Rational design of composition and activity correlations for pH-sensitive and glutathione-reactive polymer therapeutics

Journal of Controlled Release 101 (2005) 47 – 58 www.elsevier.com/locate/jconrel

Rational design of composition and activity correlations for pH-sensitive and glutathione-reactive polymer therapeutics Mohamed E.H. El-Sayed, Allan S. Hoffman*, Patrick S. Stayton* University of Washington, Department of Bioengineering, Box: 352255, AERL 338, Seattle, WA 98195-2255, United States Received 15 June 2004; accepted 26 August 2004

Abstract Limited cytoplasmic delivery of enzyme-susceptible drugs remains a significant challenge facing the development of protein and nucleic acid therapies that act in intracellular compartments. bSmartQ pH-sensitive, membrane-destabilizing polymers present an attractive approach to shuttle therapeutic molecules past the endosomal membrane and into the cytoplasm of targeted cells. This report describes the use of a new functionalized monomer, pyridyl disulfide acrylate (PDSA), to develop pHsensitive, membrane-destabilizing, and glutathione-reactive polymers by copolymerization with several pH-sensitive and hydrophobic monomers. The activity of the carriers is described as a function of (a) the influence of increasing the length of the hydrophobic alkyl group substituted onto the pH-sensitive monomer and (b) of the effect of incorporating a hydrophobic monomer such as butyl acrylate (BA) on the pH sensitivity and membrane-destabilizing activity of new polymer compositions. The membrane-destabilizing activity of different polymer compositions was evaluated as a function of pH and polymer concentration using the red blood cells (RBC) hemolysis assay. Hemolysis results show that the increase in the hydrophobic character of polymer backbone results in a shift in the pH sensitivity profile and an increase in the membrane-destabilizing activity. Results show that the observed hemolytic activities and pH sensitivity profiles could be designed across a range that matches the properties needed for drug carriers to enhance the cytoplasmic delivery of therapeutic cargos. D 2004 Elsevier B.V. All rights reserved. Keywords: Smart polymers; Intracellular delivery; Polymer therapeutics; Antisense delivery

1. Introduction * Corresponding authors. Patrick S. Stayton is to be contacted at University of Washington, Department of Bioengineering, Box: 351721, Bagley Hall 421, Seattle, WA 98195-2255, United States. Tel.: +1 206 685 8148; fax: +1 206 685 8526. Allan S. Hoffman, Tel.: +1 206 543 9423; fax: +1 206 543 6124. E-mail addresses: [email protected] (A.S. Hoffman)8 [email protected] (P.S. Stayton). 0168-3659/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jconrel.2004.08.032

Recent advances in drug design have led to the development of several classes of therapeutic macromolecules including peptides, proteins, plasmid DNA (pDNA), silencing RNA, antisense oligodeoxynucleotides (asODN), and immunotoxins. Activity of such macromolecules depends on their ability to reach the

48

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

cytoplasm of targeted cells. However, cellular uptake of macromolecules typically occurs by passive or receptor-mediated endocytosis. Internalized macromolecules subsequently accumulate in the endosomal compartment, which ultimately evolves into the lysosomal compartment where degradation of the entrapped molecules and loss of therapeutic activity occurs. It is apparent that limited cytoplasmic delivery of enzyme-susceptible drugs remains a major challenge to therapeutic approaches such as gene- and antisense-based therapies. Viral vectors proved efficient in the cytoplasmic delivery of pDNA [1], which is a result of the pHdependent membrane-destabilizing activity of fusogenic proteins present on the viral coat [2]. Hemagglutinin of the influenza virus is an example of such fusogenic proteins [2]. Hemagglutinin and similar fusogenic proteins acquire an ionized, hydrophilic conformation at physiologic pH values, which becomes hydrophobic and membrane-destabilizing in response to the acidic pH gradients of the endosome [2–5]. This results in rupture of the endosomal membrane and escape of endosomal contents into the cytoplasm [2–5]. Several endosomal-destabilizing peptides have been designed to mimic natural fusogenic protein sequences and have been shown to increase cytoplasmic gene delivery using polycationic carriers [6]. Despite the endosomolytic activity of such peptides, potential toxicity and immunogenicity limit their clinical utility. Our group has described the use of bsmartQ pHsensitive, membrane-destabilizing polymers to enhance the cytoplasmic delivery of therapeutic macromolecules [7]. These bsmartQ polymers are characterized by their unique ability to bsenseQ the changes in environment pH where they undergo a change from a hydrophilic, stealth-like conformation at physiologic pH to a hydrophobic and membranedestabilizing one in response to endosomal pH gradients. Thomas et al. reported the pH-dependent disruption of synthetic lipid vesicles using poly(ethylacrylic acid) (PEAA) in acidic environments of pH 6.3 or lower [8,9]. PEAA has been characterized as an additive to enhance gene delivery in cationic lipid formulations through its pH-dependent membranedestabilizing activity [10]. Our group developed several copolymers of acrylic acid and alkyl acrylates that exhibited pH-dependent

membrane-destabilizing activities, which increased with the increase in the proportion of the alkyl acrylates in polymer composition [11]. In addition, our group has characterized several alkyl acrylic acid polymers including poly(propylacrylic acid) (PPAA), which showed a pH-dependent membrane-destabilizing activity that is one order of magnitude higher than that of PEAA [11,12]. Furthermore, PPAA increased the transfection efficiency of pDNA in cationic lipid formulations both in vitro [13] and in vivo [14]. PPAA was also shown to increase the endosomal escape and enhance the cytoplasmic delivery of model antibodytargeted protein complexes in Jurkat T-cells [15]. We recently developed a new glutathione-reactive monomer, pyridyl disulfide acrylate (PDSA), which was copolymerized with methylacrylic acid (MAA) and butyl acrylate (BA) monomers to develop a new pH-sensitive, membrane-destabilizing, and glutathione-reactive polymer [16]. The research reported here extends our previous work to examine the influence of modifying the composition of PDSAcontaining polymers on their pH-sensitive membranedestabilizing activity. A series of new polymer compositions was prepared by copolymerizing the PDSA monomer with different pH-sensitive monomers including methylacrylic acid (MAA), ethylacrylic acid (EAA), and propylacrylic acid (PAA). The second series of PDSA polymers was prepared by incorporating the hydrophobic BA monomers in the polymer backbone. The relationship between polymer compositions and their pH-dependent membranedestabilizing activities was evaluated in a standard red blood cells (RBCs) hemolysis assay. Systematic variations of polymer composition will identify key compositional features essential for pH sensitivity and membrane-destabilizing activity of PDSA polymers.

2. Materials and methods 2.1. Materials Diethyl ethyl malonate, diethyl propyl malonate, MAA, BA, 2,2V-Azo-bis(isobutyronitrile), and buffer salts were purchased from Sigma-Aldrich (Milwaukee, WI). Both MAA and BA monomers were distilled prior to use. 2,2V-Azo-bis(isobutyronitrile) (AIBN) was recrystallized from methanol prior to use. All NMR

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

solvents were purchased from Cambridge Isotope Laboratories (Andover, MA). The peptide sequence [(Cys)-(Gly)3-(Lys)6] was custom synthesized by United Biochemicals. (Seattle, WA). The phosphorothioate sequence (5V-CCCCCCGGCCATGGCTGC– 3V, M W 5699 Da) was designed as a specific asODN for the IL-1 receptor-associated kinase-1 (IRAK-1) gene and was synthesized by Qiagen (Valencia, CA).

49

2.2. Methods

mer were pooled together and were concentrated by rotary evaporation at room temperature. The total yield of PDSA monomer was approximately 42% w/ w. The 1H-NMR spectrum of pure PDSA monomer in deuterated dimethylsulfoxide (DMSO-d6) solvent showed the following peaks: 2.2 ppm (pentet, 2H), 2.9 ppm (triplet, 2H), 4.3 ppm (triplet, 2H), 5.9 ppm (doublet, 1H), 6.2 ppm (quartet, 1H), 6.5 ppm (doublet, 1H), 7.1 ppm (multiplet, 1H), 7.7 ppm (multiplet, 2H), and 8.5 ppm (multiplet, 1H).

2.2.1. Synthesis of ethylacrylic acid (EAA), propylacrylic acid (PAA), and pyridyl disulfide acrylate (PDSA) monomers Both EAA and PAA monomers were synthesized following the method described elsewhere starting from diethyl ethyl malonate and diethyl propyl malonate, respectively [17]. The syntheses and purity of EAA and PAA monomers were verified by examining the 1H-NMR spectra of the products on a Bruker 499 MHz spectrometer (Billerica, MA). The PDSA monomer was synthesized following the method reported earlier (Fig. 1A) [16]. Briefly, PDSA monomer synthesis started by the synthesis of hydroxypropyl pyridyldisulfide (HPPDS) following the procedure described elsewhere [18]. A total of 1.56 mmol of HPPDS was placed in a dry roundbottom flask followed by 10 ml of THF directly distilled from sodium metal, under nitrogen, into the flask. The flask was kept in an ice bath (2 8C) and a total of 2.9 mmol of TEA was added to the reaction mixture before capping the flask. A total of 3.1 mmol of acryloyl chloride was added drop wise to the chilled reaction mixture while stirring. The reaction flask was filled with nitrogen and the reaction was allowed to proceed for 6 h at 2 8C then for an additional 18 h at room temperature. Formation of PDSA monomer was examined by TLC using a 35:65 ethyl acetate/hexane solvent mixture as a mobile phase (R f =0.62). The reaction mixture was filtered and the filtrate was collected and concentrated by rotary evaporation at room temperature. The dry product was purified by fractionation on a 60-g silica gel column using a 35:65 ethyl acetate/hexane solvent mixture as a mobile phase. Column fractions were tested for their PDSA content by TLC using the same mobile phase composition described above. Column fractions containing high yield of pure PDSA mono-

2.2.2. Synthesis of PDSA-containing polymers The first series of PDSA-containing polymers was prepared by free radical polymerization of PDSA monomers with different pH-sensitive monomers including MAA, EAA, and PAA using AIBN as an initiator. The molar feed ratio of PDSA and pH-sensitive monomers was adjusted to 5% and 95%, respectively. The molar ratio of AIBN used in the polymerization reactions of MAA and EAA with PDSA monomers was 2 mol% of total monomer concentration and 1 mol% for the polymerization reaction of PAA with PDSA monomers. A total of 10% v/v DMF solvent was added to the polymerization reactions of MAA and EAA with PDSA monomers but the polymerization reaction of PAA with PDSA monomers was carried out in bulk. The second series of PDSA-containing polymers incorporated the hydrophobic BA monomer with the PDSA and pH-sensitive monomers utilized in the first series. The molar feed ratio of PDSA, BA, and pHsensitive monomers was adjusted to 5%, 25%, and 70%, respectively. The molar ratio of AIBN used in the polymerization reactions of MAA/EAA with BA and PDSA monomers was 2 mol% of total monomer concentration and 1 mol% for the polymerization reaction of PAA with BA and PDSA monomers. A total of 10% v/v DMF solvent was added to the polymerization reactions of MAA/EAA with BA and PDSA monomers but the polymerization reaction of PAA with BA and PDSA monomers was carried out in bulk. In a typical polymerization reaction, the calculated amounts of monomers, AIBN, and DMF were mixed in a flask. The reaction flask was then freeze-thawed under vacuum for three consecutive cycles for removal of air from the reaction mixture. The reaction

50

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

Fig. 1. The synthesis schemes for: (A) PDSA monomer, and (B) poly(RCOOH-co-BA-co-PDSA) polymers.

flask was then transferred to an oil bath at 65 8C to start the polymerization reaction, which was allowed to proceed for 6–8 h (Fig. 1B). Methanol was used to

dissolve the formed polymers followed by drop wise addition of the polymer solution to approximately 1 l of diethyl ether for precipitation of the pure polymer

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

and removal of unreacted monomers, initiator, and short polymer chains. The precipitated polymer was filtered under vacuum, washed twice with diethyl ether, and dried under vacuum at room temperature for 18–24 h. The collected polymer was dissolved in methanol and precipitated in diethyl ether twice for a total of three precipitations/filtrations to obtain the pure polymer, which was dried under vacuum at room temperature for 48 h. Polymer yields were in the range of 30–40% w/w. 1 H-NMR spectroscopy was used to confirm the purity of the synthesized polymers and to examine their compositions. Characteristic 1H-NMR peaks of the synthesized polymers in DMSO-d6 solvent are: 4.0 ppm (2H, O-CH2-R), 7.2 ppm (1H, aromatic), 8.5 ppm (1H, aromatic), and 12.4 ppm (1H, COOH). The characteristic peaks of each monomer including the COOH peak of the pH-sensitive monomer, the aromatic peaks of the PDSA monomer, and the OCH2-R peak of BA and PDSA monomers were used to determine the exact composition of each of the new polymers. PDSA content in each polymer was further confirmed by reacting a known amount of the polymer with excess dithiothreitol (DTT) in 100 mM phosphate buffer of pH 7.4 while monitoring the UV absorbance of released pyridine-2-thione at 343 nm (e 343=8.08103 M 1 cm 1) [19]. Molecular weight of the synthesized polymers was determined by gel permeation chromatography using Waters UltrahydrogelR 1000 and 250 columns connected in series to a Waters GPC system with a Waters 510 HPLC pump, a Waters 410 differential refractometer, and a Spectra Physics SP4290 integrator. A 100 mM phosphate buffer solution with pH 8.0 was used as a mobile phase. The molecular weights of different polymers were determined against a calibration curve of PEG standards of molecular weights 1.4, 4.2, 12.6, 26, 46, 95, 170, 250, 510, and 885 kDa (Polysciences, Warrington, PA). 2.2.3. Hemolysis assay Human blood was collected in EDTA-containing vacutainers, which were centrifuged at 13,500g to separate the red blood cells (RBC). The plasma supernatant was discarded and the RBCs were washed three times using a 150-mM saline solution. After the third wash, the RBC solution was equally divided into three vacutainers and resuspended in

51

100-mM phosphate-buffered saline (PBS) solutions with the appropriate pH values (i.e., 5.8, 6.6, and 7.4). The RBC solutions were diluted 10 folds using PBS solutions with the corresponding pH values to reach a concentration of 108 RBC per 200 Al solution. Stock polymer solutions were prepared by dissolving the polymer in 10–20 Al of 1 N sodium hydroxide followed by diluting with PBS solution of pH 7.4 to reach a final concentration of 2 mg/ml. The hemolytic activity of the synthesized polymers, polymer-cationic peptide conjugates, and ionic complexes of asODN with polymer–cationic peptide conjugates was evaluated at polymer concentrations of 5, 10, and 20 Ag/ ml. The appropriate volume of the polymer stock solution was added to 800 Al of PBS solution and 200 Al of RBC solution with the appropriate pH to reach the desired polymer concentration. The RBC solutions were gently inverted several times for mixing with the added polymer solution then incubated for 60 min in a 37 8C water bath. Membrane-destabilizing activity of a given polymer was measured in terms of its ability to rupture the cell membranes of RBC allowing the release of hemoglobin into the solution. At the end of the incubation time, RBC solutions were centrifuged at 13,500g for 5 min resulting in intact and ruptured RBC to pellet out leaving the hemoglobin in the supernatant solution. Absorbance of hemoglobin in the supernatant was measured at its characteristic wavelength, 541 nm. The observed hemolysis of RBC in PBS solutions with different pH values and in 1% v/v Triton X-100 solution were used as negative and positive controls, respectively. The observed hemolytic activity of a given polymer at a given concentration and pH value was normalized to that of the positive control, 1% v/v Triton X-100 solution. All hemolysis experiments were carried out in triplicates. 2.2.4. Preparation of poly(PAA-co-BA-co-PDSA)– cationic peptide conjugate and complexation with asODN Based on the hemolysis results, poly(PAA-co-BAco-PDSA) terpolymer was considered a promising composition that can be further utilized as a pHsensitive membrane-destabilizing carrier for delivery of therapeutic asODN. The cationic peptide sequence, Cys-(Gly)3-(Lys)6, was covalently conjugated to the polymer backbone via disulfide linkages to the PDSA

52

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

Fig. 2. The endosomal barrier to intracellular delivery of therapeutic macromolecules.

units. The polymer was dissolved in 10–20 Al of 1 N sodium hydroxide then diluted with a 100 mM bicarbonate buffer of pH 10.0/200 mM NaCl solution to reach a concentration of 20 mg/ml. The peptide to PDSA molar ratio was adjusted to 2:1. The calculated amount of the peptide was dissolved in a 100-mM bicarbonate buffer of pH 10.0/200-mM NaCl solution, added to the polymer solution, and allowed to react at room temperature for 24 h. Conjugation of the peptide sequence to the polymer backbone was monitored spectrophotometrically by measuring the absorbance of the pyridine-2-thione released in the reaction medium at 320 nm. A total of 94.6% of PDSA units on the polymer backbone were conjugated to the Cys(Gly)3-(Lys)6 cationic peptide sequences. The reaction mixture was diluted 2–3 folds with a 100-mM bicarbonate buffer of pH 10.0 and run through a series of PD-10 desalting columns (Amersham Biosciences, Piscataway, NJ) equilibrated with the same buffer solution to remove excess unreacted peptide in the polymer–peptide solution. The final polymer-peptide solution was lyophilized for 48 h and stored at 20 8C until used. The amine content of the polymer-peptide conjugate was determined using the 2,4,6-trinitrobenzene sulfonic acid (TNBS) assay (Pierce, Rockford, IL) following the procedure described elsewhere [20]. The polymer–peptide conjugate was used to prepare ionic complexes with asODN at an amine to phosphate (N/P) ratio of 1:1 by mixing in a 100-mM phosphate-buffered saline solution of pH 7.4 for 15 min.

3. Results and discussion 3.1. Design, synthesis, and characterization of PDSAcontaining polymers The current work in carrier design was aimed at enhancing the cytoplasmic delivery of nucleic acid drugs (Fig. 2). Therapeutic nucleic acids including asODN can be ionically complexed to cationic peptides grafted onto the polymer backbone via disulfide linkages to the PDSA units (Fig. 3). This unique design will allow for (a) disruption of the endosomal membrane and diffusion of the carrier– ODN ionic complex into the cytoplasm and (b) release of the disulfide-conjugated cationic peptides with the complexed asODN into the cytoplasm by the reducing

Fig. 3. The design and mechanism of endosomal release of PDSA polymer–ODN ionic complexes.

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

Fig. 4. The reducing effect of cytoplasmic glutathione enzymes on polymer–drug disulfide linkage.

action of glutathione enzyme, commonly present in the cytoplasm (Fig. 4) [21]. The pH-responsiveness of PDSA-containing polymers is due to the carboxylic acid groups of the pHsensitive monomers incorporated in the polymer backbone, with the equilibria between the ionized and the protonated states determined by the environmental pH values relative to their pKa [22–24]. Amphiphilic polymers with carboxylic acid groups proved to destabilize organized lipid bilayers in a pHdependent fashion [22,24]. Design of smart polymeric carriers for intracellular drug delivery requires careful selection of the hydrophilic and hydrophobic monomers as well as optimization of their molar ratios to achieve high membrane-destabilizing activity at the desired pH values. Our group has developed a new glutathionereactive PDSA monomer to allow convenient conjugation of therapeutic peptides, proteins, and asODN to the polymer backbone via stable but reducible disulfide linkages [21,25,26]. In addition, this design allows the incorporation of targeting moieties onto the polymer backbone through the PDSA units using several chemistries such as disulfide or thioether

53

linkages. The first objective of this research was to examine the relationship between the length of the hydrophobic alkyl group substituted on the pHsensitive monomer and the pH profile of membrane disruption. The substituted alkyl group onto the pHsensitive monomer progressively increased in chain length by one methylene group from MAA to EAA to PAA. The second objective of this research was to study the effect of incorporating a hydrophobic monomer such as BA on the pH sensitivity and membrane-destabilizing activity of these new polymer compositions. A preliminary hemolysis study evaluated the change in membrane destabilizing activity of PPAA as a function of molecular weight. The hemolytic activity of PPAA increased with higher polymer molecular weights in the range of 20–80 kDa (data not shown). Based on these results, the different polymer compositions were synthesized with similar molecular weights in the range of 30–35 kDa, which correlates with the size/molecular weight requirements for potential in vivo applications where renal excretion is needed. In order to achieve this desired molecular weight range, the amount of DMF solvent and AIBN initiator was tailored for each polymerization reaction. The molecular weights of the first series of alkylacrylic acid and PDSA copolymers and poly(MAA-co-BA-co-PDSA) terpolymer were found to be in the desired molecular weight range (Table 1). However, poly(EAA-co-BAco-PDSA) and poly(PAA-co-BA-co-PDSA) terpolymers had lower molecular weights (Table 1). This is possibly due to the increase in length of the alkyl chain of the pH-sensitive monomer, which may have reduced the reactivity of the growing polymer chain leading to termination of the polymerization

Table 1 Composition and molecular weights of novel PDSA polymers Poly(MAA-co-PDSA) Poly(EAA-co-PDSA) Poly(PAA-co-PDSA) Poly(MAA-co-BA-co-PDSA) Poly(EAA-co-BA-co-PDSA) Poly(PAA-co-BA-co-PDSA) a

Monomer feed ratios

Polymer compositiona

M Wb

PIb

95/5 95/5 95/5 70/25/5 70/25/5 70/25/5

98/2 86/14 94.6/5.4 78/20/2 56/37/7 67/27/6

32,980 33,323 31,736 33,443 9471 12,107

3.9 4.5 4.7 2.1 2.4 2.2

Determined by 1H-NMR of the synthesized polymers in DMSO-d6 using a Bruker 499 MHz NMR spectrometer (Billerica, MA). Weight average molecular weight (M W) and polydispersity index (PI) of the synthesized polymers determined by gel permeation chromatography. b

54

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

reaction and formation of low-molecular weight polymers. In another preliminary study, it was found that high molecular weight (z100 kDa) polymer compositions with MAA/BA molar ratio of 60:40 were highly hydrophobic and appeared to precipitate out of solution with loss of the desired membrane-destabilizing activity (data not shown). Based on these results, the molar feed ratio of pH-sensitive, BA, and PDSA monomers was adjusted to 70%, 25%, and 5%, respectively. The exact composition of the synthesized polymers was determined by 1H-NMR analyses, i.e., the COOH peak of the pH-sensitive monomer, the aromatic peaks of PDSA monomer, and the O-CH2-R peak of BA and PDSA monomers. Results show that the compositions of the synthesized polymers were generally close to the monomer feed ratios (Table 1). 3.2. Membrane-destabilizing activity of new PDSA polymers The hemolytic activity of different polymer compositions was evaluated at polymer concentrations of 5, 10, and 20 Ag/ml in different pH values ranging from 5.8 to 7.4. Hemolysis results were used to identify which polymer composition(s) should be utilized as endosomal membrane-destabilizing drug carriers based on the established correlation between the observed hemolytic activity at acidic pH values and endosomal membrane disruption [6,11]. Polymer compositions that exhibited z70% of that observed with the positive control (1% Triton X-100), at acidic pH values, were considered potential membranedestabilizing carriers. In the first polymer series, only poly(PAA-coPDSA) copolymer exhibited a pH-dependent hemolytic activity. Both poly(MAA-co-PDSA) and poly (EAA-co-PDSA) copolymers showed no hemolysis at all the examined concentrations and pH values (Fig. 5). At a concentration of 5 Ag/ml, poly(PAAco-PDSA) copolymer showed low hemolysis at the physiologic pH of 7.4; however, the observed hemolytic activity increased significantly ( PV0.05) with the drop in pH values to 5.8 (Fig. 5A). The increase in polymer concentration did not affect the observed hemolytic activity at physiologic pH of 7.4 but caused an increase in hemolysis at pH 6.6

Fig. 5. The influence of increasing the length of the hydrophobic alkyl group substituted on the pH-sensitive monomer on the hemolytic activity of poly(MAA-co-PDSA), poly(EAA-co-PDSA), and poly(PAA-co-PDSA) polymers at pH 5.8 (5), 6.6 (M), and 7.4 (n). The hemolytic activity is examined at polymer concentrations of: (A) 5, (B) 10, and (C) 20 Ag/ml.

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

(Fig. 5). The observed pH- and concentrationdependent hemolytic activity of poly(PAA-coPDSA) copolymer is similar to that of PPAA reported earlier [12] and suggests the potential of this copolymer as a candidate carrier for drug molecules required to reach the cytoplasmic compartment for therapeutic action. These results clearly show the influence of increasing the length of the hydrophobic alkyl group substituted on the pHsensitive monomer on the membrane-destabilizing activity of the polymer backbone. In the second polymer series, the incorporation of hydrophobic BA monomers had a significant effect on the pH-responsiveness and membranedestabilizing activity of these polymers. Poly(MAAco-BA-co-PDSA) terpolymer exhibited insignificant hemolytic activity at all the evaluated concentrations and pH values (Fig. 6). Both poly(EAA-co-BA-coPDSA) and poly(PAA-co-BA-co-PDSA) terpolymers exhibited pH- and concentration-dependent hemolytic activities (Fig. 6). At a polymer concentration of 5 Ag/ml, poly(EAA-co-BA-co-PDSA) terpolymer showed high hemolytic activity, approximately 80% of that observed with the positive control, at acidic pH values of 5.8 and 6.6 (Fig. 6A). Poly(EAA-co-BA-co-PDSA) terpolymer also showed a gradual increase in the hemolytic activity with the increase in polymer concentration at pH 7.4 (Fig. 6). The observed pH-dependent hemolysis profile of this terpolymer is due to the pKa of the carboxylic groups of the EAA units on the polymer backbone. Our earlier studies showed that the pKa of EAA carboxylic groups in PEAA is approximately 5 [12]. The incorporation of hydrophobic BA in the backbone is expected to reduce the acidity of the EAA carboxylic groups and increase their pKa values, which leads to a higher percentage of EAA carboxylic groups being protonated and thus hemolytic at pH 6.6 and 5.8. By comparing the hemolytic activity of PPAA to that of poly(EAA-coBA-co-PDSA) terpolymer, we find that the hemolysis of one red blood cell requires 1.37106 PPAA polymer molecules compared to 4.18106 polymer molecules of this terpolymer. However, the poly (EAA-co-BA-co-PDSA) terpolymer remains an attractive carrier due to its ability to covalently carry different drug molecules through covalent disulfide linkages.

55

Fig. 6. The influence of incorporating the hydrophobic BA monomer in the polymer backbone on the hemolytic activity of poly(MAA-co-BA-co-PDSA), poly(EAA-co-BA-co-PDSA), and poly(PAA-co-BA-co-PDSA) polymers at pH 5.8 (5), 6.6 (M), and 7.4 (n). The hemolytic activity is examined at polymer concentrations of: (A) 5, (B) 10, and (C) 20 Ag/ml.

56

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

circulation by renal excretion after delivering their therapeutic cargo. 3.3. Membrane-destabilizing activity of polymer– cationic peptide conjugate and polymer–ODN complex

Fig. 7. The hemolytic activity of poly(PAA-co-BA-co-PDSA)[(Cys)-(Gly)3-(Lys)6] conjugate as a function of pH and polymer concentration.

Poly(PAA-co-BA-co-PDSA) terpolymer exhibited a hemolysis profile similar to that of poly(EAA-co-BAco-PDSA) terpolymer. However, the poly(PAA-coBA-co-PDSA) terpolymer appeared to precipitate out of solution at highly acidic pH values of 5.8 (Fig. 6). Our earlier studies showed that the pKa of PAA carboxylic groups in PPAA is approximately 6.3 [12] and it is expected to shift towards higher pH values due to the incorporation of hydrophobic BA monomer in the polymer backbone. This shift in the pKa values of PAA carboxylic groups is expected to cause polymer precipitation with loss of the hemolytic activity at pH 5.8 and increase in the hemolytic activity at physiologic pH 7.4 as shown in Fig. 6. By comparing the hemolytic activity of PPAA to that of poly(PAA-co-BA-co-PDSA) terpolymer, we find that the hemolysis of one red blood cell requires 1.37106 PPAA polymer molecules compared to 2.8106 polymer molecules of this terpolymer. The poly (PAA-co-BA-co-PDSA) terpolymer appears to be a stronger membrane-destabilizing carrier compared to the poly(EAA-co-BA-co-PDSA) terpolymer based on polymer concentration required to cause lysis of one red blood cell. It is interesting to note that both poly(EAA-co-BAco-PDSA) and poly(PAA-co-BA-co-PDSA) terpolymers exhibited this high hemolytic/membrane-destabilizing activity at low molecular weights of 9 and 12 kDa, respectively. This is a significant finding as it suggests their potential for further in vivo evaluation where they can be easily eliminated from the systemic

Conjugation of therapeutic molecules such as peptides and asODN to the polymer backbone is expected to reduce the membrane-destabilizing activity of the carrier as the hydrophilic/hydrophobic balance of the polymeric carrier shifts towards being more hydrophilic. These effects were evaluated by examining the hemolytic activity of poly(PAA-co-BA-co-PDSA) terpolymer after the conjugation of a model cationic peptide sequence [(Cys)-(Gly)3-(Lys)6] followed by ionic complexation of a therapeutic asODN (18 bases, 5699 kDa). The poly(PAA-co-BA-co-PDSA)-[(Cys)(Gly)3-(Lys)6] conjugate showed a pH- and concentration-dependent hemolytic activity (Fig. 7). At pH 5.8 and 6.6, the hemolytic activity of the polymer-peptide conjugate was statistically ( PV0.05) less than that of the polymer backbone at a concentration of 5 Ag/ml (Figs. 6 and 7). The small decline in the hemolytic activity at pH 5.8 and 6.6 can be attributed to the hydrophilic peptide sequence, which reduced the membrane-destabilizing activity of the polymer backbone. The hemo-

Fig. 8. The hemolytic activity of ODN ionic complexes with poly(PAA-co-BA-co-PDSA)-[(Cys)-(Gly)3-(Lys)6] conjugate as a function of pH and polymer concentration. The NH2/PO4 ratio is 1:1.

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

lytic activity of the polymer–peptide conjugate at pH 7.4 was similar to that of the free polymer backbone (Fig. 7). The ionic complexes of asODN with the poly(PAA-co-BA-co-PDSA)-[(Cys)-(Gly)3(Lys)6] conjugate were prepared following a NH2/ PO4 ratio of 1:1. These ionic complexes exhibited a pH- and concentration-dependent hemolytic activity profile similar to that of the polymer– peptide conjugate (Figs. 7 and 8). The observed hemolytic activity of the polymer-peptide conjugate and their ionic complexes with asODN collectively indicate that poly(PAA-co-BA-co-PDSA) terpolymer can retain its pH-dependent membrane-destabilizing activity after the complexation of therapeutic asODN or drug loading.

4. Conclusions The influence of composition on the pH-sensitive membrane-destabilizing activity of a family of PDSAcontaining polymers was examined. The pH sensitivity and hemolytic activity could be tuned by controlling the length of the hydrophobic alkyl group substituted on the pH-sensitive monomer and by the incorporation of a hydrophobic BA monomer. While the ratios of the monomers and the molecular weight of the polymers were kept similar in this study to isolate the effects of monomer composition, these two parameters provide further control over carrier properties. The results demonstrate that the membrane destabilizing activity is related to the pKa of the carboxylic groups incorporated in the polymer composition, which in turn dictates the hydrophilic/ hydrophobic balance of the polymer backbone. In addition, the delicate hydrophilic/hydrophobic balance of the polymer backbone is expected to vary with the quantity of the hydrophilic drug carried per polymer chain. Several new promising pH-sensitive, membrane-destabilizing, and glutathione-reactive polymer compositions were identified as a result of these structure-activity studies including poly(EAA-co-BAco-PDSA) and poly(PAA-co-BA-co-PDSA) terpolymers, and poly(PAA-co-PDSA) copolymer. Incorporation of the PDSA monomer in these polymer compositions allows for convenient conjugation of therapeutic molecules and/or targeting moieties to the polymer backbone via disulfide linkages, which

57

allows their utilization as drug carriers that can be further evaluated in vitro and in vivo.

Acknowledgments This work was funded by NIH Grant R01 EB299101, National Science Foundation Grant EEC 9529161 (UWEB ERC), and a National Cancer Center Postdoctoral Fellowship (MEHE).

References [1] D. Grimm, M.A. Kay, From virus evolution to vector revolution: use of naturally occurring serotypes of adenoassociated virus (AAV) as novel vectors for human gene therapy, Curr. Gene Ther. 3 (2003) 281 – 304. [2] J.J. Skehel, D.C. Wiley, Receptor binding and membrane fusion in virus entry: the influenza hemagglutinin, Annu. Rev. Biochem. 69 (2000) 531 – 569. [3] D.C. Wiley, J.J. Skehel, The structure and function of the hemagglutinin membrane glycoprotein of influenza virus, Annu. Rev. Biochem. 56 (1987) 365 – 394. [4] F.M. Hughson, Structural characterization of viral fusion proteins, Curr. Biol. 5 (1995) 265 – 274. [5] J. Ren, J.C. Sharpe, R.J. Collier, E. London, Membrane translocation of charged residues at the tips of hydrophobic helices in the T domain of diphtheria toxin, Biochemistry 38 (1999) 976 – 984. [6] C. Plank, B. Oberhauser, K. Mechtler, C. Koch, E. Wagner, The influence of endosome-disruptive peptides on gene transfer using synthetic virus-like gene transfer systems, J. Biol. Chem. 269 (1994) 12918 – 12924. [7] A.S. Hoffman, P.S. Stayton, O. Press, N. Murthy, C.A. Lackey, C. Cheung, F. Black, J. Campbell, N. Fausto, T.R. Kyriakides, P. Bornstein, Design of bsmartQ polymers that can direct intracellular delivery, Polym. Adv. Technol. 13 (2002) 992 – 999. [8] J.L. Thomas, D.A. Tirrell, Polyelectrolyte-sensitized phospholipid vesicles, Acc. Chem. Res. 25 (1992) 336 – 342. [9] J.L. Thomas, S.W. Barton, D.A. Tirrell, Membrane solubilization by a hydrophobic polyelectrolyte: surface activity and membrane binding, Biophys. J. 67 (1994) 1101 – 1106. [10] P. Schoen, A. Chonn, P.R. Cullis, J. Wilschut, P. Scherrer, Gene transfer mediated by fusion protein hemagglutinin reconstituted in cationic lipid vesicles, Gene Ther. 6 (1999) 823 – 832. [11] N. Murthy, I. Chang, P.S. Stayton, A.S. Hoffman, pH-sensitive hemolysis by random copolymers of alkyl acrylates and acrylic acid, Macromol. Symp. 172 (2001) 49 – 55. [12] N. Murthy, J.R. Robichaud, D.T. Tirrell, P.S. Stayton, A.S. Hoffman, The design and synthesis of polymers for eukaryotic membrane disruption, J. Control. Release 61 (1999) 137 – 143.

58

M.E.H. El-Sayed et al. / Journal of Controlled Release 101 (2005) 47–58

[13] C.Y. Cheung, N. Murthy, P.S. Stayton, A.S. Hoffman, A pHsensitive polymer that enhances cationic lipid-mediated gene transfer, Bioconjug. Chem. 12 (2001) 906 – 910. [14] T.R. Kyriakides, C.Y. Cheung, N. Murthy, P. Bornstein, P.S. Stayton, A.S. Hoffman, pH-Sensitive polymers that enhance intracellular drug delivery in vivo, J. Control. Release 78 (2002) 295 – 303. [15] C.A. Lackey, O.W. Press, A.S. Hoffman, P.S. Stayton, A biomimetic pH-responsive polymer directs endosomal release and intracellular delivery of an endocytosed antibody complex, Bioconjug. Chem. 13 (2002) 996 – 1001. [16] V. Bulmus, M. Woodward, L. Lin, N. Murthy, P.S. Stayton, A. Hoffman, A new pH-responsive and glutathione-reactive, endosomal membrane-disruptive polymeric carrier for intracellular delivery of biomolecular drugs, J. Control. Release 93 (2003) 105 – 120. [17] M. Ferrito, D.A. Tirrell, Poly(2-ethylacrylic acid), Macromol. Synth. 11 (1992) 59 – 62. [18] N. Murthy, J. Campbell, N. Fausto, A.S. Hoffman, P.S. Stayton, Bioinspired polymeric carriers that enhance intracellular delivery of biomolecular therapeutics, Bioconjug. Chem. 14 (2003) 412 – 419. [19] I.B. Nagy, A. Dancs, G. Koczan, G. Mezo, F. Hudecz, Conjugation of HS-oligopeptides with polymeric branched chain polypeptides containing multiple amino groups, J. Bioact. Compat. Polym. 15 (2000) 139 – 154.

[20] A.F.S.A. Habeeb, Determination of free amino groups in proteins by trinitrobenzenesulfonic acid, Anal. Biochem. 14 (1966) 328 – 336. [21] F.Q. Schafer, G.R. Buettner, Redox environment of the cell as viewed trough the redox state of the glutathione disulfide/ glutathione couple, Free Radic. Biol. Med. 30 (11) (2001) 1191 – 1212. [22] K. Borden, K. Eum, K. Langley, D. Tirrell, On the mechanism of polyelectrolyte-induced structural reorganization in thin molecular films, Macromolecules 20 (1987) 454 – 456. [23] A. Olea, J. Thomas, Fluorescence studies of the conformational changes of poly(methacrylic acid) with pH, Macromolecules 22 (1989) 1165 – 1169. [24] J. Linhardt, D.A. Tirrell, pH-induced fusion and lysis of phoshatidylcholine vesicles by the hydrophobic polyelectrolyte poly(2-ethylacrylic acid), Langmuir 16 (2000) 122 – 127. [25] E.P. Feener, W.C. Shen, H.J. Ryser, Cleavage of disulfide bonds in endocytosed macromolecules. A processing not associated with lysosomes or endosomes, J. Biol. Chem. 4 (5) (1990) 18780 – 18785. [26] G.D. McIntyre, C.F. Scott Jr., J. Ritz, W.A. Blattler, J.M. Lambert, Preparation and characterization of interleukin-2gelonin conjugates made using different cross-linking reagents, Bioconjug. Chem. 5 (1) (1994) 88 – 97.