Selective targeting by preS1 domain of hepatitis B surface antigen conjugated with phosphorylcholine-based amphiphilic block copolymer micelles as a biocompatible, drug delivery carrier for treatment of human hepatocellular carcinoma with paclitaxel



Using dithioester-capped 2-methacryloyloxyethyl phosphorylcholine (MPC) as a macro chain transfer agent, a diblock copolymer was synthesized with n-butyl methacrylate (BMA) as hydrophobic core-forming blocks. The MPC–BMA unit was copolymerized with an immobilizable unit, p-nitrophenylcarbonyloxyethyl methacrylate (NPMA). The NPMA moiety then was modified by the addition of preS1 domain of hepatitis B surface antigen (HBsAg). This micelle-forming nanoparticle, the poly (MPC-co-BMA-co-NPMA) (PMBN) conjugated with preS1 enables solubilization of paclitaxel (PTX) with increased hepatotropism. The 50% inhibitory concentration (IC50) values of PTX and PTX/PMBN-preS1 against the human hepatocellular carcinoma cell line, HepG2, were 1,008 and 131 nM, respectively (p < 0.05). Conjugation of preS1 to PMBN enhanced strongly the synergistic inhibitory effect of paclitaxel on HepG2 cells in vitro, whereas such a change in IC50 was not detected against the human squamous cell carcinoma cell line, A431. Tumor growth rates of a HepG2 xenograft in Balb/c nude mice after intraperitoneal injection of PTX, PTX/PMBN and PTX/PMBN-preS1 were +97.9%, −74.3% and −96.2%*, respectively (*p < 0.05 versus PTX). The local paclitaxel levels after administration of the PMBN-preS1 conjugate were determined in the xenografts by high-performance liquid chromatography and were 8 times higher than that after administration of paclitaxel alone. No side effects attributable to PMBN-preS1 were observed histologically in vital organs, and body weight loss was significantly less in the PTX/PMBN-preS1 group. These studies demonstrate that PMBN-preS1 may be used as a human hepatocyte-specific drug delivery carrier without serious adverse effects. © 2008 Wiley-Liss, Inc.

Hepatocellular carcinoma (HCC) is one of the most prevalent cancers worldwide.1 Hepatic resection by partial hepatectomy and orthotopic liver transplantation are the only curative treatments and complete cures may be achieved in fewer than 30% of HCC patients.2 Various nonsurgical treatments have been offered to patients whose tumors have been deemed unresectable but all have proved to be palliative. In particular, the results of systemic chemotherapy are most disappointing, with a response rate of less than 20%.3

This prominent chemoresistant feature of HCC might be attributable to its spheroid-like architecture, which leads to lower drug uptake and the high rate of expression of the multidrug resistance gene in most HCCs.4 In addition, the cancer cells originate from the liver, which is responsible for the detoxification of exogenous and endogenous chemicals.5 In overcoming the microenvironment, coupled with these intrinsic factors, a selective and effective drug delivery system is crucial to enhance antitumor activity. Recently, various molecular targeting therapies have been developed for HCCs and tested widely in preclinical studies.6 Unfortunately, there is a discrepancy between the very promising experimental data and the results obtained in patients. Sorafenib, the most promising molecular targeting agent, is an oral multikinase inhibitor with activity against several tyrosine (VEGFR2, PDGFR, c-Kit receptors) and serine/threonine (b-Raf) kinases but resulted in a median overall survival of only 10.7 months in a randomized phase III trial.7, 8

Paclitaxel (PTX) has been approved for treating breast, ovarian and lung cancer. Although some reports have already shown an antitumor effect of PTX against HCC, both in vitro and in vivo,9–11 clinical application of its use has yet to be established. The main reason for this is the high-grade hydrophobicity of PTX, which makes the drug delivery to the target organ quite difficult.

Micelle-forming nanoparticles have been developed for drug delivery purposes because of increased water solubility of sparingly soluble drugs, such as PTX, and improved bioavailability.12, 13 In particular, amphiphilic block copolymers have proved to be one of the most promising families of materials for use in micellar drug formulation and delivery.14, 15 Using dithioester-capped phosphorylcholine (MPC) as a macro chain transfer agent, the diblock copolymer may be synthesized with n-butyl methacrylate (BMA) as hydrophobic core-forming blocks.16 Phosphorylcholine-based polymers mimic the surface of natural phospholipid membrane bilayers and effectively suppress any unfavorable interactions with biocomponents, improving the blood compatibility of the nanoparticles. The MPC–BMA unit may be copolymerized with an immobilizable unit, p-nitrophenylcarbonyloxyethyl methacrylate (NPMA). The NPMA moiety contains active ester groups which can be replaced with preS1 domain of hepatitis B surface antigen (HBsAg) to generate specificity for liver cells.

In our study, poly-(MPC-co-BMA-co-NPMA) (PMBN) was used as a framework for specific drug delivery. PMBN forms a polymeric aggregate in water through a hydrophobic interaction and can combine with hydrophobic proteins such as PTX. Therefore, it was hypothesized that PMBN with hydrophobic monomer units would be able to solubilize hydrophobic drugs and possibly enhance their water solubility. Moreover, because the active ester group of the PMBN was replaced by preS1, it was predicted that PMBN conjugated with preS1 would bind selectively to liver cells. We aimed to deliver PTX to the liver specifically using the PMBN-preS1 conjugate and analyzed, in vitro and in vivo, the antitumor effects against human hepatoma cells (HepG2) of PTX, PTX with PMBN and PTX with PMBN-preS1 conjugate. In addition, we examined any possible side effects which PMBN and PMBN-preS1 conjugate might cause.


BMA, n-butyl methacrylate; EDTA, ethylenediaminetetra-acetic acid; HBsAg, hepatitis B surface antigen; MPC 2-methacryloyloxyethyl phosphorylcholine; MTT, 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyl-2H tetrazolium bromide; NPMA, p-nitrophenylcarbonyloxyethyl methacrylate; PDGFR, platelet-derived growth factor receptor; PMBN, MPC-co-BMA-co-NPMA; PTX, paclitaxel; VEGFR2, vascular endothelial growth factor receptor 2.

Material and methods


PTX was obtained from Wako Pure Chemical Industries (Tokyo, Japan). Recombinant HBsAg (pre-S1 antigen, 108 amino acids, 13.5 kDa) was obtained from Biodesign International (Saco, ME). Fetal bovine serum (FBS) was obtained from JRH Bioscience (Lenexa, KS). Penicillin and streptomycin were obtained from Biochrom (Cambridge, UK). Amicon ultra centrifugal filter devices were obtained from Millipore (Bedford, MA). d5-PTX was obtained from Moravek Biochemicals (Brea, CA). Six and 96 multiwell plates were purchased from Sumitomo Bakelite (Tokyo, Japan). Tissue culture flasks (250-ml) were purchased from Falcon-Becton Dickinson (Heidelberg, Germany).

All other reagents not described above were purchased from Sigma (St. Louis, MO).

Cell lines

The HepG2 cell line, derived from a human HCC, was purchased from the American Type Culture Collection, Rockville, MD, and the A431 cell line, derived from a human squamous cell carcinoma, was purchased from Riken, Saitama, Japan. These cells were seeded in 250-ml tissue culture flasks at 37°C under 5% CO2 in Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% (vol/vol) heat-inactivated FBS, penicillin (100 U/ml), streptomycin (100 U/ml) and L-glutamine (4 mM). After 7 days of culture (cultured to about 80% confluence), the cells were harvested with 0.05% trypsin/0.02% EDTA and suspended for the subsequent experiments. Cell viability was determined by exclusion of 0.1% trypan blue greater than 95% at the beginning of each experiment.


Female, 6 to 8-week-old BALB/c athymic nude mice (CLEA Japan, Tokyo, Japan) weighing 18–20 g were used. The animals were maintained under specific pathogen-free conditions with constant 12-hr light/dark cycles. Food and bedding were sterilized and water was acidified. The animal study was approved by the Committee on the Use of Live Animals in Teaching and Research of Keio University.

Preparation of HBsAg conjugated MPC polymer and PTX

PMBN was prepared by a conventional radical polymerization of the corresponding monomers.16, 17 The composition of PMBN was ∼40 mol % MPC, 50 mol % BMA and 10 mol % NPMA, and its molecular weight was about 5.1 × 104 determined by gel permeation chromatography with polyethylene oxide standards. The chemical structure is shown in Figure 1.

Figure 1.

Chemical structure of PMBN. PMBN was constructed from ∼40 mol % MPC, 50 mol % BMA and 10 mol % NPMA, with a molecular weight of around 47,000 and molecular size of around 200–250 nm.

PMBN-preS1 conjugate was prepared as follows. Ten milligrams of PMBN and 800 μg of preS1 were allowed to dissolve in 10 ml DMEM with 10% FBS (pH 7.8), with mixing by continuous gentle rotation for 48 hr at 4°C and p-nitrophenol was removed after the reaction between the NPMA unit in the PMBN and preS1. The replacement reaction was complete when all the ester units of the NPMA had been replaced by preS1, which was determined by the absorption spectra of p-nitrophenol (wavelength: 400 nm). PMBN-preS1 conjugate was purified with Amicon Ultracentrifugal filter devices were used for molecular weight 30,000 at 5,000g for 15 min, which removed free preS1 and p-nitrophenol from the PMBN-preS1 conjugate.

PTX was dissolved in dimethyl sulfoxide (DMSO) for a stock solution, which was then diluted with the desired volume of DMEM. The final concentration of DMSO was kept less than 0.1% in all experiments.

Cell growth inhibition assay

Viable HepG2 or A431 (1 × 104) cells per well were seeded with 200 μl complete medium onto a 96-well microtiter plate and incubated for 24 hr until >50% confluent. The medium was replaced with 100 μL of fresh medium containing various concentrations (100, 101, 102, 103, 104 and 105 nM) of PTX, with or without 100 μL PMBN-preS1 conjugate or PMBN (in total of 200 μL each). PMBN and PMBN-preS1 conjugate concentrations were adjusted to 10−4, 10−3, 10−2, 10−1, 100 and 101 mg/ml, according to PTX concentration (10−4 mg/ml for 100 nM of PTX) and the plates were incubated continuously under the same conditions. After 48 hr for HepG2 and 24 hr for A431, the medium was discarded and the plates were washed twice with phosphate buffered saline. Cell viability was measured using a classical MTT calorimetric assay, as described previously.18 Briefly, the medium was removed at the end of the incubation and 0.5 mg/ml of MTT (3-(4,5-dimethylthiazol-2yl)-2,5-diphenyl-2H tetrazolium bromide) was added to each plate. After 4 hr incubation at 37°C in 5% CO2, the medium was discarded and 100 μl of DMSO was added to each well and the plates were incubated for 30 min at room temperature. A 96-well microtiter plate reader (BMG LabTechnologies, Germany) was used to determine the optical density (OD) at 570 nm with background subtraction at OD690. Cell survival was expressed as a percentage of control untreated cells. The 50% inhibitory concentration (IC50) was calculated from a dose–response curve derived by nonlinear algorithms from at least 5 experimental points. Within each experiment, determinations were performed in triplicate and experiments were repeated at least 3 times for reproducibility.

In vivo hepatoma growth inhibition assay

Suspensions of 1 × 107 HepG2 cells or 1 × 105 A431 cells were inoculated subcutaneously into the right flank of the nude mice to establish the xenografts. The lengths (A, mm) and widths (B, mm) of the tumors were measured twice weekly with a Vernier caliper and the body weights were measured once a week. Treatment was started when the xenograft length reached at least 5 mm and the mice were randomly assigned to treatment and control groups of 5 mice per group. The mice were then treated by intraperitoneal (i.p.) injection of PTX or PTX/PMBN or PTX/PMBN-preS1 and saline (as control), which were adjusted to a total of 200 μg of PTX in 400 μl DMEM with 10% FBS per mouse, on days 1 to 5 consecutively. Tumor volume (TV, mm3) was calculated according to the following formula:

equation image

Tumor growth rate (TGR) was calculated using the following formula:

equation image

where TVx is the TV on day x and TV1 is the TV on day 1. The animals were killed on day 21 by deep anesthesia and body and tumor weight were measured and recorded. To validate the accuracy of TV measurements, the correlation of actual tumor weight at the end of the treatment with estimated TV was assessed. A correlation coefficient of 0.92 (p < 0.001) was obtained. Assuming 1 unit as the specific gravity of tumor tissue, the body weights of the animals were estimated by the formula: Final net body weight (g) = body weight at the end (g) − estimated final tumor weight (g).

Mice in which the total 2-dimensional size of tumors had not grown beyond 40 mm were killed humanely via CO2 inhalation after 6 weeks. Within each experiment, determinations were performed in triplicate and experiments were repeated at least twice for reproducibility.

Tissue distribution of PTX in BALB/c mice

PTX alone or PTX/PMBN-preS1 were administered intraperitoneally to the xenografted nude mice. After administration, the mice were successively killed by deep anesthesia with diethyl ether at 4, 8 and 24 hr (3 mice each at each time point). The tumors and livers were excised immediately and washed with 10 mM Na2HPO4 buffer and homogenized. PTX concentrations in tumor and liver were determined using high-performance liquid chromatography (HPLC).19, 20 Briefly, frozen standards and samples were allowed to thaw at room temperature, homogenized further by vortex-mixing, and a volume of 500 μl of each sample was aliquotted into a microcentrifuge tube. Volumes of 100 μl of d5-PTX as the internal standard and 5 ml of an extraction solution (1-chlorobutane) were added to each tube. Tubes were then vortex-mixed for 20 sec, followed by centrifugation for 5 min at 3,000 rpm. Next, the organic layer was collected in a glass tube and evaporated under a gentle stream of nitrogen. The residue was redissolved in 200 μl mobile phase [25 mM potassium dihydrogen phosphate (KH2PO4), pH 3.5 adjusted with 85% phosphoric acid (H3PO4), and acetonitrile (80:20, vol/vol)]. Finally, a volume of 50 μl of this solution was injected into the HPLC system.

Adverse effect on vital organs and histological study

The animals (n = 5 per group) were killed on day 21 by deep anesthesia. Plasma was obtained to measure aspartate aminotransferase (AST), alanine aminotransferase (ALT), blood urea nitrogen (BUN) and creatinine (Cr), and the tumor, liver, kidneys, spleen, pancreas and lungs were surgically resected. Tissue samples were fixed in 10% phosphate-buffered formalin (pH 7.0) and embedded in paraffin. The blocks were cut into 5 μm sections and stained with hematoxylin and eosin using standard techniques. The histological sections were observed under a light microscope with various magnifications and micrographs under microscope were taken for record and comparison.

Statistical analysis

The experiment data were presented as mean ± standard deviation. Unpaired Student's t-test was used for statistical analysis between 2 groups. The IC50 values were estimated using nonlinear algorithms, which enabled IC50 values to be derived with 95% confidence intervals. Analyses were carried out using the following software: Statview 5.01 (SAS Institute, NC) statistical package, Excel 2003 (Microsoft, Redmond, WA) and GraphPad Prism 5.0 (GraphPad Software, San Diego, CA). Statistical significance was defined at p < 0.05 (2-sided).


Inhibition of proliferation of HepG2 and A431 cells by PTX/PMBN-preS1 in vitro

To explore the possibility of enhancing the cytotoxic effect of PMBN-preS1, we examined initially in vitro a combination treatment of PMBN-preS1 and the chemotoxic agent PTX. Inhibition of cell growth by PTX, with or without PMBN-preS1, was evaluated using the MTT assay. PTX showed a dose-dependent growth-inhibitory effect on HepG2 cells. The incorporation of PMBN-preS1 with PTX led to a significant increase in the growth-inhibitory effect. The IC50 values of PTX, PTX/PMBN and PTX/PMBN-preS1 against HepG2 cells were 1,008, 890 and 131 nM, respectively (Table I). The IC50 of PTX with PMBN against HepG2 cells was slightly lower than for the unconjugated drugs alone, but there were no statistically significant differences. The conjugation of preS1 to PMBN enhanced strongly the synergistic inhibitory effect of PTX on the human hepatocellular cell line and the IC50 of PTX was reduced markedly to about one-eighth of that of PTX alone (p < 0.05).

Table I. IC50 Values for Inhibition of Proliferation of HepG2 and A431 Cells
IC50 (nM)95% CI (nM)p valueIC50 (nM)95% CI (nM)p value
  • The IC50 values were estimated using nonlinear algorithms, which enabled IC50 values to be derived with 95% confidence intervals (95% CI).

  • 1

    p values were calculated from 2 dose–response curves by nonlinear algorithms of PTX and PTX/PMBN.

  • 2

    p values were calculated from 2 dose–response curves by nonlinear algorithms of PTX and PTX/PMBN-preS1.

PTX1,008201.6–5,040 13850.74–374.3 

On the other hand, the IC50 values of PTX, PTX/PMBN and PTX/PMBN-preS1 against A431 were 138, 277 and 175 nM, respectively (Table I). The effects observed on the HepG2 cell line were not apparent on the A431 cell line. PTX alone had a strong cytotoxic effect on A431 in vitro, and the addition of the MPC polymers as a transfectant showed no increased cytotoxic effect.

As controls, PMBN and PMBN-preS1 without drug loading did not show any significant cytotoxicity when used at doses close to the IC50 values of PTX/PMBN and PTX/PMBN-preS1 (data not shown).

Hepatoma growth inhibition by PTX/PMBN-preS1 in vivo

The in vitro anti-proliferation activity of PTX combined with PMDN-preS1 was evaluated further using human tumor models xenografted in athymic mice. HepG2 xenografts appeared in ∼70% of the nu/nu mice between 14 and 20 days after subcutaneous injection, whereas A431 xenografts appeared in ∼90% between 5 and 12 days after injection. TGR was suppressed significantly with addition of PMBN and even further with addition of PMBN-preS1. In particular, almost complete growth inhibition was achieved with administration of PTX/PMBN-preS1. The TGR for this group was −96.2% (p < 0.05 versus PTX) (Fig. 2a). The range of TGR in this group was from −91.3 to 100% and complete response (disappearance of the tumor pathologically) was shown in 3 out of 5 mice.

Figure 2.

Hepatoma growth inhibition by PTX/PMBN-preS1 in vivo on (a) HepG2 and (b) A431: (—□—, open rectangle) PTX; (—○—, open circle) PTX/PMBN; (— • —, filled circle) PTX/PMBN-preS1. Statistical analysis was performed with the Student's t-test, *p < 0.05 versus PTX.

On the other hand, no statistical difference in TGR for both PTX/PMBN and PTX/PMBN-preS1 was shown versus PTX alone in the A431 inoculated mice (Fig. 2b).

This implied that the outcome of PMBN's drug-carrying effect is more evident on HepG2 xenografts than A431 xenografts and PMBN-preS1 is only targeted to human liver cell lines, as seen in vitro.

Tissue distribution of PTX in BALB/c mice

The distribution of PTX in the tumor and liver was analyzed using HPLC. The PTX levels in subcutaneous hepatoma were increased significantly at 4 hr by PMBN-preS1, compared to PTX alone (PTX 131 ± 78 ng/g, PTX/PMBN-preS1 1,105 ± 608 ng/g) (Table II). However, the change in PTX concentration was only moderate in the A431 xenograft models at 4 hr (PTX 386 ± 57 ng/g, PTX/PMBN-preS1 712 ± 229 ng/g). The PTX levels in tumors 4 hr after administration of PMBN-preS1 were 8-fold higher than that of PTX alone in the HepG2 group but only 1.8-fold higher in the A431 group. The PTX levels in the murine liver remained slightly higher with PMBN-preS1 delivery for both groups (data not shown).

Table II. Tissue Concentration of Paclitaxel in HepG2 or A431 Xenografts
Time (hr)Concentration (ng/g)
  • nd, not detected. Values are given in mean ± SD of samples from 3 mice at each time.

  • 1

    p < 0.05 versus PTX group.

  • 2

    No statistical significant difference was shown between PTX and PTX/PMBN-preS1.

4131 ± 781,105 ± 6081386 ± 57713 ± 2292
840 ± 31268 ± 971153 ± 46204 ± 602
24ndnd6 ± 531 ± 24

Adverse effect on vital organs and histological study

No significant changes in serum AST, ALT, BUN and Cr were found in the groups of HepG2- and A431-bearing mice, but AST was slightly elevated uniformly in the groups to which PTX had been administered (Table III). Body weight loss was significantly lower in the PTX/PMBN-preS1 group than the control group and the final net body weight was clearly higher in the PTX/PMBN-preS1 treated HepG2-bearing mice (Table IV). Body weight tended to increase in all the groups in A431-bearing mice due to rapid growth of cancerous tissue (Table IV). No signs of morbidity were observed in either group throughout the experiment period, which could be considered to reflect the anti-cancer activity of PTX, with PMBN taking precedence over the toxicity to athymic mice.

Table III. Serum Bun/Cr and AST/ALT Levels After 21 Days
  1. BUN, blood urea nitrogen; Cr, creatinine; AST, asparate aminotransferase; ALT, alanine aminotransferase. Values are given in mean ± SD.

BUN31.4 ± 4.2 (mg/dL)36.4 ± 6.329.5 ± 3.934.6 ± 3.722.4 ± 4.0
Cr0.08 ± 0.01 (mg/dL)0.11 ± 0.010.08 ± 0.010.11 ± 0.010.09 ± 0.00
AST76 ± 22 (IU/L)194 ± 102112 ± 24307 ± 138101 ± 5
ALT32 ± 2 (IU/L)7 ± 58 ± 13 ± 215 ± 2
Table IV. Effects of Paclitaxel With/Without PMBN-Related Carriers on (a) HepG2-Bearing Nude Mice and (b) A431-Bearing Nude Mice Body Weight, Tumor Volume and Tumor Weight
GroupsBody weight (g)BW change (g)Final tumor volume (cm3)Final net BW (g)Tumor growth rate (%)
  • BW, body weight; BW change = (BW at start) − (BW at the end); final net BW = (BW at the end) − (estimated final tumor weight); tumor growth rate = [(final tumor volume) − (initial tumor volume)]/(initial tumor volume) × 100. Values are given in mean ± SD.

  • *

    p < 0.05 versus control;

  • **

    p < 0.05 versus PTX.

Control (saline)20.8 ± 0.116.7 ± 0.5−4.060.563 ± 0.07216.1 ± 0.5800 ± 229
PTX21.2 ± 1.118.7 ± 0.6−2.480.155 ± 0.07918.6 ± 0.797.9 ± 25.5
PTX/PMBN20.0 ± 0.518.6 ± 0.9−1.420.012 ± 0.00518.5 ± 0.9−74.3 ± 7.5
PTX/PMBN-preS119.9 ± 0.720.0 ± 1.9+0.03*0.004 ± 0.00419.9 ± 1.9*−96.2 ± 3.6**
Control (saline)21.2 ± 0.526.3 ± 0.7+5.105.59 ± 1.3222.1 ± 1.54,160 ± 712
PTX20.0 ± 0.823.4 ± 0.4+3.421.48 ± 0.3622.0 ± 0.51,135 ± 334
PTX/PMBN21.7 ± 1.123.2 ± 1.0+1.58*0.83 ± 0.6822.4 ± 1.386.3 ± 53.0
PTX/PMBN-preS121.3 ± 0.422.5 ± 0.3+1.23*0.32 ± 0.1222.1 ± 0.3120 ± 153

To assess the histological nonspecific toxicity caused to several organs by PMBN-preS1, mice were killed and samples of liver, kidneys, spleen, pancreas and lungs obtained, and HE-stained sections prepared. All of the histological evaluations revealed unremarkable toxicity. A few megakaryocytes, probably derived from the spleen via the splenic vein or remnant of embryonic blood production in the fetal liver, were observed close to the zone 3 area, but no parenchymal change was seen at the zone 2 area (Fig. 3a). No zonal necrosis was observed at all in any xenografts in the absence of PTX (Fig. 3b). Zonal necrosis was sporadically seen at the outer side of HepG2 just inside of the fibrous capsule after administration of PTX (Fig. 3c). Zonal necrosis with apparent hemorrhagic changes was evident and distributed widely after administration of PTX with PMBN-preS1 conjugate (Fig. 3d).

Figure 3.

(a) Liver: A few megakaryocytes might be observed close to the zone 3 area but no parenchymal change was seen at the zone 2 area (hematoxylin–eosin, original magnification ×400). (b) HepG2 nontreated: A typical, moderately differentiated HCC structure was noted. A compact trabecular growth pattern with slightly basophilic cytoplasm and sinusoidal vasculature was maintained and no zonal necrosis was shown (hematoxylin–eosin, original magnification ×200). (c) HepG2 treated with PTX: Zonal necrosis was sporadically seen at the outer side of HCC just inside of the fibrous capsule (hematoxylin–eosin, original magnification ×200). (d) HepG2 treated with PTX/PMBN-preS1: Zonal necrosis accompanying with clustering erythrocytes was spread out and distributed widely (hematoxylin–eosin, original magnification ×200).


These results show that PMBN can dissolve PTX readily and greatly increase hepatotropism by HBsAg. The cytotoxicity of PTX with PMBN-preS1 conjugate in vitro was ∼8 times higher than PTX alone (Table I). Furthermore, the in vivo growth inhibition reached nearly 100% for PTX delivered by PMBN-preS1 conjugate (Fig. 2a).

Particle size is the critical property that governs the tissue distribution process, such as passage through the capillaries (about 5 μm) or through the fenestrae between discontinuous endothelial cells (30–500 nm). The internalization of particles into cells by adsorptive endocytosis or receptor-mediated endocytosis is also a size-limited process.21 The size of the micelle formed by PMBN with PTX was about 50 nm, which is small enough (<200 nm) to bypass filtration by interendothelial cell slits and accumulate effectively in tumors via the enhanced permeability and retention effect.22 The actual internalization of particles inside the micelle formed by PMBN into cells had been proved previously by the specific expression of green fluorescent protein (GFP) in HepG2 cells from a GFP-encoded plasmid enclosed within the PMBN micelle.22

Another important factor that determines the stability of the micelle is its molecular weight. PMBN weighed ∼50 kDa in a previous experiment.23 Micelles should be sufficiently large to avoid rapid renal excretion (>50 kDa), but PMBN will be conjugated with preS1 domain of HBsAg and dissolve with PTX, resulting in a slightly higher weight; also the shape flexibility may assist in overcoming the weight shortage.

These superior “passive targeting” properties may be increased by additional “active targeting” through conjugation of preS1. The passive targeting vector (PMBN) was combined with an active targeting component (preS1), which make it feasible to achieve effective drug delivery to the liver. An MPC unit was copolymerized with an immobilizable unit, p-NPMA, and the NPMA moiety was replaced by preS1. This reaction was confirmed by the absorption spectrum of p-nitrophenol (wavelength 400 nm) and excess free preS1 was removed by dialysis. Recombinant HBsAg (Biodesign International, USA) is derived from yeast (Saccharomyces cerevisiae) and used widely as a reference standard for immunological studies because of its stable antigenicity.24, 25 HBsAg contains small (S), medium (M; S + pre-S2) and large (L; S + pre-S2 + pre-S1) proteins. The pre-S1 peptide, the N-terminal amino acid residues 108–119 of the L protein, functions as the specific ligand that binds to specific receptors on human hepatocytes, and this enables “receptor-mediated endocytosis” of the PMBN/preS1 conjugate. Unlike the S and M proteins, attempts to efficiently synthesize L proteins and assemble them into L protein particles in various eukaryotic cells have been unsuccessful, probably because of the presence of the pre-S1 peptide with an unknown function which appears to be inhibitory to the host secretory apparatus.26 In this study, recombinant Pre-S1 antigen derived from Escherichia coli was used to deliver PTX specifically to hepatocytes. We reported previously that micelle formation of PMBN was not disturbed, despite any additional conjugate attached to the polymer.27 There are only a few reports using preS1 as an active targeting component to hepatocytes and all of these are viral vector with preS1 genetic sequence inserted.28, 29 This is the first report of nonviral vector which contains preS1 domain as an active targeting component to hepatocytes.

The tissue specificity of PMBN could be altered in other ways by replacing the hepatocyte-specific moieties with other biorecognition molecules (e.g., ligands, receptors and antibodies) without modifying the assembly of PMBN. In a previous publication, we reported that tissue specificity can be switched from hepatocytes (HepG2) to squamous cells (A431) by changing the ligand from preS1 to epidermal growth factor (EGF).30 Hence, the precise opposite result would be expected if EGF was conjugated to PMBN, instead of preS1.

It is known widely that HepG2 cells are relatively less susceptible to PTX in clinically achievable concentrations compared to other hepatoma cell lines, such as Hep3B or N1S1.10 Because HepG2 cells were derived from a less aggressive hepatoblastoma with the capacity to express a variety of hepatic genes, including the cytochrome P450 family, they can metabolize PTX.9 However, the chemoresistance of HepG2 cells to PTX was overcome readily by the polymer-HBsAg conjugate delivery system, and this result proved the importance of drug delivery efficacy for anti-cancer toxicity. The PTX-resistant nature of HepG2 cells was taken into account and the dosage of PTX was set slightly higher (350 mg/m2 for 5 consecutive days) than that for clinical practice use in humans (210 mg/m2). It is well known that animals commonly used in preclinical drug studies (i.e. mice, rats, rabbits, monkeys and dogs) do not eliminate drugs at the same rate as humans.4 Small mammals usually eliminate drugs faster than large mammals and the dose regimens used for mice usually are higher than for humans (usually 15 times the general human dose).4 However, a linear extrapolation is not possible for dose estimation. Regarding to the dose of PTX (350 mg/m2) used for mice in our study, we referred to previous experiments in which a 3-hr PTX infusion at 135 mg/m2 yielded a maximum plasma concentration of 2.5 μM. The IC50 of PTX against HepG2 was about 1 to 4.3 μM, so that a dosage (350 mg/m2) should be necessary for HepG2-bearing mice.

The antitumor effect of PTX depends upon sustained therapeutic concentrations of the drug rather than maximal plasma concentrations, because PTX-toxicity requires entry of cells into the M phase. Sustained drug concentrations are achieved with prolonged PTX infusions9 and various alternative drug delivery systems instead of Cremophor®, such as albumin-bound nanoparticles,31 microspherical polymers,32 cholesterol-rich nanoemulsions,33 or gelatin nanoparticles34 have been introduced. All these particles, including our PMBN, are intended to solubilize PTX to increase its bioavailability and to let the drug remain in the systemic circulation long enough to provide gradual accumulation in the required area. In parallel, our in vitro study showed a greater than 24-hr requirement for PTX-induced tumor cell death. However, because continuous administration to nude mice is impractical technically, daily consecutive administration was chosen. In our model, fractionated daily administration showed a significantly more distinct antitumor activity than a single transient administration (data not shown). A daily addition of PTX might have the advantage of minor or less severe side effects, and consequently such a therapeutic regimen might be tolerated better. The ideal pathway of administration should be the intravenous (i.v.) route for clinical settings, but, for nude mice, i.v. administration is restricted due to the likelihood of caudal venous thrombosis that prevents further injections. More large animal experiments are required to ascertain the efficacy of the PMBN complex via i.v. administration.

Currently, a considerable proportion of the world population has been immunized against HBsAg with yeast-derived hepatitis B virus (HBV) vaccines. Conventional HBV vaccines are produced by recombinant DNA technique mainly composed of S protein.35 However, around 10% of adults have no response to current commercial vaccines, and 5–10% show only a low response.35 Neurath et al.36 have shown that T cell immune response to preS antigens (preS1 and preS2) in nonresponder mice could circumvent the genetic nonresponsiveness to the S antigen. Therefore, “a third-generation vaccine” containing preS1 and preS2 antigens may enhance the immunological therapeutic efficacy. Various preS containing vaccines have been introduced and evaluated, but none of them are currently sufficient for clinical use.37–39 If this third-generation vaccine was successfully developed in near future, our PMBN-preS1 conjugate could be unfortunately neutralized readily by vaccine-induced antibodies. In addition, 22–23% of HBV patients exhibit detectable levels of antibodies with specificity for preS1 region.40 So there will be a need for some manipulation of the preS1 component to avoid this immunity, such as by a point mutation at the S protein seen in HBV escape mutant.41

There are some limitations in this study. First, the antitumor effect was determined by an MTT assay in vitro and tumor size in vivo, both of which measure a degree of tumor necrosis rather than apoptosis. PTX binds microtubules and inhibits cell replication by inducing arrest of mitosis. Mitotic arrest of PTX-treated cells has been considered to cause cell apoptosis.10 Therefore, the precise measurement of PTX-induced cell death should be determined by measuring the apoptosis index rather than tumor necrosis. Second, the problem of particle size, and particle charge of the PMDN-preS1 conjugate, might be determined for this specific experiment. It remains doubtful in this specific delivery mechanism as a nanoparticle. Thirdly, it is thought that further experiments are necessary to evaluate chemosensitivity, pharmacokinetics and biodistribution of PTX-PMBN conjugate in large animals, especially in humans.

In summary, the PMBN-preS1 conjugate showed a substantially enhanced capacity to carry PTX without any adverse side effects and could well provide the basis of a human hepatocyte-specific drug delivery system in clinical settings.


The authors thank Mr. Toshihide Muramatsu, Ms. Yuki Inaba, Mr. Manabu Nakano and Ms. Yuko Uchida for their excellent technical support.