Acid-cleavable poly(oxazoline) surfactants

Abstract

The acidic tumor microenvironment and late endosomes present a promising target for stimuli-responsive nanotherapeutics. Acid-cleavable surfactants, particularly those with hydrazone linkages, offer enhanced stability outside the cell while enabling efficient intracellular payload release. Their acid-triggered cleavage and cationic byproducts facilitate endosomal escape, making them attractive for cancer nanomedicine. Herein, we report the synthesis of a new hydrazone-linked poly(oxazoline)-based diblock copolymer surfactant. This surfactant cleaves in a pH-dependent manner going from pH 7.4 down to pH 5.0, where after 21 h, 80% ± 3% of the hydrazone-linked polymer remained at pH 7.4 compared to 17% ± 2% at pH 5.0. We then demonstrate the ability of nanoemulsion encapsulated payloads to partition into cell membrane mimics only after cleavage of the surfactant. Through this system, we were able to increase the amount of payload release from 26% to 47% over 42 hours through pH changes. In all, this work demonstrates a viable route to create POx-based nanomaterials with controlled release capabilities in biologically relevant conditions and is a promising platform for advancing the endosomal escape and cancer targeting of nanomaterials.

1. Introduction

Noncovalent encapsulation of poorly soluble drugs into nanomaterials has been a prolific strategy for improving the utility of drugs over the past few decades.^1–3 The core of this strategy involves the use of surfactants – polymers with hydrophilic and hydrophobic segments self-assembling around the drug due to the hydrophobic effect.⁴ While nanomaterials have seen great success in solubilizing drugs and altering their pharmacokinetic properties, limitations in endosomal escape and controlled release of the payload are still prevalent. A strategy to overcome both these challenges is to incorporate stimuli-responsive groups into the polymer surfactant composing the nanoparticle. The responsive group will then facilitate a chemical or conformational change leading to release of the payload and/or endosomal escape.

There are many stimuli which have been explored for responsive nanomaterials which broadly fall into two categories of endogenous or exogenous stimuli. While exogeneous stimuli such as light or ultrasound have been successfully used in pre-clinical models, they have historically been limited by poor tissue penetration⁵ and damage to tissue,^6,7 respectively. There is a plethora of endogenous stimuli that can instead be used to facilitate controlled release and are not limited by the challenges exogenous stimuli face. These endogenous stimuli should be present in appreciably high concentrations within the cell to illicit a chemical response. In the presence of the stimulus, the nanomaterial undergoes a significant conformational change or cleavage, leading to release of the encapsulated payload (Fig. 1A). Some of the most popular stimuli utilized in controlled release mechanisms include glutathione,⁸ reactive oxygen species,^9,10 metabolic enzymes such as esterases,^11,12 as well as the low pH of the late endosome and tumor microenvironment.^13–15

Fig. 1.

P(Ox) nanomaterials for endosomal escape. (A) Schematic showing how controlled release of a payload in a stimuli-responsive nanoemulsion leads to endosomal escape. (B) Summary of previous work regarding glutathione and acid cleavable poly(oxazoline) surfactants. (C) Overview of this work – acid-responsive perfluorocarbon nanoemulsions achieved through a hydrazone-linked, acid-cleavable poly(oxazoline) surfactant.

Of the aforementioned stimuli, nanomaterials which respond to a pH change provide solutions to both the challenges of payload release and endosomal escape. Furthermore, they can be utilized for targeted chemotherapeutic delivery due to the decreased pH in tumors. Consequently, a vast array of acid-cleavable functional groups with varying stabilities have been made, including orthoesters,^16–19 ketals/acetals,^20–23 imines,^24–29 hydrazones,^30–33 semicarbazones,^34,35 and oximes.^36–38 While all have been used in payload delivery into cells, the imine-like linkages offer further opportunities to promote endosomal escape because the nitrogen-containing byproduct is cationic in aqueous environments. Cationic materials are known to facilitate endosomal escape by rupturing the endosomal membrane.^39–41 However, direct introduction of cationic materials is cytotoxic.^42–44 Thus, responsive nanomaterials which become cationic, sometimes referred to as “charge switching” materials, are particularly advantageous.^45–49

Poly(oxazoline) (P(Ox)) has emerged in the field as a desirable polymer scaffold to constitute nanoparticles. Poly(2-methyl-2-oxazoline) (P(MeOx)) and poly(2-ethyl-2-oxazoline) (P(EtOx)) specifically display high antifouling abilities, which prevent nonspecific interactions with endogenous biomolecules.^50–52 Additionally, the size of the polymers can be easily controlled through the monomer : initiator stoichiometry, and P(Ox) based surfactants can be easily synthesized in one pot through the addition of a second hydrophobic monomer.^53–56 Lastly, P(Ox) functionality is highly modular; novel functional groups can be installed as nucleophilic terminating agents or electrophilic initiators. Though there are plenty of examples of unfunctionalized P(Ox)-based surfactants used in the formulation of nanoparticles for payload delivery, there are relatively few which contain stimuli-responsive moieties.^24,57–61 In a report from our group in 2021, we showed the synthesis of a P(MeOx) and poly(2-nonyl-2-oxazoline) (P(NonOx))-based diblock copolymer surfactant where the two blocks were linked by a disulfide bridge (Fig. 1B).⁶⁰ This report used glutathione to facilitate payload release of solubilized enhanced green fluorescent protein plasmid DNA out of perfluorocarbon (PFC)-in-water nanoemulsions through the degradation of the surfactant at the interface. Although this report was successful in demonstrating increased payload release into cells when comparing the cleavable and noncleavable versions, one limitation was the need to add glutathione into the cell culture media for emulsion cleavage to occur in the endosome. Thus, the need for an alternative linkage that responds directly upon endocytosis was necessary.

P(Ox) surfactants with acid-labile moieties have also been reported. Kempe and coworkers in 2024 reported the synthesis and use of P(Ox) surfactants bearing aldehyde side chains (Fig. 1B).²⁴ These surfactants were used to formulate core-crosslinked micelles with 1,4-phenyldiamine, forming an acid-labile imine linkage. Their payload of interest bore an amine and was also covalently linked into the micelle core through an imine. Though they were able to demonstrate lability in harsh acidic conditions (pH 2), no other pH besides 7.2 was tested. In pH 7.2, there was also appreciable payload release (up to 31%) over the course of 24 h. A more stable linkage would be desirable to prevent premature payload release.

To solve the need for a polymer that responds to physiologically relevant endogenous pH changes, we report a P(MeOx) – P(NonOx) based diblock copolymer with a hydrazone linkage (Fig. 1C). Employing a hydrazone linkage instead of an imine linkage endows the material with additional stability but remains labile enough to cleave within low pH environments.^62,63 Furthermore, placing the hydrazone between the two blocks of the surfactant maximizes the impact that individual cleavage events have on the nanoparticle stability. We use the previously reported hydrazine end cap⁶⁴ in conjunction with an unreported protected aldehyde-initiated hydrophilic block to generate hydrazone-linked surfactants. After surfactant synthesis, we formulate PFC-in-water nanoemulsions and demonstrate the pH dependent emulsion degradation and payload release mechanisms in vitro.

2. Results and discussion

2.1. Surfactant synthesis

We began the synthesis of the hydrazone-linked diblock copolymer through the synthesis of each respective block (Fig. 2A). To synthesize the aldehyde containing hydrophilic block, we used commercially available 2-(2-bromoethyl)-1,3-dioxolane as an initiator to polymerize 2-methyl-2-oxazoline (1). Although this initiator has not been reported, we employed the rates of a similar, reported alkyl bromide initiator to determine the appropriate reaction times.⁶⁵ Notably, additional equivalents of initiator were required in order to obtain the desired molecular weights. Nonetheless, dioxolane-bearing P(MeOx) (2) was obtained in large scale and a low dispersity (Đ = 1.12) after purification by dialysis. Then, a reported protocol for acetal deprotection was modified to yield aldehyde-bearing P(MeOx) (3) after dialysis.²⁵ In order to obtain a hydrazine-containing hydrophobic block, methyl trifluoromethanesulfonate was used to initiate the polymerization of 2-nonyl-2-oxazoline (4). This polymerization was terminated with commercially available t-butyl carbazate, adapting previously reported conditions to yield the Boc-hydrazine bearing P(NonOx) (5).⁶⁴ Polymer 5 was treated with 50 vol% TFA in DCM in order to remove the Boc group. After aqueous quenching and washes of the organic phase, hydrazine-bearing P(NonOx) (6) was obtained. Interestingly, while MALDI analysis of polymers 5 and 6 shows monomodal character of both polymers, SEC analysis shows bimodal character present in both polymers. This bimodality appears after the termination step with boc hydrazine to form 5 and it was reproduced in every batch of polymer. While we were unable to identify this second polymer species, we are confident that 66% of the polymer contains the desired hydrazine moiety (determined by the Boc integral in the ¹H NMR spectrum of 5). Lastly, polymers 3 and 6 were mixed together in the presence of acetic acid and magnesium sulfate to facilitate the condensation of water and the coupling of the polymers. Polymer 6 was used in slight excess to push the reaction to conversion because this homopolymer was the easier of the two to remove following post-polymerization modification. Polymer 6 was reproducibly bimodal, but did not impart bimodality onto surfactant 7. The reaction was monitored by size exclusion chromatography (SEC) and was complete after 18 hours. After filtering off salts, a workup protocol was devised which capitalized on the amphiphilic nature of the product. An emulsion phase was made by manually shaking the solution with saturated sodium bicarbonate to quench the acid. The emulsion was then washed with water to remove 3, or organic solvent to remove 6 (Fig. S1†), thus yielding acid-labile P(MeOx)-Hz-P(NonOx) (7). As a negative control for future experiments, a noncleavable surfactant P(MeOx)-b-P (NonOx) (8) was synthesized according to literature protocol.⁶⁶ We also synthesized an amine bearing P(NonOx)NH₂ (Schemes S1 and S2†) to synthesize an imine-linked acid-cleavable surfactant (Fig. S2A and S3†), however the unstable nature of the imine caused severe degradation in the emulsion purification protocol and prevented isolation (Fig. S2B†). Therefore, only surfactants 7 and 8 were compared for this study. The polymers obtained throughout the synthesis all displayed low dispersities (Đ ≤ 1.2) (Fig. 2B). The purity of 7 was determined through SEC, ¹H nuclear magnetic resonance spectroscopy (¹H NMR), and matrix assisted laser desorption ionization (MALDI) mass spectrometry. By SEC, we observe an increase in the M_n of 7 from 3 of 900 Da, which is approximately the M_n of 6. Furthermore, only small amounts of P(NonOx) homopolymer that could not be removed were present (Fig. 2C) (2% by peak area). By ¹H NMR, a block ratio of 3.2 : 1 was observed, which is close to the theoretical ratio of 3 : 1 (Fig. 2D). MALDI analysis shows 7 with small amounts of 3 and 6 present – though the exact amount cannot be quantified using this mode of characterization (Fig. 2E). While small amounts of homopolymer are present within the sample, we do not expect them to affect the performance or function of 7. Lastly, we demonstrated that 7 is interfacially active by formulating perfluorocarbon-in-water nanoemulsions while varying the concentration of 7 (Fig. S3†).

Fig. 2.

Synthesis and characterization of acid-cleavable surfactant 7. (A) Forward synthesis of acid-cleavable surfactant 7, along with structure of noncleavable surfactant 8. (B) Table of characterization for polymers 2–8. (C) SEC trace of 3 (blue), 6 (grey), and 7 (black). (D) NMR of surfactant 7 showing the appropriate peaks and blocks in the expected ratio of 3 : 1. (E) MALDI spectrum of 3 (blue), 6 (grey), and 7 (black).

2.2. Surfactant degradation

With cleavable surfactant 7 in hand, we sought to characterize its degradation profiles in differing pH buffers (Fig. 3A). Citrate phosphate buffer (CPB) was chosen, as its buffering range spans the desired pHs we aimed to test. A 1 : 1 ratio of this buffer and MeOH was used to ensure complete solubility of the surfactant. We incubated the samples at 37 °C to reflect cell incubation conditions. First, we incubated 7 in 1 : 1 MeOH/pH 5.0 CPB for 23 h. Gratifyingly, complete degradation of 7 was observed by MALDI (Fig. 3B). We then quantified the degradation by SEC, monitoring the increase in the retention time going from 7 back to 3 upon hydrolysis of the hydrazone. We generated a calibration curve using mixtures of 3 and 7 to quantify the reaction progress (Fig. S4†). As expected, we observed minimal degradation of 7 at early time points in pH 7.4 buffer, where 89% ± 2% of the surfactant persisted after 3 h (Fig. 3C and Fig. S5†). Importantly, increased degradation profiles were observed when using more acidic buffer, where 81% ± 1%, 85% ± 1% and 68% ± 1% of 7 remained in pH 7.0, pH 6.0, and pH 5.0, respectively, after 3 h. After 21 h, 80% ± 3%, 60% ± 1%, 41% ± 3%, and 17% ± 2% of 7 remained in pH 7.4, 7.0, 6.0, and 5.0 buffers, respectively. Although similar values were observed between the different pH buffers at early time points, the differences become drastic over prolonged periods of time.

Fig. 3.

Acid-mediated degradation of surfactant 7. (A) Schematic depicting surfactant degradation in a 1 : 1 mix of buffer and MeOH at 37 °C. (B) MALDI trace of 7 after treatment in 1 : 1 MeOH : CPB at pH 5.0 for 23 h. (C) pH dependent degradation of surfactant 7 in pH 7.4, 6.0, or 5.0 citrate phosphate buffer (CPB). Error bars represent the standard error of three experimental replicates. For assessment of the statistical significance of differences, a one-tailed student’s t-test assuming unequal sample variance was employed. Results were considered significant/not significant different per the following definitions: ns = p > 0.05, significant = p ≤ 0.05, * = p ≤ 0.05, ** = p ≤ 0.01.

After establishing a pH-dependent degradation mechanism, we then sought to investigate how this degradation affects PFC-in-water nanoemulsions, our chosen nanoparticle medium. To visualize the effect of surfactant degradation on the emulsions, we formulated emulsions using 7 or 8 and encapsulated fluorous rhodamine⁶⁷ into the PFC phase (Fig. 4A). By using this dye, we can easily see the emulsion destabilization through phase separation. First, we established that empty emulsions formulated from 7 or 8 were of similar size and polydispersity (PDI) by dynamic light scattering (DLS) (Fig. 4B). Then, emulsions containing fluorous rhodamine were formulated and resuspended in either CPB pH 7.4 or CPB pH 5.0 and left undisturbed at 37 °C. Gratifyingly, emulsions made from 7 did not show degradation until 21 h at pH 7.4, whereas the degradation is very noticeable at pH 5.0 as early as 3 h (Fig. 4C). Additionally, we observed no demulsification in either pH buffer from the emulsions stabilized with 8, suggesting that this mechanism is dependent on emulsion cleavage in a low pH environment. Zeta potentials of empty emulsions made from 7 show a pH dependent response, suggesting that the emulsions are more positively charged in low pH environments, likely due to the presence of hydrazine (Fig. 4D). This same response was not observed in emulsions made from 8. Empty emulsions were then monitored by DLS over a 21 h time frame where the size, PDI, and zeta potentials were tracked at pH 7.4, 6.0, and 5.0 (Fig. S6A–C†). Surprisingly, for any given condition there was little to no change in any of the parameters. This is particularly unexpected considering the degree to which we observe demulsification by eye. One possible explanation is that partially cleaved emulsions have flocculated out of solution and are thus immeasurable by DLS. SEC quantification of the remaining surfactant from the emulsions after 24 h showed 83% ± 1%, 33% ± 4%, and 25% ± 1% retention for pH 7.4, 6.0, and 5.0 buffers respectively (Fig. S6D†). Lastly, this assay was used to assess the stability of the nanoemulsions in commercial Dulbecco’s modified eagle medium (DMEM cell culture media without phenol red) at 37 °C, which was used as purchased and the pH was measured to be 8.4. Only a minimal amount of demulsification was observed with 7-stabilized emulsions after 21 h (Fig. S7†).

Fig. 4.

Acid-mediated degradation of surfactant 7 in emulsions. (A) Schematic showing the formulation of PFC-in-water nanoemulsions stabilized by either 7 or 8, followed by demulsification. (B) DLS intensity plots of PFC-in-water nanoemulsions formulated using 7 or 8. (C) Photos of fluorous rhodamine containing PFC-in-water nanoemulsions stabilized by 7 or 8 taken over 24 h. (D) Zeta potential measurements of emulsions stabilized by 7 or 8 in CPB at pH 7.4 or 5.0. Error bars represent the standard deviation of three measurements.

2.3. Payload release

Having established conditions for emulsion degradation, we sought to characterize the release of payloads encapsulated within nanoemulsions stabilized by 7 or 8. We began with the release of fluorous rhodamine dye in a partition assay against 1-octanol (a lipid membrane mimic) (Fig. 5A). Emulsions encapsulating fluorous rhodamine were diluted in CPB pH 7.4, 6.0, or 5.0 buffer, then 1-octanol was added on top. The biphasic solution was gently rocked at 37 °C and payload release was quantified by measuring the fluorescence observed in the 1-octanol layer. Noncleavable emulsions stabilized by 8 showed minimal release, with up to 11% ± 1% of the payload being released at 7 h and up to 22% ± 2% released after 42 h, with no apparent pH dependency (Fig. 5B). Cleavable emulsions stabilized by 7 on the other hand displayed a prominent pH-dependent payload release profile. At pH 7.4, the release was comparable to the noncleavable version, with 17% ± 1% cleavage at 7 h and 26% ± 1% cleavage at 42 h. At pH 6.0 the differences became more pronounced, with 24% ± 1% being released at 7 h and 32% ± 1% being released after 42 h. Lastly, at pH 5.0, 7-stabilized emulsions showed 25% ± 2% release at 7 h and 47% ± 1% release after 42 h, a drastic increase compared to the noncleavable counterpart. These results demonstrate the utility of this acid-cleavable surfactant for mediating selective payload release without the additional input of external stimuli.

Fig. 5.

Payload release of fluorous rhodamine from acid-cleavable emulsions. (A) Schematic showing the formulation of PFC-in-water nanoemulsions stabilized by either 7 or 8, followed by fluorophore partitioning into 1-octanol after surfactant cleavage. (B) Quantified release profiles of emulsions in CPB pH 7.4, 6.0, or 5.0 measured by fluorescence. Error bars represent the standard error of three experimental replicates. For assessment of the statistical significance of differences, a one-tailed student’s t-test assuming unequal sample variance was employed. Results were considered significant/not significant different per the following definitions: ns = p > 0.05, significant = p ≤ 0.05, * = p ≤ 0.05, ** = p ≤ 0.01.

3. Conclusion

We have successfully incorporated an acid-labile hydrazone linkage in between the blocks of a poly(oxazoline) diblock copolymer surfactant. The hydrophilic block of the surfactant was synthesized using a new, protected aldehyde initiator. After deprotection of both blocks, the surfactant was isolated using an unconventional emulsion purification technique which separated the polymer surfactant from the homopolymer impurities. The surfactant was first demonstrated to degrade in acidic media, as evaluated using SEC. A pH dependent trend in the degree of surfactant cleavage was observed using pH 7.4, 7.0, 6.0, and 5.0 buffer. This surfactant was then used to formulate perfluorocarbon-in-water nanoemulsions, where emulsion degradation trends were apparent. Lastly, fluorous rhodamine was encapsulated into the perfluorocarbon-in-water nanoemulsions and its release was measured in vitro using a 1-octanol partition experiment to mimic release into lipophilic cellular environments. The cleavable emulsions showed approximately double the payload release in pH 5 buffer than pH 7.4 buffer over 21 hours. These assays demonstrate the ability of the acid-cleavable surfactants to perform their function in the pHs relevant to the tumour microenvironment and late endosome.

In all, this work lays the foundation for future experiments targeting cancer and facilitating endosomal escape. This chemistry can be combined with readily established post-polymerization methods for appending biologically relevant cancer targeting small molecules^68–70 such as cyclic RGD tripeptide or folate. Furthermore, the catalogue of fluorous soluble diagnostic imaging agents and prodrugs is rapidly expanding, adding increased utility for perfluorocarbon-in-water nanoemulsions and elevating the value of having a degradable surfactant to enable payload release.^66,67,71,72 This work is an important step towards P(Ox)-based nanotherapeutics that can target cancer selectively and can escape the endosome before the payload is degraded by the lysosome.

Data availability

The data supporting this article have been included as part of the ESI.^†