Bioresponsive poly(amidoamine)s (PAA)s are currently under development as endosomolytic polymers for intracellular delivery of proteins and genes. Here for the first time, small-angle neutron scattering (SANS) is used to systematically investigate the pH-dependent conformational change of an endosomolytic polymer, the PAA ISA 23. The radius of gyration of the ISA23 was determined as a function of pH and counterion, the aim being to correlate changes in polymer conformation with membrane activity assessed using a rat red blood cell haemolysis assay. With decreasing pH, the ISA23 radius of gyration increased to a maximum (R(g) approximately 80 A) around pH = 3, before subsequently decreasing once more. At high pH and therefore high ionic strengths, the polymer is negatively charged and adopts a rather compact structure (R(g) approximately 20 A), presumably with the dissociated carboxylic groups on the exterior of the polymer coil. At low pH, the coil again collapses (R(g) < 20 A), presumably due to the effects of the high ionic strength. It is concluded that the nature of the salt form has no direct bearing on the size of the polymer coil, but it does indirectly determine the prevailing pH and, hence, polymer conformation. Pulsed-gradient spin-echo NMR measurements were in good agreement with the SANS estimates of the radius of gyration, although ISA23 polydispersity does complicate the data interpretation/comparison. These results support the proposed mode of action of PAAs, namely a coil expansion on passing from a neutral pH (extracellular) to an acidic pH (endosomal and lysosomal) environments. The results do, however, suggest that the charge on the polymer shows a closer correlation with the haemolysis activity rather than the polymer conformation.

No abstract

Polymers are appealing as pH-responsive elements of multicomponent systems designed to promote cytosolic delivery of macromolecular drugs (including proteins and genes), but so far the delivery efficiency achieved has been relatively modest. Therefore, the aim of this study was to apply several physicochemical techniques that are well established in the colloid field (surface tension measurements, small-angle neutron scattering (SANS), and electron paramagnetic resonance (EPR)) to probe the mechanism of endosomolytic polymer-surface interaction over the pH range 7.4 to 5.5 using the poly(amidoamine) (PAA) ISA23 x HCl and a series of "model" micelle surfaces. These micellar models were chosen to represent increasing complexity from simple, single surfactant sodium dodecylsulfate (SDS) micelles, surfactant mixtures containing bulky malono-bis-N-methylglucamide headgroups, or highly extended ethylene oxide headgroups. Spherical micelles composed of 1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine (lyso-PC) were also used. Changes in the onset of micellization, micelle surface fluidity, and in selected cases, the overall micelle shape and size were all quantified as a function of pH in the presence and absence of ISA23 x HCl. This amphoteric PAA is negatively charged at pH 7.4 and becomes gradually more protonated on exposure to lower pH values representative of the endosomal-lysosomal pathway. As expected, the strength of polymer interaction with anionic micelles increased with a decrease in pH, while for cationic micelles the opposite was observed. Addition of bulky, nonionic surfactant headgroups led to weaker interactions. The observations from surface tension and SANS studies showed a complex pattern of interaction with both an electrostatic and hydrophobic component. Using EPR it was confirmed that ISA23 x HCl perturbed the micelle palisade layer leading to a decrease in fluidity of the interface with a lower degree of headgroup hydration, and a significant change in micelle morphology. Surprisingly, there was no interaction between ISA23 x HCl and globular micelles formed from lyso-PC (a more biologically relevant model), and this suggests that the PAA structure could be better optimized to promote rapid interaction with endosomal membranes at the physiologically relevant pH 6.5.

Small-angle neutron scattering (SANS) has been employed to characterize mixed micelles comprising the surfactants 1,2-diheptanoyl-sn-phosphatidylcholine (DHPC)/sodium dodecyl sulfate (SDS) and 1,2-diheptanoyl-sn-phosphatidylcholine (DHPC)/dodecyltrimethylammonium bromide (DTAB) using the classical model of a micelle consisting of a hydrocarbon core surrounded by a polar shell. Various constraints are applied to the data fitting, specifically the volume of the hydrocarbon core using the aggregation number determined from time-resolved fluorescence quenching (TRFQ) and the degree of micelle hydration as determined from an electron paramagnetic resonance (EPR) experiment. The morphologies of the DHPC/SDS and DHPC/DTAB mixed micelles are largely invariant with composition-as might be expected given the similarity in the tail volumes of the respective surfactants, and the observed behaviour of the aggregation number-with a radius comparable to the fully extended length of a dodecyl chain, i.e. the longer of the two hydrophobic moieties. The shell thickness is also largely invariant with composition, indicating that the phosphatidylcholine headgroup lies flat at the interface. For both cases, the degree of counter-ion dissociation extracted from the fitting of the SANS data increases significantly on addition of DHPC, in a manner very similar to that observed for related binary surfactant mixtures, provided there is only a small change in micelle aggregation number (micelle curvature). The coincidence of the degrees of counter-ion dissociation for these quite different systems suggests a common underlying behaviour in which the nonionic species merely dilutes the surface charge associated with the ionic headgroups.

The mechanism of ISA23 · HCl interaction with model membrane vesicles (80-100 nm in diameter) was investigated using EPR in conjunction with SANS. For EPR, 16-DSE was dissolved in the vesicle membrane to measure its dynamics and polarity, whereas a spin-labeled (Tempo)-ISA 23 analogue was used to give a measure of the polymer flexibility. When ISA23 was added to the external vesicle surface, no interaction was found. This observation conflicts with the reported ability to lyse RBC, but is in agreement with recent studies that showed no effect on membrane permeability when a PAA was added to an incubation medium containing isolated lysosomal vesicles. The vesicle-mimetic models used here provide a new and useful tool for studying endosomolytic polymer/membrane interactions.