Development of a novel systemic gene delivery system for cancer therapy with a tumor-specific cleavable PEG-lipid

Abstract

For successful cancer gene therapy via intravenous (i.v.) administration, it is essential to optimize the stability of carriers in the systemic circulation and the cellular association after the accumulation of the carrier in tumor tissue. However, a dilemma exists regarding the use of poly(ethylene glycol) (PEG), which is useful for conferring stability in the systemic circulation, but is undesirable for the cellular uptake and the following processes. We report the development of a PEG-peptide-lipid ternary conjugate (PEG-Peptide-DOPE conjugate (PPD)). In this strategy, the PEG is removed from the carriers via cleavage by a matrix metalloproteinase (MMP), which is specifically expressed in tumor tissues. An in vitro study revealed that the PPD-modified gene carrier (Multifunctional Envelope-type Nano Device: MEND) exhibited pDNA expression activity that was dependent on the MMP expression level in the host cells. In vivo studies further revealed that the PPD was potent in stabilizing MEND in the systemic circulation and facilitating tumor accumulation. Moreover, the i.v. administration of PPD or PEG/PPD dually-modified MEND resulted in the stimulation of pDNA expression in tumor tissue, as compared with a conventional PEG-modified MEND. Thus, MEND modified with PPD is a promising device, which has the potential to make in vivo cancer gene therapy achievable.

Introduction

Gene therapy is a promising technology for intractable diseases including cancer. Non-viral gene carriers that are applicable via the systemic circulation are especially desired to expand the clinical applications of gene vectors to treat a metastatic cancer. Recently, great efforts have been made toward a breakthrough in the establishment of systemically administrable in vivo gene carriers for tumor treatment.^{1, 2}

For successful systemic gene carrier development, it is essential to optimize both the pharmacokinetics and intracellular trafficking. To achieve this purpose, a Multifunctional Envelope-type Nano Device (MEND) has been developed, in which the core of a plasmid DNA, condensed by a polycation, is encapsulated by a lipid envelope. This structure is advantageous to achieve the a concept of Programmed Packaging.³ In this design, devices to regulate endosomal escape and nuclear delivery can be modified on the lipid envelope and polycation, respectively, thus allowing them to function at the correct time and at the appropriate location.^{3, 4} As a result, the MEND exhibited high pDNA expression in vitro, owing to the control of intracellular trafficking.^{3, 5, 6}

However, its rapid elimination from the systemic circulation, owing to the high rate of clearance by the reticuloendothelial system (RES) or other organs, prevents the use of the MEND in in vivo applications for tumors via intravenous (i.v.) administration, as also found with lipoplexes.⁷ It is generally accepted that extended circulation can be achieved by the surface modification of carriers with PEG.^{8, 9} Moreover, limiting the size of the carriers to less than 200 nm enables them to efficiently accumulate in tumor tissue, by an enhanced permeability and retention (EPR) effect.^{10, 11} In contrast once they are taken up by the tumor tissue, the PEG inhibits interactions between the gene carriers and the tumor cells, which results in a significant loss in transfection activity.¹² Thus, the use of PEG represents a dilemma, in that the PEG is useful for controlling the pharmacokinetics of the carriers, but is undesirable for the cellular association of gene carriers with tumors. To resolve this dilemma, various devices have been developed. Ambegia et al.¹³ developed a strategy of exchangeable PEG-lipids. In the present study, we propose alternative strategy, involving a novel gene delivery system for cancer gene therapy, using a MEND modified with a tumor – specifically cleavable PEG-lipid. A cleavable PEG-lipid triggered in a specific manner would be desirable. Most of the current cleavable PEG devices have been designed to be cleaved in response to the intracellular microenvironment, such as low pH or a specific enzyme in the endosome/lysosome and the reductive conditions in the cytoplasm.^{14, 15, 16, 17, 18, 19}

Here, we report the development of a novel gene delivery system for cancer gene therapy, using a MEND modified with an enzymatically cleavable PEG-lipid. The cleavable PEG-lipid is composed of a PEG/matrix metalloproteinase (MMP)-substrate peptide²⁰/1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) ternary conjugate (PPD), which is specifically cleaved by MMP in the extracellular space in tumor tissues. MMP is expressed in high levels in tumor cells, and is secreted into the extracellular space.²¹ MMP is involved in the angiogenesis, invasion and metastasis of malignant tumors via its ability to degrade the extracellular matrix (ECM). MMP transcript levels are low in normal cells, but are induced at high levels in tumor cells.²² A schematic diagram (Figure 1) shows the strategy for overcoming the dilemma associated with the use of PEG. By modifying the MEND with PPD (PPD-MEND), systemic circulation is prolonged as a result of the presence of PEG, and the MEND then can accumulate in tumor tissue by the EPR effect. After extravasation from capillaries in the tumor tissue, the PPD is cleaved by the secreted MMP, and the PEG dissociates from the MEND. Finally, the PEG-uncoated MEND can interact efficiently with the tumor cells, resulting in high transfection activity. To address the utility of this strategy, the enzymatic cleavage of PPD and the transfection activity of PPE-MEND with in vitro and in vivo tumors were evaluated along with the pharmacokinetics.

Figure 1

A schematic diagram illustrating the strategy used to resolve the dilemma associated with the use of PEG. (i) By modifying the gene carrier with PPD, the systemic circulation is prolonged, and the accumulation in tumor is increased by the EPR effect. (ii) After extravasation from capillaries in the tumor tissue, the PPD is cleaved by an extracellular MMP secreted from tumor cells. (iii) PEG dissociates from the gene carrier, and the naked carrier can then associate efficiently with the tumor cell surface.

Results

Synthesis of PPD

The PPD (2) was synthesized as described in the synthetic diagram (Figure 2a). A functional PEG, bearing an activated ester group at the ω-end (MeO-PEG-CONHS: M_n=2192, M_w/M_n=1.02), was synthesized as described previously.²³ The N-terminal region of the peptide was reacted with MeO-PEG-CONHS (1 eq.) to yield the PEG-Peptide conjugate (1). The PEG-Peptide conjugate gave a unimodal peak at a high molecular position in the matrix assisted laser desorption/ionization-time of flight mass spectroscopy (MALDI-TOF MS) spectrum (obsd. M_n=3268, M_w/M_n=1.01, calcd. M_n=3256), as compared to the MeO-PEG-CONHS. To obtain the PPD, the PEG-Peptide was conjugated with DOPE (2 eq.) by activating the terminal carboxylic group of the PEG-Peptide conjugate using N-hydroxysuccinimide (NHS) and N,N″-dicyclohexylcarbodiimide (DCC). After the conjugation reaction, the resulting PPD was purified by size exclusion chromatography (SEC). The PPD was located at a high molecular position in the MALDI-TOF MS spectrum (obsd. M_n=3955, M_w/M_n=1.01, calcd. M_n=3950), as compared to the PEG-Peptide conjugate, indicating an increased molecular weight due to the formation of the PPD. In the ¹H NMR spectrum (Figure 2b), the peak corresponding to a double bond residue, assignable to the unsaturated fatty acid of DOPE, was clearly observed at δ=5.4 ppm, along with tyrosyl peaks at δ =6.7 and 7.1 ppm and a methoxy peak at δ =3.4 ppm. The integral ratios between the double bond peak, the tyrosyl peaks and the methoxy peak confirmed that the DOPE was quantitatively introduced at the C-terminus of the PEG-Peptide conjugate.

Figure 2

Synthetic route to PPD (2) and ¹H NMR spectrum (a) (i) Et₃N(2 eq.), DMSO, 3 h, Y. 95%; (ii) NHS(2 eq.), DCC(2 eq.), Et₃N(2 eq.), CHCl₃, 24 h, Y. 40%. (b) ¹H nuclear magnetic resonance (NMR) spectrum of the purified PPD(2).

PPD cleavage in response to MMP on the liposomes

We then investigated the cleavage of PPD on the surface of liposomes. It was previously shown that DOPE forms a hexagonal (H_II) phase and readily aggregates under physiological conditions.²⁴ However, when the PEG-conjugated lipid was included in the lipid component with DOPE, a stabilized liposomal structure was formed in aqueous media. Similarly, stable liposomes with diameters of around 150 nm were prepared by hydrating a lipid film composed of DOPE and PPD. These liposomes were incubated with no treatment, bovine serum albumin (BSA) or MMP-2 at several concentrations (Figure 3a). After 24 h, a significant increase in size (>2000 nm) was observed, when the PPD-modified liposomes were treated with MMP-2 at 14 and 56 nM. This result indicated that the PPD is cleaved in an MMP-2-dependent manner, leading the liposomes to an H_II transition. In contrast, an increase in size was not observed with any treatment of the control liposomes composed of DOPE and conventional PEG-lipid (Figure 3b). The cleavage of PPD in response to MMP-2 was also confirmed by a MALDI-TOF MS analysis. After the incubation with MMP-2, the peaks representing PPD were detected at a lower position (M_n=3953–1980). Collectively, these results confirm that the PPD is cleavable in an MMP-2-dependent manner, even when its substrate peptide is shielded by the PEG layer.

Figure 3

Time profiles for the size of DOPE liposomes modified with PPD or conventional PEG-lipid after incubation with MMP-2. The liposomes were modified with PPD (a) and conventional PEG-lipid (b). PPD-liposomes (a) and PEG-liposomes (b) were treated with MMP-2 at several concentrations or with BSA (56 nM), and the diameters were measured at 3, 6 and 24 h by DLS analysis.

Physical properties of MENDs

The impact of the MMP-dependent cleavage of PPD for use in a gene delivery system was then addressed. We recently established the MEND, in which a condensed DNA/polycation complex is coated with a lipid bilayer.³ In the present study, luciferase-encoding plasmid DNA (pDNA) was condensed with protamine as a nuclear targeting device,^{25, 26} and then was encapsulated with a lipid bilayer composed of DOPE:1,2-dioleoyl-3-trimethylammonium-propane (DOTAP):Cholesterol. The physical characteristics of the prepared MENDs are shown in Table 1 and Figure 4. The PEG-MEND and PPD-MEND displayed comparable sizes, ζ-potentials and polydispersity indexes by a dynamic light scattering (DLS) measurement, suggesting that particles with similar morphology had been prepared (Table 1). We also confirmed that the non-PEG-modified MEND (MEND) and the PEG-MEND existed as a single particle population (Figure 4a). Therefore, heterogeneous aggregates or lipopolyplexes were probably not formed. The encapsulation of pDNA in a lipid envelope was supported by the ζ-potential data. The ζ-potential of the core particle prepared with protamine and pDNA was −26 mV. After the hydration and sonication of the lipid film, the ζ-potential of the particles was inverted to +54 mV. The inversion of the ζ-potential indicated that the core particle was wrapped with the lipid film, as demonstrated previously.³ More direct evidence of encapsulation of pDNA, but not of lipopolyplex formation, was obtained with electron micrographs. After the purification of MEND and PEG-MEND on a discontinuous sucrose density gradient, freeze-fracture electron micrographs were taken (Figure 4b). These micrographs revealed that the MEND and PEG-MEND were both spherical, with a diameter of approximately 200 nm, which is consistent with the diameters measured by DLS measurements.

Table 1 Physical properties of MEND, PEG-MEND, PPD-MEND

Figure 4

Characterization of MEND and PEG-MEND (a) The size distributions of MEND and PEG-MEND were obtained by a histogram analysis of the DLS measurements. (b) The MEND was separated from the empty vesicles by discontinuous sucrose density gradient, and then was visualized by freeze-fracture electron microscopy. The bars indicate 200 nm.

In vitro transfection study

In an in vitro study, HT1080 (fibrosarcoma cells) and HEK293 (human embryonic kidney cells) were used as model cells, in which MMP is highly and poorly expressed, respectively (Figure 5a).^{27, 28} The MEND exhibited high transfection activity in both cell types (Figure 5b and c). When the MEND was modified with 5 and 15% of conventional PEG-lipid (5 and 15% PEG-MEND), the transfection activities were decreased to less than 1%, as compared to the MEND in both cell types. The transfection activity of 5% PPD-MEND was enhanced by 35-fold, as compared to that of PEG-MEND at the same PEG density in HT1080 cells (Figure 5b). It should be noted that the transfection activity of the PPD-MEND was similar to that of MEND. In contrast, the transfection activity of 5% PPD-MEND was less than 5% of that of MEND in HEK293 cells (Figure 5c). The transfection activity of 15% PPD-MEND was comparable to that of 15% PEG-MEND, presumably because of the effect of steric hindrance. To obtain a certain evidence that the PPD-MEND was activated by soluble MMP, the transfection activity was also evaluated with three additional cell types (Supplemental data, Figure 1). In the MMP-positive cell lines (MG-63 and HOS), the transfection activity of 5% PPD-MEND was more than 10 times higher than that of 5% PEG-MEND, whereas the enhancement was less prominent in MMP-negative cells (PC-3). These data support the conclusion that PPD-MEND was activated in response to the MMP expression. Furthermore, the transfection activity of 5% PPD-MEND with the MMP-negative HEK293 cells was increased more than 14-fold, as compared to the 5% PEG-MEND, by a co-incubation with recombinant MMP-2 or conditioned medium from HT1080 cells (Supplemental data, Figure 2). These results also suggested that the soluble MMP secreted from HT1080 cells activated the PPD-MEND. Taking these observations into consideration, the transfection activity is closely related to the cellular expression level of MMP and the PEG density on the surface of the MEND.

Figure 5

The in vitro transfection activities of PPD- or a conventional PEG-lipid modified MEND. (a) The expression level of MMP-2 in the supernatant of HT1080 and HEK293 cells was evaluated by an ELISA. (b, c) Luciferase activities of MEND, PEG-MEND and PPD-MEND in HT1080 cells (b) and in HEK293 cells (c) were evaluated at 48 h after transfection. Luciferase activities are expressed as relative light units (RLU) per mg of protein. A value of 10⁶ RLU represents the luciferase activity of 5 ng luciferase protein. Each bar represents the mean±s.d., n=3. ^*P<0.05, ^**P<0.01.

In vivo pharmacokinetics study of MENDs

The ability of PPD to improve the stability in the systemic circulation and the tumor accumulation of MENDs was confirmed after i.v. administration. To measure the elimination profiles from the systemic circulation, the lipid bilayer of MEND was labeled with [³H]cholesteryl hexadecyl ether (CHE), and the MEND was administered to mice via the tail vein.²⁹ The half-lives of the blood concentrations were comparable between PEG-MEND and PPD-MEND (Figure 6a). As a result, the calculated area under the curve (AUC) values of 5% PEG-MEND, 15% PEG-MEND, 5% PPD-MEND and 15% PPD-MEND were 19.6-, 37.2-, 8.6- and 14.6-times higher than that of MEND (Supplemental data, Table 1). Therefore, PPD is a potent asset in the elongation of the systemic circulation of MEND, although it was less potent than PEG. To investigate the tumor accumulation of MEND, MEND labeled with [³H]CHE was injected into tumor-bearing mice via the tail vain. The tumor accumulations of PPD-MEND were enhanced, as compared to those of MEND (Figure 6b). These results show that the PPD on MEND potently prolongs its systemic circulation, presumably by preventing entrapment by RES or other organs, which results in the enhancement of the tumor accumulation of MEND.

Figure 6

Stability in systemic circulation, the tumor distribution and transfection activities of MENDs in vivo. (a) Blood concentrations of [³H]CHE-labeled MEND were evaluated at 1, 6 and 24 h after i.v. injection. Data represent as the %ID per milliliter of blood. (b) Tumor distribution of MEND at 24 h in tumor-bearing mice, expressed as the %ID per gram tumor. (c) Luciferase activity of MEND at 48 h, expressed as RLU per gram tumor. (d) Specific activity of the MEND in the tumor. A value of 10⁶ RLU represents the luciferase activity of 5 ng luciferase protein. Each bar represents the mean±s.d., n=3. ^**P<0.01.

In vivo transfection study in tumor tissue

The in vivo luciferase activity of the tumor was evaluated at 48 h after the i.v. administration of MENDs at a dose of 25 μg pDNA/mouse. The transfection activity (Figure 6c) was barely detectable with the MEND. The tranfection activities of 15% PEG-MEND and 15% PPD-MEND were comparable. However, it is noteworthy that the tumor accumulation of 15% PEG-MEND was approximately four times higher than that of 15% PPD-MEND. Therefore, the specific activity of the particles of 15% PPD-MEND, which is denoted as the in vivo transfection activity normalized by the amount of tumor-delivered particles (relative light units (RLU)/%ID), was threefold higher than that of 15% PEG-MEND (Figure 6d). Moreover, a comparison between 15% PPD-MEND and 5% PEG-MEND, which displayed comparable systemic circulation and tumor accumulation, revealed that 15% PPD-MEND exhibited 55-fold higher transfection activity of particles than 5% PEG-MEND. The calculated specific activity of the particles of 15% PPD-MEND was 50 times higher than that of 5% PEG-MEND. Therefore, 15% PPD-MEND exhibited more prominent activity in delivering the pDNA to the nucleus after it reached the tumor tissue. These results show that PPD has the ability to stabilize MEND, even in the systemic circulation, and is advantageous for the in vivo transfection by its cleavable nature in response to MMP in tumor tissue.

Further improvement of the in vivo transfection activity in the tumor by the combination of PEG and PPD

The poor stability of PPD-MEND in the blood circulation was compensated by an additional modification of conventional PEG (Figure 7). A one to one mixture of PEG and PPD was attached to MEND (PEG/PPD-MEND), in which the total content of PEG and PPD was adjusted to 5, 15 and 20%. For the PEG-MEND, the modification with 15% PEG was sufficient to achieve maximum stability in the systemic circulation and tumor accumulation (Figure 7a and b). In contrast, in the case of PEG/PPD-MEND, the blood concentration and the tumor accumulation monotonically increased, depending on the total PEG/PPD content. As a result, comparable tumor accumulation was achieved when the MEND was modified with 20% of PEG or PEG/PPD (1:1 mole ratio). In this situation, the PEG/PPD-MEND exhibited more than 65 times higher transfection activity, as compared with the PEG-MEND (Figure 7c). Moreover, the specific activity of the PEG/PPD-MEND was more than 90 times higher than that of the PEG-MEND (Figure 7d). This result indicated that the combining PEG with the PPD-MEND is an intelligent strategy to improve both the pharmacokinetics and in vivo transfection activity.

Figure 7

Further evaluation of the transfection activities of MENDs in the in vivo tumor. The MEND was modified with PEG and PPD (1:1 mole ratio) at 5, 15 and 20% total PEG densities. (a, b) Blood concentration and tumor distribution of MENDs at 24 h in tumor-bearing mice, represented as the %ID per milliliter of blood and %ID per gram of tumor. (c) Luciferase activity of MENDs at 48 h, expressed as RLU per gram of tumor. (d) Specific activity of the MENDs in the tumor. A value of 10⁶ RLU represents the luciferase activity of 5 ng luciferase protein. Each bar represents the mean±s.d., n=3. ^**P<0.01.

Discussion

Many studies have sought suitable non-viral carrier for cancer gene delivery. To improve the stability in the systemic circulation and the tumor accumulation, various groups have devised non-viral gene carriers with PEG.^{13, 14, 15, 16, 17, 18, 19, 30, 31, 32} Recently, cleavable PEG systems were developed, which are cleaved in response to an intracellular environment (e.g. low pH in endosome/lysosomes, thiolytically small molecules, etc.^{14, 15, 16, 17, 18, 19}) and an extracellular environment³³ (e.g. low pH in tumor tissue³⁴) to enhance endosomal escape and cellular association, respectively. Therefore, a strategy for triggering the specific cleavage of PEG on the exterior of cancer cells should be an alternate option for improving tumor selectivity.

In the present study, we propose a novel strategy for controlling the triggered release of PEG for gene carriers in response to enzymatic cleavage by MMP, which is specifically secreted from tumor cells (Figure 1). To realize this strategy, a peptide, which is cleaved by MMP, was used as a linker between the PEG and the lipid.

The cleavage of PPD in response to MMP-2 was confirmed by MALDI-TOF MS and DLS analyses (Figure 3). In the DLS analysis, 10% of the PPD was modified, as liposomes composed of DOPE and PPD could not stably form below this density. Under these conditions, aggregation was observed when the liposomes were incubated with more than 14 nM of MMP-2. It is difficult to estimate the MMP concentration in the extracellular space of an in vivo tumor. However, it can be roughly calculated from the in vitro studies. In this calculation, we considered two factors derived from the in vivo and in vitro experimental conditions. One is the difference in the expression level of MMP-2. The other is the difference in the extracellular space. The expression of MMP-2 in the in vivo tumor and the in vitro cultured cells was evaluated as approximately 0.4 and 4 ng/mg protein, respectively, by means of an enzyme-linked immunosorbent assay (ELISA). Concerning the extracellular space, 4 × 10⁴ cells were commonly cultured in 500 μl of medium. Therefore, the secreted MMP was diluted in a large volume of medium. In contrast, in the case of the in vivo tumor, 1 g of tissue consisted of 1 × 10⁹ cells, which are surrounded by approximately 200 μl of extracellular space.³⁵ Therefore, the extracellular space per cell is 2.5 × 10⁴ times higher in the in vitro culture cells. In other words, the secreted MMP in vivo is much more effectively concentrated (∼2.5 × 10⁴-fold), as compared with the in vitro cell culture system. Concerning the latter factor, the MMP-2 concentration in the in vivo tumor and the in vitro-cultured cells were evaluated as approximately 0.4 and 4 ng/mg protein, respectively. Therefore, even the though cellular expression level of MMP in the tumor tissue was about 1/10 of that of the in vitro culture system, the extracellular concentration of MMP in the in vivo tumor is calculated to be ∼2.5 × 10³ times higher than that of in vitro cell culture system. As the concentration of MMP-2 in the cell culture medium was 0.4 nM (Figure 5a), the extracellular concentration in the in vivo tumor is estimated as approximately 1 mM. Therefore, in the in vivo situation, the concentration of MMP-2 should be sufficient to cleave the PPD.

We examined the availability of PPD for cancer gene delivery, as compared to PEG-MEND. The particle size of the MEND was controlled to the optimal size for an EPR effect (100–200 nm) by modification with PEG or PPD (Table 1). In vitro transfection revealed that the PPD-MEND exhibited high transfection activity in the MMP-positive HT1080 cells, even in medium containing serum (Figure 5b). Furthermore, the transfection activity with the MMP-negative HEK293 cells was increased by supplementation with HT1080 conditioned medium and recombinant MMP-2, indicating that the MMP secreted by the HT1080 cells was responsible for the cleavage of PPD. However, in the HT1080-conditioned medium, the MMP-2 was present at a concentration of approximately 30 ng/ml (0.4 nM), which is 1/30 of the threshold concentration of recombinant MMP-2 required to cleave the PPD, as determined by the light scattering experiment (14 nM) (Figure 3). This discrepancy can be accounted for by the following explanation: first, the linker peptide was designed to be cleavable by other types of MMP (e.g. MMP-7 and -9) along with MMP-2. Although recombinant MMP-2 was used in the light-scattering experiments, HT1080 cells may secrete various types of MMPs,²⁸ which would be synergistically responsible for the cleavage of PPD. Second, the concentration of MMP-2 in the conditioned medium was determined after the medium in the well was mixed. It is possible that the PPD was effectively cleaved in the close proximity of the cellular surface, where the secreted MMP was locally concentrated. Finally, the PPD density on the liposomes was 10% (Figure 3) in the DLS analysis, whereas 5% PPD-MEND was used in the in vitro transfection experiment (Figure 5b); therefore, the MMP may be more easily accessible to the substrate peptide under the PEG shield.

The modification with PPD improved the stability in the systemic circulation and the tumor accumulation of MEND, whereas the potential of PPD was slightly inferior to that of the conventional PEG-lipid combination (Figure 6a and b, Supplemental data, Table 1). To investigate the mechanism concerning the reduced tumor accumulation of PPE-MEND, the fractions of PPD-MEND and PEG-MEND bound to the blood cells and free in the plasma were compared. The fraction of 5% PEG-MEND free in the plasma was 4.2 times higher than that of 5% PPD-MEND. Similarly, the fraction of 15% PEG-MEND free in the plasma was 5.0 times higher than that of 15% PPD-MEND (Supplemental data, Table 2). Therefore, PPD-MEND can bind to the blood cells, and this may contribute to its poor tumor accumulation (Figure 6b). However, the PPD modification was advantageous for the gene delivery for the in vivo tumor. In the tumor, the 15% PPD-MEND exhibited 100-fold higher transfection activity than the 5% PEG-MEND, even though the tumor accumulations of both MENDs were almost equivalent. Furthermore, PPD-MEND exhibited more robust activity in delivering the pDNA to the nucleus after it reached the tumor tissue.

Overall, the in vivo transfection activity is inconsistent with the in vitro studies. PPD-MEND was only slightly activated under the in vitro conditions when a high density of PPD was modified, whereas it was significantly activated in the in vivo tumor. This discrepancy may be accounted for by the different concentrations of MMP in the extracellular space of cultured cells in vitro and in vivo tumor cells. As mentioned above, the concentration of MMP in the in vivo tumor was calculated to be higher than that of the in vitro medium of the cultured cells (approximately 1 μ M vs. 0.4 nM). Therefore, the PPD may be cleaved by the MMP secreted by the tumor, even when the PPD was highly modified.

Finally, further trials to improve the in vivo transfection activity were performed. As mentioned above, the stability in the systemic circulation and the subsequent tumor accumulation of PEG-MEND were higher than those of PPD-MEND (Figure 6a and b). In the in vitro situation, the PEG modification just perturbs the transfection activity of MEND. Therefore, all of the PEG should be exchanged to PPD to achieve maximum in vitro transfection activity. In contrast, in the in vivo situation, the pharmacokinetics (e.g. the stability in blood circulation and the tumor accumulation) and the specific activity of the particle should be balanced to maximize the in vivo transfection activity.

Based on this concept, the poor stability of PPD-MEND in the blood circulation was compensated by the additional modification of conventional PEG (Figure 7), whereas 15% of PEG density generated maximum stability in the systemic circulation and the tumor accumulation with PEG-MEND, those of PEG/PPD-MEND increased depending on the total PEG/PPD content. The 20% PEG/PPD-MEND combination exhibited the highest transfection activity in tumor tissue, among all of the combination of MENDs. The results suggest that the balance between the improvement of the pharmacokinetics of MEND and the PPD cleavage in tumor tissue seemed to be better than other conditions. Therefore, the additional modification of PPD-MEND with PEG is an intelligent strategy to compensate for the poor stability of PPD in the systemic circulation. Further optimization concerning the effects of the ratio of PEG and PPD and the length of the PEG chain is ongoing in our laboratory.

Collectively, the use of PPD may provide a valuable function for in vivo cancer gene delivery, by virtue of its ability to be cleaved in response to an MMP-rich environment.

Materials and methods

Materials

DOPE, DOTAP, Cholesterol, PEG-DSPE and PEG-DOPE were purchased from AVANTI Polar Lipids (Alabaster, AL, USA). [³H]CHE was purchased from Perkin–Elmer Life Sciences, Japan (Tokyo, Japan). MMP-2 was obtained from SIGMA-Aldrich (St Louis, MO, USA). Protamine sulfate salmon milt was purchased from Merck KGaA (Dramstadt, Germany). The MMP cleavable peptide (>70% purity, obsd. M_w=1178.9, sequence: IndexTermGGGVPLSLYSGGGG) was obtained from Thermo Electron GmbH (Ulm, Germany). The pcDNA 3.1 (+)-luc plasmid (7037 bp) was constructed by inserting the firefly luciferase gene (HindIII–XbaI fragment) of the pGL3-Control plasmid (Promega, Madison, WI, USA) into the pcDNA 3.1 (+) plasmid (Invitrogen, Breda, Netherlands).

Polymer analysis

SEC, in 10 mM Tris-HCl (pH 7.4), was performed with a Model1321H1 (GILSON) instrument equipped with a UV/VIS-155 detector and a Superdex 75 10/100 GL column. MALDI-TOF MS spectroscopy was performed on a Bruker MALDI-TOF-MS Reflex II. ¹H NMR (400 MHz) spectra were obtained in CD₃OD with a JOEL EX400 spectrometer.

Synthesis of PEG-Peptide conjugate (1)

The peptide (24.6 mg, 20.4 μmol) and MeO-PEG-CONHS (41.2 mg, 20.4 μmol, 1 eq.) were dissolved in dimethyl sulfoxide (DMSO) (1 ml) in the presence of Et₃N (5.9 μl, 41 μmol, 2 eq.). The reaction mixture was stirred for 3 h at room temperature. The crude conjugate was purified by dialysis against distilled, deionized water (MW cutoff 1000), and then was freeze-dried to give the PEG-Peptide conjugate (1) (66.8 mg, 95% yield). MALDI-TOF MS: obsd. M_n=3268, M_w/M_n=1.01 (calcd. M_n=3256); ¹H NMR (CD₃OD): δ=3.38 (s, 3H, CH₃O), 3.72 (s, 195H, PEG-backbone), 6.84 (d, J=8 Hz, 2H, aromatic ring of Tyr), 7.16 (d, J=8 Hz, 2H, aromatic ring of Tyr). The ¹H NMR was incompletely assigned because of the complex signals of the 14 peptide residues.

Synthesis of PPD (2)

A solution of DOPE (24.7 mg, 33.2 μmol, 2 eq.) in CHCl₃ (1 ml) was added to a mixture of PEG-Peptide conjugate (1) (54.7 mg, 16.6 μmol), NHS (3.8 mg, 33.2 μmol, 2 eq.) and DCC (6.9 mg, 33.2 μmol, 2 eq.) in CHCl₃ (3 ml). The reaction mixture was stirred for 24 h at room temperature. The crude PEG-Peptide-DOPE conjugate (PPD) (2) was isolated after evaporation of the solvent. The reaction mixture, including PPD, was loaded on a size exclusion chromatographic column, which was eluted with 10 mM Tris-HCl (pH 7.4). The pure PPD was obtained in 40% yield. MALDI-TOF MS: obsd. M_n=3955, M_w/M_n=1.01 (calcd. M_n=3950); ¹H NMR (CD₃OD): δ=3.4 (s, 3H, CH₃O-), 3.7 (s, 195H, PEG-backbone), 5.4 (m, 4H, double bond of DOPE), 6.7 (d, J=8 Hz, 2H, aromatic ring of Tyr), 7.1 (d, J=8 Hz, 2H, aromatic ring of Tyr). The ¹H NMR was incompletely assigned because of the complex signals of the 14 peptide residues and the DOPE.

Diameter, ζ-potential and polydispersity analysis

The diameter, the ζ-potential and the polydispersity index of the liposomes and the MEND were determined by DLS (ELS-8000, Otsuka Electronics, Japan).

Preparation of liposomes

Liposomes were prepared by the hydration method. A lipid film was prepared by the evaporation of a chloroform solution of lipids (125 nmol DOPE and 12.5 nmol PPD or PEG-DOPE). Then, a solution of 20 mM HEPES-buffered saline (HBS), 140 mM NaCl and 2 mM CaCl₂ (pH 7.4) was added to the lipid film, which was incubated for 10 min at room temperature to hydrate the lipid. To obtain the liposomes, the lipid film was sonicated for about 1 min in a bath-type sonicator (125 W, Branson Ultrasonics, Danbury, CT, USA).

Evaluation of the cleavage of the synthesized PPD by MMP on the surface of liposomes by DLS analysis

The liposomes (0.5 mM), containing 10% PEG or PPD, were incubated with no treatment, BSA (56 nM) or MMP-2 (2, 7, 14 and 56 nM) at 37°C. The concentration of recombinant MMP-2 was measured by an MMP-2 ELISA system. The diameter of the liposomes was determined by DLS, at 3, 6 and 24 h after incubation.

Cell culture

HT1080 and HEK293 cells were grown in DMEM, supplemented with 10% fetal bovine serum, penicillin (10000 units per ml) and streptomycin (10 mg per ml), at 37°C in a humidified 5% CO₂ atmosphere.

Preparation of MENDs

The MEND was prepared as described previously.³ Briefly, pDNA (0.1 mg per ml) was condensed with protamine (0.1 mg per ml), at a nitrogen/phosphate (N/P) ratio of 1.0, in 250 μl of HBS. The average diameter, the ζ-potential and the polydispersity index of the condensed DNA were 92 nm, −26 mV and 0.12, respectively. A lipid film was formed by the evaporation of a chloroform solution of 125 nmol total lipids, composed of DOTAP, DOPE and Cholesterol (3:4:3 mole ratio). The condensed DNA was applied to the lipid film, which was incubated for 10 min at room temperature to hydrate the lipids. For the preparation of the MEND modified with PEG-lipid or PPD, a lipid film was prepared by evaporation with a certain amount of PEG-lipid or PPD. To coat the condensed DNA with the lipid, the lipid film was then sonicated for about 1 min in a bath-type sonicator.

Purification of MEND by sucrose density gradient centrifugation

The MEND, containing rhodamine-labeled DNA (10% of the total DNA) and N-(7-nitro-2,1,3-benzoxadiazol-4-yl) (NBD)-labeled DOPE (1 mol% of the total lipids), was layered on a discontinuous sucrose density gradient (0–40%), and was centrifuged at 160 000 g for 2 h at 20°C. Aliquots (1 ml) were collected from the top, and the fluorescence intensities of rhodamine and NBD in each fraction were measured. The fraction containing both the rhodamine and NBD fluorescence peaks was used as the purified MEND.

Freeze-fracture electron microscopy

The MEND, purified by sucrose density gradient centrifugation, was rapidly frozen at −196°C, and fractured at −120°C with a Freeze replica apparatus (FR-7000B, Hitachi, Japan). An electric discharge was applied to deposit Pt/C, then C, on the surface at angles of 45° and 90°. The replicas were mounted on 300-mesh Ni grids and were observed with an electron microscope (JEM200 CX, JEOL, Tokyo, Japan).³

Evaluation of MMP-2 expression by ELISA

The expression of MMP-2 in the supernatants of HT1080 and HEK293 cells and in the in vivo tumor tissues was evaluated by the MMP-2, Human, Biotrak ELISA System (Amersham Biosciences, Uppsala, Sweden). Samples were prepared according to the ELISA System protocol.

The in vitro luciferase reporter assay

To examine the in vitro transfection activity of MEND, 4 × 10⁴ cells were seeded in 24-well dishes 1 day before transfection. The MEND, corresponding to 0.4 μg of DNA, was applied to 0.25 ml of DMEM-containing serum, followed by an incubation at 37°C for 3 h. Then, 0.75 ml of DMEM was added to the cells, which were incubated further for 45 h. The cells were then lysed in Reporter Lysis Buffer (Promega, Madison, WI, USA) and the luciferase activity was assayed using the Luciferase Assay System (Promega). The protein concentration was determined with a BCA Protein Assay Kit (Pierce, Rockford, IL, USA). Luciferase activities are expressed as RLU per mg of protein. A value of 10⁶ RLU represents the luciferase activity of 5 ng luciferase protein.

Evaluation of the stability of MENDs in the systemic circulation

The lipid bilayers of MEND, PEG-MEND and PPD-MEND were labeled with [³H]CHE, and then the MENDs were administered to male DDY mice (5–6 weeks old, CLEA Japan, Tokyo, Japan) via the tail vein, at a dose of 12.5 μg of DNA (0.1 μmol of lipids). At 1, 6 and 24 h post-injection, the radioactivity in the blood was measured as described previously.³⁶ Briefly, at the indicated times, the mice were killed and the blood was collected. A 100 μl sample of blood was solubilized in 1 ml of Soluene-350 (Perkin–Elmer Life Sciences) for 1 h at 40°C, and then was decolorized by H₂O₂. The radioactivity was determined by liquid scintillation counting, after adding 10 ml of scintillation fluid. The blood concentration is represented as the % injected dose (%ID) per ml of blood.

Preparation of tumor-bearing mice

Tumor-bearing mice were prepared by the subcutaneous inoculation of 10⁶ HT1080 cells to male BALB/c nude mice (5–6 weeks old, CLEA Japan).

Evaluation of tumor accumulation of MENDs

The MENDs, labeled with [³H]CHE, were injected via the tail vein into tumor-bearing mice, with tumor sizes of 12–18 mm in diameter, at a dose of 12.5 μg of DNA (0.1 μmol of lipids). After 24 h, the tumor tissues were collected and then the radioactivity was determined as described previously.³⁶ Tumor accumulation is represented as the %ID per gram of tumor.

The in vivo luciferase reporter assay

When the tumors reached sizes of 12–18 mm in diameter, the MENDs were injected via the tail vein, at a dose of 25 μg of DNA (0.2 μmol of lipids). At 48 h post-injection, the mice were killed and the tumor tissues were collected. Tumor tissues were homogenized in Reporter Lysis Buffer (1 ml), using a POLYTRON homogenizer (KINEMATICA, Littan, Switzerland), followed by centrifugation at 13 000 r.p.m. for 5 min at 4°C to obtain the supernatant. The luciferase activity in the supernatant (20 μl) was assayed using the Luciferase Assay System. Luciferase activities are expressed as RLU per gram of tumor. A value of 10⁶ RLU represents the luciferase activity of 5 ng luciferase protein.

Statistical analysis

Comparisons between multiple treatments were made with the analysis of one-way analysis of variance (ANOVA), followed by the Student–Newman–Keuls test. Comparisons between treatments were made using a two-tail Student's t-test. A P-value of <0.05 was considered as a significant difference.