean diameter of BSA and n is the amount of BSA molecular in one nanosphere. d0 was determined to be 5.1 nm by DLS at room temperature with the concentration of 5.0 mg/mL. From Eq. (3), we can obtain that n ∼ = 16,900 because of d̄ = 131 nm. So that, we can calculate that the amount of AN is 8.96 × 10−13 mol within 1.0 mg AN (approximating that 1.0 mg AN was consisted of 1.0 mg BSA). Therefore, the amounts of PNIPAM-AAm-AA on the surface of AN were calculated to be 3.23, 3.38, 4.90, 5.98, 1.92, 1.79 and 3.37 mol/mol AN for ADR-PAN1–7, respectively. They were 300–400 mol polymer/mol AN in the previous study, and were enhanced to 10,000–60,000 mol polymer/mol AN in this study by increasing the amounts of amino groups at the end of PNIPAM-AAm-AA. Furthermore, the low drug loading ratio and entrapment efﬁciencies of ADR-PAN were ascribed to the loss of drug during the conjugation process. From Fig. 3f, it is obvious that PNIPAM-AAm-AA was not visible by TEM and the morphology of ADR-PAN1 was similar to that of ADR-AN5. Fig. 6 – Temperature dependence of Dh of PNIPAM-AAm-AA-conjugated albumin nanospheres in water for PAN1, PAN2, PAN3 and PAN4 (mean ± S.D., n = 3). Fig. 7 – Effect of hairy density (a), hairy length (b) and temperature (c and d) on ADR release from ADR-PAN in NaCl solution (pH 7.4, 10 mM) containing 2.0 mg/mL trypsin. Error bars represent calculations of standard error on the basis of quintuplicate determinations. 280 e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 3.7. Hydrodynamic diameter of PNIPAM-AAm-AA-conjugated albumin nanospheres As shown in Fig. 6, the Dh of PAN decreased with increasing temperature from 30.0 to 43.0 ◦ C because of the change in conformation of PNIPAM-AAm-AA chains from random coil to compact globule, and then increased with increasing temperature from 43.0 to 60.0 ◦ C because of aggregation of nanoparticles. In the previous paper (Shen et al., 2008), the increasing behavior of Dh of PAN was not found with increasing temperature that was because no nanoparticle aggregation formed at higher temperature due to the low conjugating amounts of PNIPAM-AAm-AA. Furthermore, compared with the previous results, with increase of temperature from 30.0 to 43.0 ◦ C, the decreasing value of Dh of PAN was small. That was probably because part of nanoparticles aggregated during the change in conformation of PNIPAM-AAm-AA chains from random coil to compact globule due to high conjugating amounts. 3.8. ADR release from PNIPAM-AAm-AA-conjugated albumin nanospheres In the previous paper (Shen et al., 2008), only two kinds of PNIPAM-AAm-AA were conjugated onto AN, and the effects of hairy density and hairy length on drug release have not been investigated. In this paper, it became possible to investigate the effect of hairy density and hairy length on ADR release from PAN by successfully increasing the conjugating amounts of PNIPAM-AAm-AA on AN particle. Fig. 7a shows the ADR release from ADR-PAN1–4 with different hairy density at 37 ◦ C in pH 7.4 NaCl solution containing 2.0 mg/mL trypsin. Apparently, the ADR release rate decreased with the increase of the hairy density. The hairy density was almost the same for ADR-PAN2 and ADR-PAN7 with 3.38 × 104 and 3.37 × 104 mol PNIPAM-AAmAA/mol AN, respectively. But the hairy length was different for ADR-PAN2 and ADR-PAN7 with Mw of 1.96 × 104 and 2.84 × 104 . As shown in Fig. 7b, the ADR release from ADR-PAN2 was faster than that from ADR-PAN7 at 37 ◦ C, which indicated that the longer hairy length led to slower drug release rate. Moreover, as shown in Fig. 7c and d, ADR release from both ADR-PAN2 and ADR-PAN7 were faster at 43 ◦ C than at 37 ◦ C. The release rate of ADR from ADR-PAN decreased with increase of the hairy density and hairy length, which suggested that the existence of a steric hydrophilic barrier on the surface of the AN made digestion of the AN more difﬁcult. Moreover, the release of ADR from ADR-PAN above Tcp of PNIPAM-AAm-AA was faster than that below Tcp , implying that thermosensitive polymers on the surface of ADR-PAN were not so dense and the interspace between the thermosensitive polymer chains increased after they shrunk due to the high temperature, so that the biodegradable albumin nanospheres were attacked and degraded easily by trypsin, and correspondingly the drug release rate from the PAN increased. From above results, it was expected that ADR-PAN can circulate in blood stream with a longer half-life after intravenous injection and less drug is released, while the drug is released faster after ADR-PAN is accumulated in tumor cells. Fig. 8 – LSCM images of crosslinked BSA nanoparticles observed at different wavelengths: (a) 510–540 nm, (b) 570–600 nm and (c) overlay. The margin of the suspension was taken as a control to conﬁrm the ﬂuorescent property of these crosslinked BSA nanoparticles. The nether panel was the higher magniﬁed images of upper. e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 281 Fig. 9 – Bioadhesion of PAN under different temperature: (a) HepG2 cells incubated with PAN2 for 0.5 h at 37 ◦ C; (b) HepG2 cells incubated with PAN2 for 0.5 h at 43 ◦ C; (c) the higher magniﬁed image of (b). 3.9. Autoﬂuorescent property and selective bioadhesion of PAN on HepG2 cells PAN2 was added in the culture medium of HepG2 cells to testify their bioadhesion ability under different temperatures. We ﬁrstly found that the nanospheres crosslinked with glutaraldehye showed great ﬂuorescent under LSCM as shown in Fig. 8. One possible explanation is that the C N bonds formed between the free amino groups of BSA and the aldehyde groups from the crosslinker might contribute to the ﬂuorescent phenomenon. Similar result for crosslinked chitosan microsphere has been found by us (Wei et al., 2007). With the advantage of autoﬂuorescent property, we can directly observe these nanospheres by LSCM without conjugating with any ﬂuorescent reagent. As shown in Fig. 9, it is obvious that great numbers of PAN2 adhered on the surface of the cell membrane under 43 ◦ C, whereas few PAN2 was found under 37 ◦ C. This result might reﬂect the potential clinic application of PAN for anti-cancer therapy. Animal experiments are being carried out to prove the advantage of this new drug carrier. increase of drug release rate at the target sites (cancer cells or cancer tissues) after intravenous injection. From these results, it was expected that PAN can be used as thermal targeting anticancer drug carrier and can eliminate undesirable side effects generated by free drugs for cancerous chemotherapy. Acknowledgments Financial support from the Knowledge Innovation Program Pilot Project of the Chinese Academy of Sciences (No. KJCX2.YW.M02) and the National Nature Science Foundation of China (contract nos. 20536050, 20221603 and 20376082) are gratefully acknowledged. The authors thank the Ministry of Education, Culture, Sports, Sciences, and Technology of Japan for providing scholarship to the ﬁrst author to pursue his research at the Department of Biological and Chemical Engineering, Faculty of Engineering, Gunma University. Appendix A. Supplementary data 4. Conclusion In this paper, a novel thermal targeting anti-cancer drug carrier PAN was developed by conjugating PNIPAM-AAm-AA on the surface of AN, which was prepared by desolvation technique method and showed good hydrophilicity of particles. DLS measurement showed good temperature-sensitive property of PAN after enhanced the conjugating amounts of polymers. Bioadhesion experiment of PAN on HepG2 cells showed that PAN can target cancer cells above Tcp of PNIPAMAAm-AA, whereas it cannot below Tcp of PNIPAM-AAm-AA, which indicated that PAN can selectively accumulate on solid tumors that are maintained above physiological temperature due to local hyperthermia. Furthermore, ADR was entrapped into AN and PAN. ‘In vitro’ release studies proved the slow drug release from ADR-AN in the absence of the proteinase and dramatic increase in the presence of proteinase. The ADR release rate from ADR-PAN decreased with the increase of the hairy density and hairy length. Fast drug release from PAN at the temperature above Tcp of PNIPAM-AAm-AA indicated the Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ejps.2008.07.006. references Bae, Y., Fukushima, S., Harada, A., Kataoka, K., 2003. Design of environment-sensitive supramolecular assemblies for intracellular drug delivery: polymeric micelles that are responsive to intracellular pH change. Angew. Chem. Int. Ed. 42, 4640–4643. Chung, J.E., Yokoyama, M., Okano, T., 2000. Inner core segment design for drug delivery control of thermo-responsive polymeric micelles. J. Control. Release 65, 93–103. Fujishige, S., Kubota, K., Ando, I., 1989. Phase transition of aqueous solutions of poly(N-isopropylacrylamide) and poly(N-isopropylmethacrylamide). J. Phys. Chem. 93, 3311–3313. Illum, L., Davis, S.S., 1984. The organ uptake of intravenously administered colloidal particles can be altered using a non-ionic surfactant (Poloxamer 338). FEBS Lett. 167, 79–82. 282 e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 5 ( 2 0 0 8 ) 271–282 Inoue, T., Chen, G., Nakamae, K., Hoffman, A.S., 1998. An AB block copolymer of oligo(methyl methacrylate) and poly(acrylic acid) for micellar delivery of hydrophobic drugs. J. Control. Release 51, 221–229. Jeong, J.H., Kim, S.W., Park, T.G., 2003. A new antisense oligonucleotide delivery system based on self-assembled ODN–PEG hybrid conjugate micelles. J. Control. Release 93, 183–191. Kohori, F., Sakai, K., Aoyagi, T., Yokoyama, M., Yamato, M., Sakurai, Y., Okano, T., 1999. Control of adriamycin cytotoxic activity using thermally responsive polymeric micelles composed of poly(N-isopropylacrylamide-co-N,Ndimethylacrylamide)-b-poly(d,l-lactide). Colloids Surf. B: Biointerf. 16, 195–205. Kwon, G.S., Naito, M., Yokoyama, M., Okano, T., Sakurai, Y., Kataoka, K., 1995. Physical entrapment of adriamycin in AB block copolymer micelles. Pharm. Res. 12, 192–195. Lin, W., Garnett, M.C., Schacht, E., Davis, S.S., Illum, L., 1999. Preparation and in vitro characterization of HSA–mPEG nanoparticles. Int. J. Pharm. 189, 161–170. Moghimi, S.M., Muir, I.S., Illum, L., Davis, S.S., Bachofen, V.K., 1993. Coating particles with a block co-polymer (poloxamine-908) suppresses opsonization but permits the activity of dysopsonins in the serum. Biochim. Biophys. Acta 1179, 157–165. Muller, B.G., Leuenberger, H., Kissel, T., 1996. Albumin nanospheres as carriers for passive drug targeting: an optimized manufacturing technique. Pharm. Res. 13, 32–37. Neradovic, D., Nostrum, C.F., Hennink, W.E., 2001. Thermoresponsive polymeric micelles with controlled instability based on hydrolytically sensitive N-isopropylacrylamide copolymers. Macromolecules 34, 7589–7591. Pelton, R., 2000. Temperature-sensitive aqueous microgels. Adv. Colloid Interf. Sci. 85, 1–33. Rijcken, C.J.F., Veldhuis, T.F.J., Ramzi, A., Meeldijk, J.D., Nostrum, C.F., Hennink, W.E., 2005. Novel fast degradable thermosensitive polymeric micelles based on PEG-blockpoly(N-(2-hydroxyethyl)methacrylamide-oligolactates). Biomacromolecules 6, 2343–2351. Schild, H.G., 1992. Poly(N-isopropylacrylamide): experiment, theory and application. Prog. Polym. Sci. 17, 163–249. Schlemmer, M., Lindner, L.H., Abdel-Rahman, S., Issels, R.D., 2004. Principles, technology and indication of hyperthermia and part body hyperthermia. Radiologe 44, 301–309. Shen, Z.Y., Ma, G.H., Dobashi, T., Maki, Y., Su, Z.G., 2008. Preparation and characterization of thermo-responsive albumin nanospheres. Int. J. Pharm. 346, 133–142. Shen, Z.Y., Terao, K., Maki, Y., Dobashi, T., Ma, G.H., Yamamoto, T., 2006. Synthesis and phase behavior of aqueous poly(N-isopropylacrylamide-co-acrylamide), poly(N-isopropylacrylamide-co-N,N-dimethylacrylamide) and poly(N-isopropylacrylamide-co-2-hydroxyethyl methacrylate). Colloid Polym. Sci. 284, 1001–1007. Sheng, X.Z., Liu, Z.Q., Wu, L.B., Tang, J., Zhao, C.R., Kong, L.B., Wang, Q., Wang, C.D., 2006. Technical feasibility and histopathologic studies of poly(N-isopropylacryl-amide) as a non-adhesive embolic agent in swine rete mirabile. Chin. Med. J. 119, 391–396. Siegel, R.A., 1998. Drug delivery: a lesson from secretory granules. Nature 394, 427–428. Soga, O., Nostrum, C.F., Fens, M., Rijcken, C.J.F., Schiffelers, R.M., Storm, G., Hennink, W.E., 2005. Thermosensitive and biodegradable polymeric micelles for paclitaxel delivery. J. Control. Release 103, 341–353. Soga, O., Nostrum, C.F., Ramzi, A., Visser, T., Soulimani, F., Frederik, P.M., Bomans, P.H.H., Hennink, W.E., 2004. Physicochemical characterization of degradable thermo-sensitive polymeric micelles. Langmuir 20, 9388–9395. Taillefer, J., Jones, M.C., Brasseur, N., Lier, J.E., Leroux, J.C., 2000. Preparation and characterization of pH-responsive polymeric micelles for the delivery of photosensitizing anticancer drugs. J. Pharm. Sci. 89, 52–62. Topp, M.D.C., Dijkstra, P.J., Talsma, H., Feijen, J., 1997. Thermosensitive micelle-forming block copolymers of poly(ethylene glycol) and poly(N-isopropylacrylamide). Macromolecules 30, 8518–8520. Wei, W., Wang, L.Y., Yuan, L., Wei, Q., Yang, X.D., Su, Z.G., Ma, G.H., 2007. Preparation and application of novel microspheres possessing autoﬂuorescent property. Adv. Funct. Mater. 17, 3153–3158. Xu, H.X., Xu, J., Jiang, X.Z., Zhu, Z.Y., Rao, J.Y., Yin, J., Wu, T., Liu, H.W., Liu, S.Y., 2007. Thermosensitive unimolecular micelles surface decorated with gold nanoparticles of tunable spatial distribution. Chem. Mater. 19, 2489–2494. Zintchenko, A., Ogris, M., Wagner, E., 2006. Temperature dependent gene expression induced by PNIPAM-based copolymers: potential of hyperthermia in gene transfer. Bioconjug. Chem. 17, 766–772.