pH-responsive pseudo-peptides for cell membrane disruption

Journal of Controlled Release 69 (2000) 297–307 www.elsevier.com / locate / jconrel

pH-responsive pseudo-peptides for cell membrane disruption M.E. Eccleston a , M. Kuiper b , F.M. Gilchrist b , N.K.H. Slater a , * a

Department of Chemical Engineering, University of Cambridge, Pembroke Street, Cambridge CB2 3 RA, UK b Chemical Engineering and Applied Chemistry, Aston University, Birmingham, UK Received 12 May 2000; accepted 17 August 2000

Abstract We describe pseudo-peptides obtained by the copolymerisation of L-lysine and L-lysine ethyl-ester with various hydrophobic dicarboxylic acid moieties. In aqueous solution, when the carboxylic acid groups are charged, the polymers dissolve. When they are fully neutralised the hydrophobic moieties cause the polymer to precipitate. The pH range over which reversible precipitation occurs can be adjusted by changing the intramolecular hydrophilic / hydrophobic balance, by using a carboxylic acid moiety with a different pKa value or by changing the apparent pKa value of the polymer through chemical modifications of the backbone. These bio-degradable materials are well tolerated by a range of mammalian cell lines at physiological pH but display an ability to associate with the outer membranes of these cells, which they rupture to varying degrees at pH 5.5. Relative to the degree of lysis displayed by poly( L-lysine iso-phthalamide), lysis was reduced by partial esterification and increased by replacing the aromatic iso-phthaloyl moiety with a long chain aliphatic dodecyl moiety. Similar behaviour was observed for the pH-dependent rupture of human erythrocytes, where poly( L-lysine dodecanamide) displayed enhanced cell lysis at pH values ,7.0 relative to poly( L-lysine iso-phthalamide).  2000 Elsevier Science B.V. All rights reserved. Keywords: Amphiphilic; Pseudo-peptides; Membrane disruption.

1. Introduction Low uptake across the plasma membrane of cells is a principal barrier to the effective intracellular delivery of macromolecules [1]. In nature, membrane disruption by amphiphilic peptides plays a central role in the pathogenesis of certain viral and bacterial toxin infections [2–4]. These peptides contain both hydrophobic and weakly-charged hydrophilic amino acid residues and upon activation become integrated into the extracellular or endosomal lipid bilayer *Corresponding author. E-mail address: nigel [email protected] (N.K.H. Slater). ]

membranes. In the case of viruses activation is a result of receptor-mediated binding to the extracellular cell surface and pH changes within the endosome during endocytosis. The reduction in pH causes such peptides to undergo a conformational transition whereby the hydrophobic groups begin to aggregate and become integrated into the lipid bilayer of the endosomal membrane. Subsequent formation of transient pores within the lipid bilayer allows escape of the viral capsid into the cytoplasm. The pH-dependent membrane lysis characteristics of such peptides have been used for the disruption and fusion of liposomes [5,6]. They have also been exploited for DNA delivery and conjugates displaying amphiphilic

0168-3659 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0168-3659( 00 )00316-3

298

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

sequences from the influenza virus hemagglutinin HA2 peptide have been used to enhance the efficiency of gene transfection [7,8]. We were interested in establishing whether synthetic amphiphilic pseudo-peptides might display similar membrane disruption behaviour towards cell membranes. Synthetic hydrophobically-modified polyelectrolytes have been used in-vitro to confer pH sensitivity to liposomes in order to allow the selective release of their contents. Tirrell et al. (e.g. [9,10]) has carried out extensive research into the membrane solubilisation effects of poly(amethacrylic acid), poly(a-ethacrylic acid) and copolymers of the two. The pH-dependent disruption of liposomal membranes was clearly demonstrated by the sharp fall in optical density of a suspension of multilamellar dilauroylphosphatidylcholine vesicles in 0.1% poly(a-ethacrylic acid). This was attributed to the disruption of the large liposomal vesicles and the formation of much smaller mixed lipid / polymer micelles [11]. The pH at which membrane disruption occurred was dependent on polymer concentration, ionic strength of the medium, tacticity and molecular weight of the polymer and could be controlled by co-polymerising a-methacrylic acid and a-ethacrylic acid [10]. The potential analytical and therapeutic applications of such responsive vesicles are widely recognised and have been reviewed [12]. It is proposed that the mode of action of these polymers is similar to that of the peptidic fusion proteins, whereby a structural reorganisation of the polymer, in response to changes in pH, results in increased interaction of its hydrophobic moieties with the liposome bilayer. The conformational transition exhibited by poly(a-ethacrylic acid) in response to a decrease in pH is known as hypercoiling [13] and several hypercoiling vinyl polymers exist with the potential to interact with lipid bilayers (e.g. [14– 16]).

The similarity between extracellular and liposomal membranes prompted us to question whether synthetic hypercoiling pseudo-peptides can disrupt cell membranes in an acidic environment. From a biomedical viewpoint, the molecular weight of vinyl polymers is limited to values below the renal threshold to prevent bioaccumulation. Furthermore, such polymers are not readily degradable in-vivo. This limitation could be overcome by utilizing a noncytotoxic, biodegradable polymer. Fenyo et al. [14] synthesised a polymer based on L-lysine that was condensed through its amine functions with 1,3benzene sulphonyl dichloride to give a hypercoiling poly(sulphonamide). With this in mind, we have condensed aromatic and aliphatic diacid chlorides with naturally occurring diamino acids and their esterified derivatives. We anticipate that the resulting functional polyamides would degrade to physiologically acceptable compounds and show pH-dependent solubility. Thus, the synthesis and membrane lysis properties of pseudo-peptides based on L-lysine and its esterified derivatives with aromatic diacid chlorides have been investigated.

2. Materials and methods

2.1. Polymer synthesis Polymers based on L-lysine and the aromatic diacid chloride iso-phthaloyl chloride (Fig. 1) were synthesised using a miscible mixed solvent technique in which acetone was employed as the solvent for the diacyl chloride [17]. This method was also used for the synthesis of partially-esterified copolymers and for all the work reported here, the initial reaction mixture contained a 4:1 molar ratio of L-lysine to L-lysine ethyl ester. The low hydrolytic stability of aliphatic and non-conjugated aromatic dicarboxylic

Fig. 1. Repeat unit structures of polycondensates of L-Lysine with aromatic and aliphatic di-acid chlorides.

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

acids precluded the use of a single acetone / aqueous phase reaction for these monomers. Polymers based on L-lysine and aliphatic diacyl chlorides were therefore synthesised by an interfacial technique in which chloroform was employed as the solvent for the diacyl chloride. In a typical interfacial polymerisation, L-lysine? HCl (0.02 mol), potassium carbonate (0.16 mol) and potassium chloride (0.135 mol) were dissolved in 100 ml deionised water, cooled until the appearance of ice crystals and placed in a Waring laboratory blender. Iso-phthaloyl chloride (0.02 mol) was dissolved in 100 ml of pre-cooled (2208C) chloroform. The solution of the diacyl chloride in chloroform was added to the aqueous diamine solution and stirred at full power for 30 min The resulting emulsion was diluted to a volume of 1 dm 3 with deionised water and the organic solvent removed on a rotary evaporator. For the single phase polycondensation syntheses sufficient L-lysine ethyl ester?2HCl and / or L-lysine (free base) to give a combined diamine concentration of 0.2 M were dissolved in 100 ml deionised water, cooled until the appearance of ice crystals and placed in a Waring laboratory blender. Iso-phthaloyl chloride (0.02 mol) was dissolved in 200 ml of precooled (2208C) acetone. The procedure continued as for interfacial polycondensation with the exception that the acetone was not removed prior to ultra filtration. Crude polymer solutions were concentrated to a volume of 100 ml using a Millipore MiniplateE unit containing a cellulose diafiltration membrane with a molecular-weight-cut-off of 3000 Da. They were then diafiltered with 500 ml deionised water to remove inorganic salts, low molecular weight oligomers and residual organic solvent. Diafiltration was preferred to dialysis to avoid hydrolysis of the Llysine ethyl ester moieties. Complete hydrolysis of the ester groups of the poly( L-lysine-co-L-lysine ethyl ester iso-phthalamides) co-polymers during conventional dialysis was confirmed by comparison of 1 H and 13 C NMR, which were all identical to that of poly( L-lysine iso-phthalamide). Diafiltered solutions were lyophilised to produce powdered polymer. Diafiltered polymer solutions were acidified to precipitate the polymers. Samples were isolated by Buchner filtration, washed with deionised water and

299

dried in a vacuum oven overnight at 608C. The remainders of the diafiltered polymer solutions were lyophilised to give the required polymer as the potassium salt. The polymers were characterised in DMSO at 258C by 13 C and 1 H NMR spectroscopy on a Bruker 300 NMR spectrometer (see [16]) Potentiometric titrations were carried out on a Radiometer Tim900 titration manager equipped with a Calomel combined electrode and a 5-ml automatic burette. The ionised polyamides dissolved readily in deionised water. The required amount of sodium chloride was added and the solution made up to 50 ml in a volumetric flask. Twenty-five ml of a 0.1 wt% solution of the polymer was titrated with 0.102 M hydrochloric acid in 0.0125 ml aliquots (0.25% of the maximum burette volume). Readings were taken when a critical stability of 60.05 pH units was reached or 60 s after addition of the titrant if this arbitrary stability was not achieved.

2.2. Cell culture COS1 and A2780 cells lines were used for the membrane disruption studies. COS1 cells were maintained in Dulbecco’s Modified Eagle’s medium (DMEM) and A2780’s were maintained in RPMI1640 medium. Both sets of medium were supplemented with 10% (v / v) foetal bovine serum (FBS), and made up to 2 mM L-glutamine, 100 U ml 21 penicillin and 10 mg ml 21 streptomycin. Both sets of cells were trypsinised twice-weekly using 0.25% trypsin–EDTA. COS1 and A2780 cells were seeded at a ratio of 1:6 and 1:5 respectively. Thereafter, cells were maintained in a humidified incubator at 378C and 5% CO 2 . Chinese Hamster Ovary (CHO) cells were employed for use in both fluorescence and confocal microscopy experiments. The CHO cell line was maintained in nutrient mixture F12 Ham medium that was supplemented as above. Cells were passaged two–three times weekly using 0.25% trypsin– EDTA and were seeded at a ratio of 1:10. Thereafter, cells were maintained as above.

2.3. Cytotoxicity and erythrolysis assays The membrane disruption characteristics of poly( L-lysine iso-phthalamide), poly( L-lysine ethyl ester-

300

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

co-L-lysine iso-phthalamide) and poly( L-lysine dodecanamide) were assessed. In a typical experiment, 5310 3 cells were washed with PBS and incubated in fresh, serum-free medium containing 2 mM L-glutamine at 378C for up to 60 min in the presence of 0–1 mg ml 21 of the appropriate polymer. After incubation the cells were pelleted by centrifugation at 250 g for 4 min. Lactate dehydrogenase (LDH) in either the supernatant, or from the lysed cell pellet (1% Triton X-100 for 45 min at 378C), was determined using a modified LDH assay (Promega Cytotox 96  ). The resulting UV absorption at 490 nm (Anthos Labtec instruments plate reader) was proportional to the number of cells lysed. Sheep erythrocytes were washed and resuspended in phosphate buffered saline containing the appropriate psuedo-peptide at the required concentration. The erythrocytes were then titrated with 0.1 N HCl to gradually reduce the supernatant pH using a Radiometer autotitrator and samples removed at the required pH. Two-hundred-ml samples were plated in quadruplicate in a round-bottomed 96-well plate. The plate was left for 15 min and centrifuged at 250 g for 3 min. The supernatants were transferred to a flatbottomed 96-well plate and the absorption at 540 nm was measured. The experiment was then repeated with the omission of polymer and the percentage of erythrocyte lysis determined.

3. Results and discussion

3.1. Polymer characterisation FT–IR analysis of the polymer precipitates on a

Nicolet 510M FT–IR spectrometer showed a characteristic absorption due to the carboxylic acid C=O stretch (|1710 cm 21 ) with strong absorption at |1640 cm 21 (amide band I) and at |1542 cm 21 (amide band II). In the poly( L-lysine ethyl ester-co-Llysine iso-phthalamide) samples the absorption due to the C=O stretch of the carboxylic acid at |1710 cm 21 was usually present as a shoulder on the more intense absorption of the ester carbonyl stretch at |1735 cm 21 . The molecular weight distributions of the diafiltered polymer samples were determined by gel permeation chromatography in dimethylformamide at 808C using a 30-cm polystyrene-co-di-vinyl benzene column with a differential refractometer detector. Weight average molecular weights of the diafiltered samples, expressed as polyethylene glycol / polyethylene oxide (PEG / PEO) equivalents, were typically in excess of 20 000 although the number average was much lower, indicating that the samples were polydispersed (data not shown).

3.2. Physical observations on the hypercoiling behaviour The polymers precipitate from aqueous solution at different pH’s (Table 1). Hydrophobic modification of weak polyelectrolytes can result in changes in its preferred conformation as the degree of ionisation of the weakly charged groups (e.g., carboxylate or amine groups). These conformational changes may change the solubility of the polymer. For example, the assumption of an amphiphilic structure can result in prolonged solubility of a hydrophobically modified polyelectrolyte, since the hydrophobicity is

Table 1 The experimentally-determined pH at which precipitation proceeds in aqueous solution for a range of pseudo-peptides containing pendant carboxylic acids and different hydrophobic moieties Sample

pH at onset of precipitation

poly( L-lysine ethyl-ester diethylmalonamide) poly( L-ornithine iso-phthalamide) poly( L-lysine ethyl-ester phenylmalonamide) poly( L-lysine iso-phthalamide) poly( L-lysine-co-L-lysine ethyl-ester iso-phthalamide) poly( L-lysine ethyl-ester phenylglutamide) poly( L-lysine-co-hexamethylene iso-phthalamide) poly( L-ornithine dodecanamide) poly( L-lysine dodecanamide)

4.00 4.13 4.38 4.39 4.47 4.70 4.75 4.85 5.00

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

buried within the core of the amphiphile. In the case of a polyanion, increasing the hydrophobic content may increase the pH below which the polymer assumes a collapsed conformation, rather than increase the pH at which it precipitates. This entropydriven hydrophobic association and associated conformational transition within polymers is known as hypercoiling. The phenomenon occurs when asymmetrically positioned hydrophobic moieties pendant to or within a polyelectrolyte overcome the electrostatic repulsion of the weakly-charged hydrophilic groups causing the polymer to adopt an abnormally compact conformation stabilized by hydrophobic bonding. The change in the apparent pKa of the polymers was determined as a function of their degree of ionisation from potentiometric titration data (Fig. 2) [18]. In the hypercoiled state the charged groups are drawn closer together than would be predicted by a simple electrostatic repulsion model and so exhibit a relatively high pKa . The assumption of an intramolecular amphiphilic structure is dependent on the separation of the hydrophobic and hydrophilic moieties. In the case of a polymer with asymmetrically positioned hydrophobic groups this is relatively easily achieved. However, potentiometric titration of some of the aliphatic lysine containing polyamides

301

revealed deviations similar to the poly(isopthalamides). This indicated that polyelectrolytes with aliphatic segments above a critical length were also capable of undergoing some kind of conformational transition that was triggered by a decrease in the degree of ionisation of the carboxylate functions. We speculate that as the degree of ionisation decreases then the aliphatic chains begin to contract and that when they contain sufficient methylene units, a hydrophobic tail can form. Such hydrophobic tails would no longer be symmetric and could collapse into an amphiphilic structure. That the tail formation and assumption of the amphiphilic structure is probably a concerted process is suggested by the relatively high degree of ionisation at which the collapse begins. Direct evidence for a conformational transition for poly( L-lysine-co-lysine ethyl ester iso-phthalamides) came from studies on the diafiltration behaviour of this material as a function of pH. A series of measurements of the degree of polymer transmission through the diafiltration membrane, as a function of solution pH, demonstrated a significant reduction in transmission at pH values close to those at which hypercoiling occurs (Fig. 3). This suggests that the polymer may be ‘snaking’ through the filter in an extended conformation but is effectively retained

Fig. 2. Variation of the apparent pKa of the polymers poly( L-lysine dodecanamide) (♦) and poly( L-lysine iso-phthalamide) (j) with degree of ionization.

302

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

Fig. 3. Variation in membrane transmission of poly( L-lysine iso-phthalamide) during diafiltration of solutions at different pH’s (d) pH 10.5, (♦) pH 5.6, (j) pH5.2 and (m) pH 4.8.

when coiling to an associated, globular form is initiated.

3.3. Membrane lysis The erythrolysis lysis results displayed in Fig. 4(b) indicate that in the presence of poly( L-lysine isophthalamide), the amount of erythrocyte lysis progressively increases relative to the control as the pH is reduced. This corresponds to increased cell lysis as the degree of ionisation of the polymer is reduced. The pH at which this enhancement in cell lysis occurs for poly( L-lysine iso-phthalamide) is relatively low (pH 4.6) compared to that found in the lysosomal compartment (around pH 5.5). Reduction in the density of carboxylic acid groups along the polymer backbone, by esterification of a fraction of these groups, marginally increases the pH at which the polymers precipitate (from pH 4.39 for poly( Llysine iso-phthalamide) to pH 4.47 for poly( L-lysine ethyl ester-co-L-lysine iso-phthalamide), Table 1). Corresponding to this change we observed the onset of erythrocyte lysis at slightly higher pH values in the presence of poly( L-lysine ethyl ester-co-L-lysine iso-phthalamide), Fig. 4(c). Increase in the hydrophobicity of the polymer through co-polymerisation with a dodecyl dicarboxylic acid leads to a more substantial increase in the pH at which precipitation occurs (Table 1). As a result, in the presence of

poly( L-lysine dodecanamide) erythrocyte lysis is promoted at pH values as high as pH 6.5–7.0 (Fig. 4(a)). In this case the gradual increase in lysis is attributed to the progressive association of the dodecanamide segments and their subsequent interaction with the lipid bilayer. To establish whether the polymers displayed any cell toxicity in their fully dissociated state we determined the cell lysis behaviour of the three polymers to COS1 cells after incubation for 60 min under normal cell culture conditions in the absence of serum (see Materials and methods). No toxicity was observed with any polymer at a concentration of 100 mg ml 21 or less. Cytotoxicities of 14.95% (poly( L-lysine dodecanamide)), 6.7% (poly( L-lysine ethyl ester-co-L-lysine iso-phthalamide)) and 13.7% (poly( L-lysine iso-phthalamide)) were measured at a polymer concentration of 500 mg ml 21 . However, at this concentration all the wells appeared turbid, which was ascribed to precipitation of the polymer on addition of acetic acid used to quench the enzymatic conversion of the tetrazolium salt. Thereafter, a modified version of the LDH assay was employed to avoid interference at high polymer concentrations. COS1 cells were incubated for 60 min with the fresh culture medium containing 1 mg ml 21 of each of the polymers. The cells were washed with serum-free medium, lysed with 0.9% Triton X-100 and the LDH assay performed on the clarified supernatant. At this

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

303

Fig. 4. (a) Relative erythrolysis of sheep erythrocytes with pH in the presence of poly( L-lysine dodecanamide) (b) poly( L-lysine iso-phthalamide) and (c) poly( L-lysine ethyl-ester-co-L-lysine iso-phthalamide) determined by adsorption at 540 nm.

304

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

high concentration poly( L-lysine dodecanamide) was observed to be the most toxic yielding a cell viability of 74%. Polymers poly( L-lysine ethyl ester-co-Llysine iso-phthalamide) and poly( L-lysine isophthalamide) appeared to be well tolerated with cell viability values of 96.8% and 95.4% respectively. These results indicate that under normal cell culture conditions and pH the polymers display little lytic behaviour towards the COS1 cells, even when present at relatively high concentrations. In-vitro cytotoxicity testing of poly( L-lysine / Llysine ethyl ester iso-phthalamide) copolymers with other cell lines (CHO, A2780, MAC-16 and C-26) similarly indicated that they were well tolerated at pH 7.4. The viability of CHO cells was affected little by the presence of polymer except at the high concentration of 5 mg ml 21 , when a 40% reduction in viability was evident after 4 h exposure. MAC-16 cells displayed no loss in viability after exposure to 0.5 mg ml 21 for up to 48 h. C-26 murine colorectal tumour cells displayed an IC 50 in excess of 0.5 mg ml 21 for poly( L-lysine iso-phthalamide) and an IC 50 of 0.4 mg ml 21 for poly( L-lysine ethyl ester-co-Llysine iso-phthalamide) at pH 7.4. We next sought to determine whether the exposure of cells to the polymers, followed by a reduction of the culture pH to that at which hypercoiling occurs, leads to cell lysis. Such an approach was chosen to simulate the reduced pH environment within endosomes. Again, 1 mg ml 21 of polymer was added to the COS1 cell suspension, the pH was adjusted to 5.5 with 20 mM HCl and the cells were incubated at 378C for 15, 30, 45 or 60 min After incubation, the LDH content was determined. In a separate set of experiments approximately 1310 4 A2780 cells were incubated for 30 min at 378C with either 1 mg ml 21 of poly( L-lysine iso-phthalamide) or poly( L-lysine dodecanamide). The pH of the supernatant in each well was dropped to pH 5.0 by addition of 0.1 M HCl. At the appropriate time intervals (15, 30, 45 or 60 min) the acidified supernatant was removed from each well. These wells were then resuspended in fresh medium (100 ml). Thereafter, an LDH assay was carried out according to the method described previously for COS1 cells. The results for both COS1 and A2780 cells are presented in Fig. 5 as cell viabilities relative to control experiments in which cells were acidified for

equivalent periods of time with no polymer present. The overall COS1 cell viability dropped markedly for cells incubated at pH 5.5 for 60 min but clear differences both with the control (no polymer) and each other were observed in the presence of all polymers. Poly( L-lysine dodecanamide) appeared to be the most toxic again with a cell viability of 0.8% after 60 min. Poly( L-lysine ethyl ester-co-L-lysine iso-phthalamide) appeared to be the best tolerated and poly( L-lysine dodecanamide) the least. A2780 cells were similarly insensitive to polymer treatment at pH 7.8 but showed a pH-dependent sensitivity to polymers poly( L-lysine dodecanamide) and poly( Llysine iso-phthalamide) upon acidification (though the relative loss of viability was now only ca. 50% after 60 min exposure). Finally, we sought to observe the association of Cy3 conjugated poly( L-lysine iso-phthalamide) with cell membranes. Fluorescence microscopy showed that at pH 7.4 the fully disassociated polymer adsorbs very weakly onto the membrane of CHO cells following 30 min incubation. By contrast, incubation at pH 5.5 led to extensive adsorption of the polymer (Fig. 6(a)). Repeated washing cycles removed the cell bound fluorescence, suggesting a reversible binding to the outer cell membrane. Further examination by confocal microscopy (Fig. 6(b)) showed that the adsorption of polymer at pH 5.5 is largely onto the extracellular lipid bilayer, with little internalisation of the polymer apparent. Cationic polymer–hydrophobe conjugates have been shown to induce pH dependent cell lysis [19]. For these materials the adsorption of polymer onto membranes at low pH is favoured by both electrostatic and hydrophobic interactions. Intuitively, little adsorption of anionic polymers would be anticipated at pH values above their pKa range, though hydrophobic association might occur at low pH. Studies on the relationship between membrane disruption and the chemical structure of lytic peptides indicate that whereas electrostatic interactions influence initial association, hydrophobic interactions control the penetration of the peptide into the membrane and its eventual disruption [20]. Indeed, in one study the extent of lipid vesicle disruption by an amphipathic peptide was observed to increase with a reduction in electrostatic interaction [21]. Fig. 6 seems consistent with these observations. At pH 7.4 the electrostatic

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

305

Fig. 5. The viability of (a) COS1 and (b) A2780 cells relative to a control following dosing with poly( L-lysine iso-phthalamide), (shaded), poly( L-lysine ethyl-ester-co-L-lysine iso-phthalamide), (stippled), and poly( L-lysine dodecanamide), (striped), pH adjusted to 5.5 with 20 mM HCl.

interaction between the anionic polymer and cell membranes is unfavourable for adsorption, whereas progressive charge neutralisation at lower pH’s enables polymer to approach the membrane sufficiently for hydrophobic interactions to take place. Notwithstanding, potentiometric titration of the polymers shows that their precipitation occurs at pH

values below that at which cell disruption was observed here (pH 5.5). A very similar observation for the pH-dependent disruption of phospholipid vesicles by poly(2-ethacrylic acid) has been remarked upon by Thomas et al. [12], who hypothesise that membrane disruption arises from the structural rearrangement of the membrane bound polymer as a

306

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307

poly( L-lysine ethyl ester-co-lysine iso-phthalamide) show very little toxicity. At pH 5.5 a significant increase in toxicity for all polymers is seen, compared to when no polymer is present and poly( Llysine iso-phthalamide) appears to be approximately twice as toxic as poly( L-lysine ethyl ester-co-lysine iso-phthalamide) at the lower pH.

Acknowledgements The authors thank Dr. M. Read and Dr. L. Seymour of the CRC Institute for Cancer Studies, University of Birmingham for technical assistance with the cytotoxicity studies with COS1 cells and Nycomed Amersham PLC for providing confocal microscopy facilities.

References

Fig. 6. (a) Fluorescence microscopy images of CHO cells treated with 1 mg / ml Cy3 labelled poly( L-lysine iso-phthalamide) at pH 5.5. (b) Confocal microscopy image of cells as in (a).

distinct process from their solution-phase hypercoiling.

4. Conclusions In-vitro experiments suggest that these bio-degradable, hypercoiling pseudo-peptides can integrate with cell membranes and exhibit pH mediated cell membrane disruption, the extent of which can be varied by careful choice of polymer structure. Thus, whilst poly( L-lysine dodecanamide) is mildly toxic at physiological pH, poly( L-lysine iso-phthalamide) and

[1] D. Luo, W.M. Saltzman, Synthetic DNA delivery systems, Nat. Biotechnol. 18 (1) (2000) 33–37. [2] L. Carrasco, Entry of animal viruses and macromolecules into cells, FEBS Lett. 350 (2–3) (1994) 151–154. [3] H. Ostolaza, L. Bakas, F.M. Goni, Balance of electrostatic and hydrophobic interactions in the lysis of model membranes by E.-coli alpha-haemolysin, J. Membr. Biol. 158 (2) (1997) 137–145. [4] A. Aspedon, E.A. Groisman, The antibacterial action of protamine: evidence for disruption of cytoplasmic membrane energization in Salmonella typhimurium, Microbiology–UK 142 (12) (1996) 3389–3397. [5] M.H.B. Dacosta, H. Chaimovich, Structure-activity relationships in the fusion of small unilamellar phosphotidylcholine vesicles induced by a model peptide, Biochemie 79 (8) (1997) 509–516. [6] K.H. Hahn, H. Kim, Fusion and fragmentation of phospholipid-vesicles by apohemoglobin at low pH, J. Biochem. 110 (4) (1991) 635–640. [7] E. Wagner, C. Plank, K. Zatloukal, M. Cotten, M.L. Birnstiel, Influenza-virus hemagglutinin HA2 n-terminal fusogenic peptides augment gene-transfer by transferrin polylysine dna complexes — toward a synthetic virus-like gene-transfer vehicle, Proc. Natl. Acad. Sci. 89 (17) (1992) 7934–7938. [8] C. Plank, B. Oberhauser, K. Mechtler, C. Koch, E. Wagner, The influence of endosome-disruptive peptides on genetransfer using synthetic virus-like gene-transfer systems, J. Biol. Chem. 269 (17) (1994) 12918–12924. [9] J.L. Thomas, S.W. Barton, D.A. Tirrell, Membrane solubilisation by a hydrophobic polyelectrolyte: surface activity and membrane binding, Biophys. J. 67 (1994) 1101–1106.

M.E. Eccleston et al. / Journal of Controlled Release 69 (2000) 297 – 307 [10] J.L. Thomas, H. You, D.A. Tirrell, Tuning the response of a pH-sensitive membrane switch, J. Am. Chem. Soc. 117 (10) (1995) 2949–2950. [11] D.A. Tirrell, D.A. Takigawa, K. Seki, pH sensitisation of phospholipid vesicles via complexation with synthetic poly(carboxylic acids), Ann. N.Y. Acad. Sci. 446 (1985) 237– 248. [12] J.L. Tirrell, D.A. Tirrell, Polyelectrolyte-sensitized phospholipid vesicles, Acc. Chem. Res. 25 (1992) 336–342. [13] A.F. Oles, J.K. Thomas, Fluorescence studies of the conformational changes of poly(methacrylic acid) with pH, Macromolecules 22 (1989) 1165–1169. [14] J.C. Fenyo, Polycondensats entre la lysine et le disulphochlorure-1,3 benzene. mise en evidence d’interactions hydrophobes en solution aqueuse, Eur. Polym. J. 10 (1974) 233–237. [15] S.P. Gasper, J.S. Tan, Conformational studies of poly(ethyl acrylate-co-acrylic acid), J. Poly. Sci. Technol. 2 (1973) 87–400. [16] N. Murthy, J.R. Robichaud, D.A. Tirrell, P.S. Stayton, A.S. Hoffman, The design and synthesis of polymers for eukaryotic membrane disruption, J. Control. Rel. 61 (1999) 137– 143. [17] M.E. Eccleston, N.K.H. Slater, B.J. Tighe, Synthetic routes

[18]

[19]

[20]

[21]

307

to responsive polymers; co-polycondensation of tri-functional amino acids with diacylchlorides, React. Func. Polym. 42 (1999) 147–161. C. Methenitis, J. Morcellet, G. Pneumatikakis, M. Morcellet, Polymers with amino acids in their side chain: conformation of polymers derived with glyclglycine and phenylalanine, Macromolecules 27 (1994) 1455–1460. R. Duncan, P. Ferruti, D. Sgouras, A. Tuboku-Metzger, E. Ranucci, F. Bignotti, A polymer-triton X-100 conjugate capable of pH-dependent red blood cell lysis: a model system illustrating the possiblity of drug delivery within acidic intracellular compartments, J. Drug Targeting 2 (1994) 341–347. A. Scheller, J. Oehlke, B. Wiesner, M. Dathe, E. Krause, M. Beyermann, M. Melzig, M. Bienert, Structural requirements for cellular uptake of alpha-helical amphipathic peptides, J. Peptide Sci. 5 (4) (1999) 185–194. M. Dathe, M. Schumann, T. Wieprecht, A. Winkler, M. Beyermann, E. Krause, K. Matsuzaki, O. Murase, M. Bienert, Peptide helicity and membrane surface charge modulate the balance of electrostatic and hydrophobic interactions with lipid bilayers and biological membranes, Biochemistry 35 (38) (1996) 12612–12622.