Elucidating the effect of amine charge state on poly(β-amino ester) degradation using permanently charged analogs

Abstract

With synthetic ease and tunable degradation lifetimes, poly(β-amino ester)s (PBAEs) have found use in increasingly diverse applications from gene therapy to thermosets. Protonatable amines in each repeating unit impart pH-dependent solution behavior and lifetimes, with acidic conditions favoring solubility yet slowing hydrolysis. Due in part to these interconnected phenomena governing pH-dependent PBAE degradation, predictive degradation models, which would enable user-defined lifetimes, remain elusive. To separate the effects of charge state and solution pH on PBAE degradation, we synthesized poly(β-quaternary ammonium ester)s (PBQAEs), which differ from their parent PBAEs only by an additional methyl group, generating polymers with pH-independent cationic charge. Like PBAEs, PBQAE hydrolysis accelerates with increasing pH, though, at a given pH, PBAE degradation outpaces PBQAE degradation. This difference is more pronounced in basic solutions, suggesting deprotonated PBAE amines accelerate hydrolysis, providing an additional tuning parameter to PBAE lifetime and informing the degradation of PBAEs and other pH-responsive polymers.

Graphical Abstract

Authors are required to submit a graphic entry for the Table of Contents (TOC) that, in conjunction with the manuscript title, should give the reader a representative idea of one of the following: A key structure, reaction, equation, concept, or theorem, etc., that is discussed in the manuscript. Consult the journal’s Instructions for Authors for TOC graphic specifications.

As the need for diverse degradable polymers becomes increasingly essential,^1–4 we study a highly tunable class of polyesters: poly(β-amino ester)s (PBAEs). PBAEs boast an impressively facile synthesis, an expansive monomer library with a variety of functional groups, and tunable hydrolysis timescales.^5–13 Yet, the thorough understanding of PBAE degradation needed to construct predictive degradation models and realize these materials in applications ranging from therapeutic delivery^6,13–17 to porogenic sacrificial materials,^18,19 remains elusive, due in large to their complex pH-dependent solution and degradation behavior. Featuring a protonatable tertiary amine in each repeating unit, PBAEs become polycations in acidic aqueous environments, which improves polymer solubility and presumably water access to backbone esters. Meanwhile, basic environments accelerate degradation, despite reduced polymer solubility when amines are deprotonated and uncharged.²⁰ Isolating the effects of pH and amine charge state on PBAE degradation requires independent control of each, and would significantly advance our understanding of PBAE degradation to leverage their full potential in a range of application spaces.

As with PBAEs, the confounding effects of protonation state and degradation as a function of pH are difficult to separate with other polymers containing both amines and esters, for example poly(2-(dimethylamino)ethyl acrylate)s (PDMAEAs). Some prior work suggests that PDMAEAs degrade via a self-catalytic mechanism involving interactions of amines with esters in addition to more traditional acid- or base- catalyzed hydrolysis.^21–24 In an effort to separate pH and charge effects on ester hydrolysis, Monteiro and coworkers quaternized PDMAEAs, yielding cationic charge regardless of pH, and observed the esters to hydrolyze slower than those in tertiary amine-containing PDMAEAs and the PDMAEA hydrolysis process to be pH-independent.²⁴ However, the concentration of polymer amine/ammonium groups (ca. 190 mM) exceeded the concentration of buffer used to control pH (150 or 80 mM buffering salts). As protonatable tertiary amines would likely raise solution pH more than their quaternized counterparts, the PDMAEA esters may have hydrolyzed faster simply because the solution was more basic than the quaternized PDMAEA solution. Later work by Stöver and coworkers noted the low buffer concentration used by Monteiro and coworkers and found PDMAEA ester hydrolysis to be highly pH-dependent when pH was well-controlled with buffer, supporting the absence of pH control in the Monteiro work. Considered together, these results highlight our limited understanding of the role amine charge state plays in amine- and ester- containing polymer hydrolysis.²³

Motivated to understand PBAE degradation, we study the effects of amine charge state on PBAE degradation as a function of solution pH, where adequate buffer is present in all cases to control pH (10:1 buffering salts: amines/ammoniums). We envision multiple scenarios in which deprotonated amines, present on PBAE chains in increasing concentration in more alkaline environments, influence ester hydrolysis. Deprotonated PBAE amines may accelerate ester hydrolysis via interactions with esters, slow hydrolysis owing to reduced polymer solubility and water access to backbone esters, or have no effect if ester cleavage is largely controlled by an acid- or base- catalyzed mechanism. As with PDMAEAs, PBAE tertiary amines can be quaternized to pH-independently cationic quaternary ammoniums via the addition of methyl iodide. Zheng and coworkers demonstrated permanently cationic quaternized versions of PBAEs bind siRNA more efficiently than their parent PBAEs, despite evidence of ester cleavage during quaternization reducing polymer molecular weight.²⁵ Modifying the Zheng procedure to avoid ester cleavage during modification, we synthesized analogous poly(β-quaternary ammonium ester)s (PBQAEs) that differ from their parent PBAEs only by the presence of a quaternary ammonium group in place of the traditional PBAE backbone tertiary amine (Scheme 1, see SI Section S1 for experimental details). Here, we examine the hydrolytic behavior of these pH-independently cationic PBQAEs compared to their parent pH-responsive PBAEs to isolate the roles of pH and amine charge state on PBAE degradation. We further anticipate these materials will be useful when cationic charge in alkaline environments is needed.

Scheme 1.

PBAE and PBQAE synthesis

While the facile nature of PBAE synthesis has been recognized, we feel we cannot overemphasize its simplicity: combining two monomers and heating their mixture without the need for a low oxygen environment or solvent yields ca. 5–15 kDa polymers after 24 h with limited-to-no byproducts and no required catalyst or solvent removal (Figure 1A). The step-growth nature of PBAE synthesis is not conducive to precise molecular weight control and narrow dispersities (Ð), as evidenced by the range of number-average molecular weights (M_n) determined from analytical size exclusion chromatography (SEC), relative to poly(methyl methacrylate) standards, of PBAEs from three independent synthesis batches (M_n = 11.9 kDa – 14.0 kDa, Ð = 1.8, Table S1, Figure S12). However, if needed, narrow molecular weight distributions can be accessed by fractionation using preparative SEC (SI, Section S2 for more details).²⁶

Figure 1.

¹H NMR (600 MHz) spectra of: (A) PBAE in the deprotonated state (directly from synthesis); (B) PBAE in the protonated state; and (C) PBQAE after purification. Vertical dashed lines are drawn as guides to the eye to emphasize the downfield shift of a and i upon amine protonation or quaternization.

PBQAE synthesis from PBAEs proceeds in yet another facile, straightforward reaction. We modified a procedure described by Zheng and coworkers to reduce ester hydrolysis during modification by using an aprotic solvent, tetrahydrofuran (THF, as opposed to isopropanol), and a lower reaction temperature (ambient room temperature, as opposed to 60 °C).^25,27 Dissolving PBAE in THF (100 mg/mL) with MeI (2 molar equivalents per backbone amine) yields a precipitated PBQAE after ca. 18 h. The precipitated polymer was purified via preparative SEC similarly to a method we previously reported for other post-polymerization modifications of PBAEs to remove remaining THF and unreacted MeI (SI Section S1 and S2).²⁶ The resultant polymer exhibited a similar analytical SEC retention time to its parent PBAE (Figure 2A), which, along with the absence of characteristic carboxylic acid peaks in the infrared spectra, supports negligible cleavage of the sensitive backbone esters during modification and purification (Table S1, Figure S12, and Figure S24), with quantitative amine to ammonium conversion as determined by ¹H NMR spectroscopy (by the relative integrations of peaks labeled i and j in Figure 1C).

Figure 2.

PBAE and PBQAE comparison: (A) analytical SEC chromatograms (signal from refractive index detector) of PBAEs and PBQAEs eluting at the same retention time, indicative of similar molecular weight; (B) reverse-phase HPLC chromatograms of PBAEs in the protonated state and PBQAEs eluting at the same retention time, indicative of similar hydrophobicity; and (C) titration curves of PBAEs and PBQAEs, where pH is monitored while known amounts of NaOH_(aq) (0.2 M added at 1 mL/min, reported as mL NaOH/g sample) are dosed into the polymer solutions.

Protonated PBAEs exhibit distinct NMR spectroscopy chemical shifts compared to PBAEs in their deprotonated state, as we reported previously,²⁶ and shown in the distinct NMR spectra (Figure 1A and B) of deprotonated PBAEs, obtained directly from synthesis, and protonated PBAEs, accessed by dissolving PBAE in water, acidifying the solution with HCl_(aq), and lyophilizing the polymer solution. We suspect the reduced electron density on protonated amines deshields neighboring protons causing downfield shifts. These downfield shifts are most apparent in the peaks corresponding to protons nearby the pH-dependent amine, labeled f, g, h, and i in Figure 1A and B (see SI, Section S3–S6 for 2D NMR spectra used to determine peak assignments and full ¹H and ¹³C NMR spectra). More subtlety, the protons labeled a, located six bonds away from the amine, shift downfield slightly when PBAE amines are protonated. As with the protonated PBAEs, the reduced electron density near the PBQAE ammonium deshields nearby protons causing a downfield shift of a, f, g, h, and i (Figure 1C). The ¹H NMR chemical shifts of the PBQAE more closely resemble the protonated PBAE spectrum than the deprotonated PBAE spectra (e.g., a, g, and i in Figure 1B and C), further supporting the chemical similarities between protonated PBAEs and their pH-independent derivative, PBQAEs. Yet, since the additional methyl group may add hydrophobic character, which can be a controlling variable in PBAE degradation,^5,6,9,26 we used reverse-phase high performance liquid chromatography (HPLC) to examine differences in PBAE and PBQAE hydrophobicity. We found our protonated PBAEs and PBQAEs had similar elution times in reverse phase high performance liquid chromatography, indicative of similar hydrophobicities (Figure 2B and Figure S22). Together, spectroscopy and chromatography point to successful functionalization of PBAEs into PBQAEs with minimal degradation during synthesis or hydrophobicity change and thus, we can isolate any changes we see in later degradation experiments to amine charge state.

While PBAEs and PBQAEs differ structurally only by the addition of a methyl group, we expected the pH-dependent charge state of PBAEs to be absent in PBQAEs as the quaternary ammonium will remain charged across a broad pH range. We examined the pH-responsive behavior of both polymers by titrating polymer solutions with NaOH_(aq) (0.2 M, Figure 2C).^28,29 To ensure any buffering behavior results from the polymer and not changes in dissolved CO₂ content in the titrant or analyte solutions, we performed titrations of a non-buffering salt, NaCl, before and after each polymer titration all of which showed nearly identical profiles indicative of minimal CO₂ content change over the course of these experiments (Figure S14). Starting in an acidic environment (pH = 2.5), where PBAEs are protonated and most readily soluble, PBAE amines donate their protons to the solution as NaOH_(aq) is added, buffering the solution (Figure 2C). PBQAEs ammoniums however, lack the ability to accept or donate protons and thus, buffer solution pH far less than their parent polymers, indicated by the faster pH increase compared to PBAEs as both polymer solutions are dosed with NaOH_(aq). Both polymers exhibit some buffering, demonstrated by the gradual pH increase between the initial and final plateaus in comparison to non-buffering NaCl_(aq), but the pH of PBQAE solution increases faster as NaOH_(aq) is added than the PBAE solution, indicating PBQAEs buffer fewer protons than PBAEs. The small buffering capacity of PBQAEs may stem from the production of deprotonatable carboxylic acids as esters hydrolyze, as observed during titrations based on infrared spectroscopy (Figure S13).^30–35 Though we cannot prevent hydrolysis from occurring during titration, which occurs rapidly at elevated pH, we ensure the dosing rate of NaOH_(aq) is constant (1 mL/min) to keep degradation relatively constant between samples. Alternatively, perhaps the cationic nature of PBQAEs may draw higher local concentrations of hydroxide ions, particularly at low pH, and therein decrease the measured bulk solution pH. The reduced buffering capacity of the PBQAE compared to its parent PBAE supports our synthesis of PBAE analogs, PBQAEs, whose charge state is independent of solution pH.

To determine how the differences in pH-dependent charge state manifest in degradation behavior, we monitored ester hydrolysis with ¹H NMR spectroscopy. We studied degradation in 90% H₂O and 10% D₂O to avoid making assumptions associated with the estimation of pD (the deuterated analog of pH), using a 1D NOESY pulse sequence to suppress the water signal.^36,37 These experiments were conducted in 150 mM buffers (sodium citrate for pH 3, potassium phosphate for pH 6 and 7, and sodium carbonate for pH 10) at a polymer concentration of 6 mg/mL (corresponding to ca. 15 mM amines or ammoniums) to ensure control of solution pH throughout the experiments.²⁰ To verify the buffer concentration was sufficient to maintain pH control throughout the degradation experiments so as to attribute any differences in degradation behavior to amine charge state, we measured pH 23 h after the PBAE and PBQAE were dissolved separately in pH 7 buffer, finding the measured pH of the solutions was 6.96 and 6.98, respectively.

While several peaks increase, decrease, or shift during hydrolysis, we monitored the integration of the protons adjacent to the ester (labeled a in Figure 1) at 4.24–4.37 ppm to track hydrolysis (Figure 3, Figure S17–S21), as done by others and in our previous work.^26,38 Like PBAEs, PBQAE hydrolysis accelerates with increasing pH, though, at a given pH, PBAE degradation outpaces PBQAE degradation. This difference between PBAE and PBQAE hydrolysis grows in alkaline environments; Figure 3B shows the percentage of remaining esters in each polymer at the 24 h timepoint (from the data points highlighted in the box of Figure 3A). In our previous work, we demonstrated basic pH environments to accelerate PBAE hydrolysis,²⁰ while here, we find the PBAE analogs, PBQAEs, with pH-independent charge states to degrade slower than traditional PBAEs, despite the potential for charged groups to increase solubility.

Figure 3.

PBAE and PBQAE pH-dependent degradation tracked by ¹H NMR (800 MHz) spectroscopy: (A) esters remaining as a function of time for PBAEs and PBQAEs degrading in 150 mM pH 3, 6, 7, or 10 buffers. All points are from one representative batch of PBAE and their PBQAE analog (see SI Section S10 and S11 for other batches); and (B) esters remaining at the 24 h timepoint for PBAEs and PBQAEs in 150 mM pH 3, 6, 7, or 10 buffers averaged over three independent synthetic batches. Error bars represent standard deviation (n = 3).

To understand the source of the differences in PBAE and PBQAE degradation, we considered the potential roles deprotonated PBAE amines may play in ester hydrolysis. First, we studied the ability of secondary amines, present on PBAE chain ends, to lyse esters. Aminolysis would lead to an amide-containing degradation product instead of the carboxylic acid degradation product expected from hydrolysis. Using infrared spectroscopy, we verified the absence of characteristic amide peaks in degrading PBAEs (Figure S24), discrediting aminolysis as the path for faster PBAE degradation compared to PBQAEs. Next, to ensure minimal changes to bulk solution pH, we measured pH after a degradation experiment that started at pH 7, and find the bulk solution pH to be 6.96 and 6.98 (within the margin of error of these measurements) for PBAE and PBQAE solutions, respectively, debunking bulk solution pH changes as the cause for PBAEs to degrade faster than PBQAEs. Though bulk pH remains constant during degradation, the experiments presented here lead us to suspect the deprotonated PBAE amines, situated only two bonds away from the labile esters, may locally increase pH or lead to changes in counterion condensation or water structure, accelerating hydrolysis (Scheme 2).^39–43 Additionally, changes in the polymer electronic environment, indicated by downfield shifts in NMR peaks upon protonation or quaternization, notably persists several bonds away from the amines/ammoniums (e.g., protons labeled a in Figure 1, located six bonds away from amines/ammoniums). These electronic environment changes or amine-catalyzed ester hydrolysis (as has been proposed for other amine and ester containing polymers)^21–24 may underlie differences in polymer behavior as a function of charge state. Likely, some combination of these factors – local pH,^41,44 counterion condensation,^42,43 water organization changes,^45–49 local dielectric changes,²¹ electron distribution, and/or amine-catalyzed hydrolysis^21–23 – results in the faster degradation of PBAEs relative to their charged, quaternary analogs.

Scheme 2.

Proposed reasoning for amine charge state effects on PBAE degradation

In summary, we find the amine charge state to influence PBAE degradation in solutions with well-controlled bulk pH, with permanently cationic PBQAEs degrading slower than their parent pH-dependent PBAEs. Since we observe these differences when bulk properties are constant, we suspect PBAE tertiary amines may act to change local properties (e.g., pH, water structure, counterion concentration) or otherwise catalyze PBAE ester hydrolysis. The slower degradation of PBQAEs relative to their parent PBAEs provides an additional polymer lifetime tuning parameter in addition to facilitating cationic charge in alkaline environments, if needed. By furthering our understanding of the phenomena involved in PBAE and PBQAE degradation as well as the role of amine charge state on ester hydrolysis, we hope this work informs future predictive degradation models, a step towards realizing these versatile polymers as sacrificial materials, drug delivery vehicles, and degradable alternatives to single-use plastics. Moreover, these findings may inform the degradation behavior of other emerging hydrolysable polymers with basic groups, including poly(α-amino ester)s,⁵⁰ poly(ester urea)s,⁵¹ and poly(2-[dimethylamino]ethyl acrylate)s and methacrylates.^21,23,24,52