Concentration Threshold for Membrane Protection by PEO-PPO Block Copolymers with Variable Molecular Architectures

Abstract

Poloxamer 188, a poly(ethylene oxide)-b-poly(propylene oxide)-b-poly(ethylene oxide) (PEO-PPO-PEO) triblock copolymer, protects cell membranes in several injury models. However, the nature of the copolymer/membrane interaction and the mechanism of membrane protection remain unknown. Systematic variations of the block copolymer architecture – including PPO-PEO-PPO triblocks and PPO-PEO diblocks – were used to probe the mechanism and evaluate the potential for alternative architectures to yield superior protection. To test the polymers, murine myoblasts were subjected to an osmotic stress, and membrane integrity was quantified by measuring lactate dehydrogenase (LDH) leakage. These experiments exposed a concentration threshold effect where all tested polymers reach 50% leakage of LDH compared to a non-treated buffer only control over a narrow concentration range of 0.8–4 μM. Differences in polymer protection at lower concentrations indicate that protection increases with the PPO-PEO-PPO molecular architecture and increasing hydrophobicity.

Graphical Abstract

1. INTRODUCTION

Amphiphilic poly(ethylene oxide)-b-poly(propylene oxide)-b-poly(ethylene oxide) (PEO-PPO-PEO) triblock copolymers, referred to as Pluronics or Poloxamers, have been shown to stabilize cellular membranes in vitro and in vivo.^1,2 Poloxamer 188 (P188; M_n = 8.4 kDa, 79 wt% PEO) is the only Poloxamer to receive FDA approval and has been shown to protect against injury across multiple animal models, including porcine studies involving acute cardiac failure³ and ischemia/reperfusion injury,^4,5 and also in dysferlin- and dystrophin- deficient (dmx) mice and dogs where P188 showed protection against ventricle dilation,⁶ muscular atrophy,⁷ ischemia/reperfusion,⁴ declining respiratory function,⁸ cardiomyopathy,⁹ repetitive muscle torque,^1,10 and laser induced injury.¹¹ The protective properties of P188 extend further to other muscular systems and beyond to other tissue types, such as electrical and ionization injury in skeletal muscle,^12,13 impact injury of cartilage, connective tissue, and spinal cord,^14–17 and ischemic-reperfusion injury in neurons and lung tissue.^18,19 Although PEO homopolymer lacks the hydrophobic PPO block of P188, it has a degree of lipophilic character which offers some protection against hypoxia/reoxygenation injury in cardiomyocytes,^19,20 stabilization of membranes of neural cells subjected to mechanical injury and oxidative stress in vivo,^21–25 and protection of intestinal epithelial cells against radiation by interacting with lipid rafts.²⁶

In vitro and in silico work has aimed to enlighten the mechanism of protection and identify molecules with superior protection efficacy. P188 was shown to insert into phospholipid monolayers in a Langmuir trough at low surface pressure.²⁷ P188 and PEO (M_n = 20 kDa) reduced dye leakage from giant unilamellar vesicles under osmotic stress.²⁸ Protection of large unilamellar vesicles from peroxidation stress was observed for P188 and PEO (M_n = 8 kDa) but not for poloxamers with substantially higher PPO content.²⁹ Similarly, neurons were protected from oxygen glucose deprivation by block copolymers of PPO and poly(methacryloyloxyethyl phosphorylcholine) with low PPO fraction but not by copolymers with high PPO fraction, although the latter exhibited higher surface pressure on phospholipid monolayers.³⁰ Increased phospholipid interaction from hydrophobic block copolymers was also revealed by pulsed field gradient-nuclear magnetic resonance (NMR) experiments on liposomal lipid bilayers.³¹ The importance of hydrophobicity was shown through the addition of a hydrophobic tertiary-butyl (t-butyl) end group to the PPO block of a PPO-PEO diblock copolymer, which aided protection of myoblasts from osmotic stress^1,32 and rescued muscle contractility in dystrophic mice.¹ In silico molecular dynamics simulations indicate the t-butyl end group resides deeper in the membrane at the timescale of simulation (110 ns).^1,33 In vitro osmotic stress experiments on C2C12 myoblasts also highlight the need for the PPO block to contain 9 to 16 repeat units to provide protection and a longer PEO block to enhance protection.³² These results lead to the hypothesis of an “anchor and chain” mechanism acting at short times, where the hydrophobic chain and its end group are buried in the membrane thereby avoiding unfavorable interactions in the water phase outside the lipid membrane.^1,32 The proposed mechanism motivates study of additional molecules to further evaluate the role of the hydrophobic components. Neutron reflectivity experiments showed that both PEO and PPO partially reside inside the phospholipid bilayer after incubation at high (4.5 mM) polymer concentrations.³⁴ Surface plasmon resonance measurements also indicated that PEO homopolymer binds to a model membrane similarly to P188.³⁵ Moreover, PEO is protective in select settings albeit less potently than PPO-based block copolymers.^19–26 Collectively, this work highlights the significance of a hydrophilic-lipophilic balance. The hydrophobic PPO block enhances engagement of the polymer with the hydrophobic lipid tail increasing affinity and therefore protection. However this protection cannot be accomplished solely by PPO as PEO plays a key role in protection and in avoiding the detrimental solubilization effect associated with high fractions of PPO.^1,28–31,34

Notwithstanding the wealth of data regarding the action of P188 and a proposed mechanism for its interaction with model membranes, how this translates to in vivo efficacy remains an open question. To evaluate polymers in vitro, we have developed an assay that employs osmotic stress to inflict a controlled amount of damage to C2C12 murine myoblasts and detection of lactate dehydrogenase (LDH) enzyme leakage for gauging membrane integrity.^1,32,36 This assay balances moderate biological complexity with relative experimental efficiency to provide a powerful tool for determining the effects of polymer composition, molecular weight, and architecture on cell protection, offering an efficient screening tool along with insight into the mechanism of action. Previously, PEO-PPO diblocks containing various end groups, and different PEO and PPO block lengths, were explored. This study expands on that work, introducing a new “inverted” PPO-PEO-PPO triblock copolymer architecture, along with a larger set of PPO-PEO diblocks with variable end groups, and several PEO homopolymers. The assay is applied across a wide range of block copolymer concentrations (0.7 to 1200 μM) in aqueous buffer, ranging from conditions offering no protection to virtually complete protection as evidenced by the extent of LDH release relative to control experiments. This range of measurements includes estimated blood concentrations associated with previously published studies involving administration of P188 to humans and animals.

Under the premise that the hydrophobic PPO block aids membrane interaction, an additional hydrophobic block may strengthen overall interaction via an avidity effect.³⁷ For example, Lu et al. showed that a poly(styrene)-block-poly(ethylene-alt-propylene)-block-poly(styrene) (PS-PEP-PS) triblock copolymer dispersed as spherical micelles in squalene, a selective solvent for PEP, exhibited a rate of molecular exchange between micelles that was nine orders of magnitude slower than for a PEP-PS-PEP triblock.³⁸ This dramatically different exchange rate was attributed to the need to simultaneously eject both PS blocks in PS-PEP-PS from the poly(styrene) micelle cores whereas only one PS block had to be removed with PEP-PS-PEP.³⁸ We hypothesized that inverting the structure of the P188 triblock to PPO-PEO-PPO would enhance interaction with cell membranes, analogous to the mechanism described by Lu et al.³⁸ Furthermore, we speculate that if a polymer associates with the membrane longer and more strongly there will be an increase in protection efficacy. Therefore, a suite of a triblocks (PEO-PPO-PEO and PPO-PEO-PPO), diblocks (HO-PPO-PEO-Me and t-butyl-PPO-PEO-Me), and three PEO homopolymers were investigated to reveal the effects of molecular architecture, composition, and hydrophilic-lipophilic balance on cell protection based on the LDH assay. PEO homopolymers were added to this study to establish the effects of the hydrophilic block. Varying the polymer dosage in this study allows for differentiation in polymer performance across these variables and comparison of the effect of polymer concentration in the osmotic stress/ LDH in vitro assay to the previously reported in vivo performance of P188.

2. MATERIALS AND METHODS

2.1. Materials.

Pluronic P188 was donated by BASF (Wyandotte, MI). Dihydroxyl PEO homopolymers (M_n = 8,000 and 20,000 Da) were purchased from EMD Millipore (Billerica, MA). Dihydroxyl PEO (M_n = 7,800 Da) was purchased from Polymer Source (Dorval, Montreal) and used in the synthesis of the inverted PPO-PEO-PPO triblock copolymer. For the synthesis of methoxy-PEO-PPO-hydroxyl chain end functionalized diblocks (denoted Me-PEO-PPO-OH), poly(ethylene glycol) methyl ether (Me-PEO-OH) with M_n = 2,000 and 5,000 Da were purchased from Polymer Source (Dorval, Montreal) and EMD Millipore (Billerica, MA), respectively. Monomers (propylene oxide and ethylene oxide, purity >99.0%), potassium tert-butoxide, 18-crown-6 ether, potassium metal and naphthalene were purchased from Sigma-Aldrich. Bioreagent grade buffer salts (sodium chloride, potassium chloride, calcium chloride, magnesium chloride, and 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)) and Triton X-100 were purchased from Sigma-Aldrich.

2.2. Block Copolymer Synthesis.

Anionic polymerization was employed to make PPO-PEO-PPO triblocks, Me-PEO-PPO-OH diblocks, and t-butyl-PPO-PEO-OH diblocks. The synthesis procedures are described in previous publications.^39–41 All polymerizations were performed under an inert argon environment free of water and air. Tetrahydrofuran (THF), dried using an alumina column, was employed as the solvent in all cases. In order to maximize conversion and reduce side reactions PPO was prepared in the presence of 18-crown-6 ether at a 2:1 molar ratio relative to potassium ions.^42,43 Synthesis of PPO-PEO-PPO began with freeze-dried, di-hydroxyl terminated PEO homopolymer (HO-PEO-OH), which was reacted with potassium naphthalenide. Propylene oxide (PO) was then added and reacted at 25 °C for 72 hr. In a similar process, Me-PEO-PPO-OH diblock copolymer was synthesized by initiating Me-PEO-OH with potassium naphthalenide, followed by addition of propylene oxide at 25 °C and allowed to react for 72 hr. Acidic methanol (10:1 methanol:37 w/w% hydrochloric acid) was used to terminate all polymerizations. Crown ether and salt complexes were removed through repeated cycles of solvent evaporation, dissolution in THF, and filtering along with preparatory size exclusion chromatography (SEC) with PD-10 desalting columns. Synthesis of t-butyl-PPO-OH was accomplished by initiation and polymerization of propylene oxide with potassium t-butoxide followed by termination with acidic methanol. Crown ether and salt complexes were removed by repeated rotary evaporation and washing with hexane/water. The t-butyl-PPO-OH was then reinitiated with potassium naphthalenide, and ethylene oxide was added and reacted at 40 °C for 48 hr followed by termination with acidic methanol to yield t-butyl-PPO-PEO-OH. The product was purified by iterative filtration and preparatory SEC.

2.3. Molecular Characterization.

The molecular characteristics of the polymers employed in this study were determined using three methods: nuclear magnetic resonance (NMR) spectroscopy, SEC, and matrix assisted laser desorption/ionization time-of-flight (MALDI TOF-TOF) mass spectrometry. A 400 MHz Bruker Avance III HD nanobay AX-400 was employed to obtain NMR spectra of the polymers in deuterated chloroform. Block copolymer composition and molecular weight were determined by peak analysis and integration. Polymer dispersity (Đ) was determined by SEC with THF as the mobile phase, refractive index detection, and calibrated with polystyrene standards. Molecular weight and Đ also were confirmed by MALDI TOF-TOF mass spectrometry (TOF 5800, Sciex, Washington, DC) with α-cyano-4-hydroxycinnamic acid (Sigma) as the matrix. Characterization data are presented in the Supplemental Information and summarized in Table 1.

Table 1.

Summary of polymers used in this study.

Polymer			PEO units	PPO units	Mn (g/mol)	Ð	%EO
Triblock Copolymer	E75P30E75†	HO-PEO75-PPO30-PEO75-OH	150	30	8400	1.07	79
	P15E200P15	HO-PPO15-PEO200-PPO15-OH	200	30	10700	1.05	83
Diblock Copolymer	P16E46	HO-PPO16-PEO46-CH3	46	16	3000	1.06	69
	P17E112	HO-PPO17-PEO112-CH3	112	17	6000	1.03	83
	P14E189	HO-PPO14-PEO189-CH3	189	14	9200	1.03	91
	P32E120	HO-PPO32-PEO120-CH3	120	32	7200	1.06	74
	P27E189	HO-PPO27-PEO189-CH3	189	27	9700	1.03	84
	t-P16E75	t-butyl-PPO16-PEO75-OH	75	16	4200	1.06	79
	t-P16E180	t-butyl-PPO16-PEO180-OH	180	16	8900	1.07	90
Homopolymer	E114ỻ	HO-PEO114-CH3	114	-	5000	1.06	100
	E181ỻ	HO-PEO181-OH	181	-	8000	1.06	100
	E455ỻ	HO-PEO455-OH	455	-	20000	1.06	100

^†Pluronic 188 provided by BASF,

^ỻPEO purchased from Millipore

2.4. Cell Culture.

C2C12 myoblasts (C2C12; American Type Culture Collection, Manassas, VA) were maintained in growth media containing high glucose Dulbecco’s modified eagle medium (DMEM), 20% fetal bovine serum, and 1% penicillin streptomycin (Gibco Invitrogen, Grand Island, NY) at pH 7.2, 37 °C, and 5% CO₂ in a humidified incubator. Cells were passaged once they reached 70% confluency to avoid differentiation into myotubes. Myoblasts after more than 10 passages were used in the cellular assays.³²

2.5. Osmotic Stress and LDH Release Assay.

The osmotic stress assay was adopted from Kim et al.³² Approximately 10⁴ cells/well were plated in a tissue culture-treated 96 well-plate and grown to 50–80% confluency. Two buffer solutions were prepared: 330 mOsm isotonic (140 mM NaCl, 5 mM KCl, 2.5 mM CaCl₂, 2 mM MgCl₂, and 10 mM HEPES) and 75 mOsm hypotonic (20 mM NaCl, 5 mM KCl, 2.5 mM CaCl₂, 2 mM MgCl₂, and 10 mM HEPES). The buffers used in the assay were adjusted to a pH of 7.2 by addition of NaOH. Polymers were added to the buffers at the desired concentration (between 0.7 and 1200 μM) and filtered for sterilization. The cells were kept at 37 °C and 5% CO₂ in a humidified incubator during stress. At the start of the assay, the growth media is removed and replaced with isotonic buffer for 30 minutes followed by hypotonic buffer for 50 minutes and then isotonic buffer for 30 minutes (Figure 1). Cells were then lysed with Triton X-100 for 50 minutes. The solution recovered at each step was stored at 4 °C before measuring the LDH concentration.

Figure 1.

Timeline of osmotic stress assay. Four collection time points are as marked. Illustrations depict a cell and the associated release of LDH at each stage of exposure to the indicated buffer.

Lactate dehydrogenase (LDH) is an intracellular enzyme whose leakage out of the cell is indicative of a decrease in membrane integrity.^44,45 LDH content in each sample was measured by monitoring LDH-catalyzed reduction of nicotinamide adenine dinucleotide (NAD) to NADH (LDH Assay, Pointe Scientific Inc., Canton, MI) at 37 °C. The rate of reduction to NADH was measured by light absorbance at 340 nm. The amount of LDH released at each step was divided by the total LDH content across all steps and normalized to the non-treated, buffer-only control. Some unprotective samples exhibit LDH release nominally in excess of 100% when normalized to buffer control, but this is not statistically significantly different from the control.

2.6. Statistics.

LDH results are presented as the mean ± standard deviation. Each data point in Figures 4–7 represents 9 replicates of the specific condition (3 biological replicates repeated 3 times). Multigroup comparisons were calculated with one-way analysis of variance with Tukey post hoc test. Graphical and statistical analyses were completed using Prism (GraphPad Software).

Figure 4.

Increased valency of PPO provides increased protection efficacy. A) In the isotonic recovery, inverted triblock architecture has statistically lower intermediate LDH release than P188 and the diblock equivalent. B) The protection of the inverted molecule benefits from the added PPO valency as the PPO-PEO-PPO structure has lower LDH release than PEO-PPO diblocks. C) In contrast, the PEO-PPO-PEO structure of P188 does not offer increased protection compared to PEO-PPO diblocks. (Statistical difference of p <0.05 between two polymers are noted as: * E₇₅-P₃₀-E₇₅ and P₁₅-E₂₀₀-P₁₅, †P₂₇-E₁₈₉ and P₁₅-E₂₀₀-P₁₅, _ P₁₇-E₁₁₂ and P₁₅-E₂₀₀-P₁₅, ‡ P₁₄-E₁₈₉ P₁₅-E₂₀₀-P₁₅, θ E₇₅-P₃₀-E₇₅ and P₁₇–₁₂₀) Several polymers are presented in multiple panels (with identical data) to facilitate focused comparisons.

Figure 7.

Increased hydrophobicity enhances protection. A and B) Protection efficacy increases with increasing the amount of PPO added to a PEO homopolymer. C) With a 15 unit PPO block there is limited benefit on shortening the PEO block. D) With 30 unit PPO block, shortening the PEO block decreases LDH leakage in intermediate concentrations. (Statistical difference of p <0.05 between two polymers are noted as: †E₁₈₁ and P₂₇E₁₈₉, * E₁₈₁ and P₁₄E₁₈₉, ‡ E₁₁₄ and P₃₂E₁₂₀, _ E₁₁₄ and P₁₇E₁₁₂, θ P₁₇E₁₁₂ and P₃₂E₁₂₀, β P₁₇E₁₁₂ and P₁₄E₁₈₉, λ P₁₆E₄₆ and P₁₄E₁₈₉, φ P₃₂E₁₂₀ and P₂₇E₁₈₉) Several polymers are presented with data also shown in previous figures to facilitate focused comparisons.

3. RESULTS

3.1. Polymer Synthesis and Characterization.

This study employed a suite of polymers varying in the number of blocks (triblock, diblock, or homopolymer), sequence of blocks, number of EO and PO units, and end group of the PPO block (Table 1 and Figure 2). Molecules are abbreviated as P for PPO and E for PEO each followed by the number of repeat units. For example, the P188 chemical structure is abbreviated E₇₅P₃₀E₇₅. Note all diblocks, P_xE_y, correspond to Me-PEO-PPO-OH unless labelled as t-butyl-P_xE_y which indicates t-butyl-PPO-PEO-OH. The number of repeat units per block and dispersity were characterized using NMR, SEC, and MALDI-TOF-TOF (Figures S1, S2, and S3) as listed in Table 1.

Figure 2.

Chemical structures of polymers: A) commercial Pluronic triblock copolymer (PEO-PPO-PEO), B) inverted triblock copolymer (PPO-PEO-PPO), C-D) PEO-PPO diblock copolymer with C) hydroxyl end-group on PPO or D) t-butyl end-group on PPO, E) PEO homopolymer. m and n indicate the degrees of polymerization of the PEO and PPO blocks, respectively. Blue signifies the hydrophilic block and red the hydrophobic block.

In previous work, E₇₅P₃₀E₇₅ and t-P₁₆E₄₆ (a more hydrophobic variant of the most hydrophobic molecule in the current study) did not exhibit micelle formation at 150 μM (Kim et al.), which is well above the functional concentrations observed in the current study.³² Moreover, with relations given in Alexandridis et al., the critical micelle concertation (CMC) of P₁₆E₄₆ can be approximated as 1400 μM.⁴⁶ All experiments in the current study are below this concentration. Notably, lysis is not observed from any of the polymers in our study at the concentrations tested (up to 1200 μM as chosen for physiological relevance; Figure S4).

3.2. Stoichiometric Limitations.

In preparation to vary concentration to differentiate between polymer protective potency, we evaluated stoichiometric limitations in the cellular assay. To estimate polymer coverage of the cells, we assumed the area of coverage can be represented by a circle with a radius equivalent to the polymer radius of gyration in solution; the available cell membrane is approximated to be the area of one well of the 96-well plate. While surface roughness would provide additional area, subconfluency partially counters this effect. More importantly, only an order of magnitude estimate is needed. These estimations give a coverage of 300% (assuming all the polymer in solution is available for adsorption) at the lowest polymer concentration and lowest molecular weight employed. Accordingly, membrane interaction should not be stoichiometrically limited. In order to validate this point, we measured LDH release from osmotically stressed C2C12 myoblasts in the presence of P188 (E₇₅P₃₀E₇₅) and E₁₈₁ under two scenarios: 1) vary the polymer solution volume added to a constant confluency of cells, and 2) vary the confluency of cells with a constant polymer solution volume. Quadrupling the volume of polymer solution (50 to 200 μL) does not substantially improve protection (Figure 3A), consistent with no significant stoichiometric limitation. Likewise, decreasing cell confluency from 50% to 30% does not aid protection (Figure 3B), suggesting that polymer availability is not a limiting factor. Yet when examining cells with 90% confluency, the polymer is unable to protect as well as at lower concentrations, which could be due to membrane availability or biochemical differences at high confluency (Figure 3B). Therefore, we performed studies with 100 μL buffer volume and 50% cell confluency. With these stress parameters, the non-treated cells release 29 ± 3% of the total LDH during the 110-minute stress period (Figure S5) and maintain 95 ± 3% viability immediately following treatment, which falls to 28 ± 3% viability 24 h later (Figure S6).

Figure 3.

Polymer protection efficacy is not stoichiometrically limited but does have limitations when cells are confluent. All polymer concentrations are 3 μM and the LDH release refers to the isotonic recovery period as compared to the non-treated buffer control (Figure S7), noted as % non-treated. A) No clear trend is seen to indicate added volume results in added protection. B) At 50% and 30% confluency, the polymer is protective. At 90% confluency protection of the cell membranes is limited. * p <0.001.

3.3. Role of Molecular Architecture.

We tested the impact of block architecture on cellular protection from osmotic stress by inverting the E₇₅P₃₀E₇₅ architecture to P₁₅E₂₀₀P₁₅ as well as assessing the compositionally equivalent diblock P₂₇E₁₈₉. C2C12 myoblasts were exposed to the stress of hypoosmotic buffer and isotonic recovery in the presence of polymer, and macromolecular leakage during isotonic recovery was quantified to assess loss of membrane integrity. P₁₅E₂₀₀P₁₅ (43% LDH release) provided more potent protection from LDH leakage at a lower concentration (ca. 1.4 μM) than either E₇₅P₃₀E₇₅ (96%, * p <0.01, f=11.6) or P₂₇E₁₈₉ (94%, † p <0.05, f=11.6) (Figure 4A). To further investigate the mechanism of enhanced protection for the inverted triblock copolymer architecture, P₁₅E₂₀₀P₁₅ was compared to two additional diblocks: P₁₄E₁₈₉, which eliminates one of the PPO blocks, and P₁₇E₁₁₂, which approximates half of the triblock (Figure 4B). P₁₅E₂₀₀P₁₅ results in significantly less LDH release than P₁₄E₁₈₉ and P₁₇P₁₁₂ at intermediate concentrations, from which we conclude that the PPO-PEO-PPO architecture enhances protection (P₁₅E₂₀₀P₁₅ (43%) vs. P₁₇E₁₁₂ (110%, _ p <0.05) and P₁₄E₁₈₉ (85%, ‡ p <0.001), at 1.4 μM (f=11.6); P₁₅E₂₀₀P₁₅ (6%) vs P₁₇E₁₁₂ (26%, _ p <0.05) and P₁₄E₁₈₉ (35%, ‡ p <0.001) at 4.4 μM (f=9.2); P₁₅E₂₀₀P₁₅ (2%) vs. P₁₄E₁₈₉ (16%, ‡ p <0.001, f=11.1) and P₁₇E₁₁₂ (3.9%, not significant) at 14 μM). As a complementary study, we evaluated the potential benefit of two PEO blocks by comparing E₇₅P₃₀E₇₅ to P₃₂E₁₂₀ and P₁₇E₁₁₂ (Figure 4C). E₇₅P₃₀E₇₅ and P₃₂E₁₂₀ display similar performance in LDH leakage over the range in concentration where the LDH release drops from ~100% to ~10%, which indicates that there is no benefit to the PEO-PPO-PEO triblock over the diblock architecture with methyl and hydroxyl end groups. P₃₂E₁₂₀ LDH leakage is not statistically different than E₇₅P₃₀E₇₅ or P₁₅E₁₀₀P₁₅ (Figure S9).

3.4. End Group.

Previous work has shown that the addition of a hydrophobic tert-butyl “anchor” end group on the PPO block in a diblock copolymer increased protection in select settings.^1,32 Here we deepen evaluation of the impact of t-butyl anchors in the context of concentration dependence and additional molecular designs. Compared to hydroxyl-P₁₄E₁₈₉, t-butyl-P₁₆E₁₈₀ had superior protection from 1 to 14 μM, reinforcing the idea that the end group hydrophobicity plays a significant role in protection efficacy (Figure 5A) (* t-butyl-P₁₆E₁₈₀ vs. P₁₄E₁₈₉: 47% vs. 110% (p <0.001, f=7.3), 3% vs. 35% (p <0.05, f=11.1), and 0.5% vs. 16% (p <0.001, f=11.4) at 1.4 μM, 4.4 μM, and 14 μM, respectively). Similarly, t-butyl-P₁₅E₇₅ has superior protection compared to the hydroxyl-P₁₇E₁₁₂ across the entire titration (Figure 5B) († t-butyl-P₁₅E₇₅ vs. P₁₇E₁₁₂: 6% vs. 26% (p <0.05, f=11.1) and 1% vs. 4% (p <0.05, f=11.4), 4.4 μM and 14 μM, respectively).

Figure 5.

Addition of a t-butyl anchoring end group increases protection efficacy. A and B) Increased hydrophobicity of the t-butyl end group helps to anchor the polymer into the membrane increasing protection efficacy compared to the less hydrophobic hydroxyl end group. (Statistical difference of p <0.05 between two polymers are noted as: * P₁₄E₁₈₉ and t-butyl-P₁₆E₁₈₀, †P₁₇E₁₁₂ and t-butyl-P₁₅E₇₅) Several polymers are presented with data also shown in Figure 4 to facilitate focused comparisons.

3.5. PEO Homopolymer.

While a hydrophobic end group or an additional PPO block aid protective efficacy, PEO has also been shown to engage^34,35 and protect³² cell membranes. We evaluated PEO homopolymer of a similar molecular weight to E₇₅P₃₀E₇₅ (M_n = 8,000 Da, E₁₈₁) as well as shorter (M_n = 5,000 Da, E₁₁₄) and longer (M_n = 20,000 Da, E₄₅₅) variants. Over concentrations from 0.7 μM to 50 μM all three molecular weights exhibit some degree of protection but with diminishing benefit at higher concentrations, with 7–20% LDH release at 50 μM (Figure 6A). This contrasts with P₁₅E₂₀₀P₁₅, which reduces release to 2% by 14 μM.

Figure 6.

(A) For PEO homopolymer protection efficacy increases with molecular weight on a molar basis. (B) On a mass per volume concentration basis all PEO homopolymers display similar protection, but the length of the molecule seems to have a nominal added benefit. (Statistical difference of p <0.05 between two polymers are noted as: †E₁₁₄ and E₄₅₅, *E₁₈₁ and E₄₅₅) Several polymers are presented in multiple panels (with identical data) to facilitate focused comparisons.

The PEO homopolymer molecular weight has a distinct effect on protection efficacy when compared at constant molarity, with E₄₅₅ performing better than E₁₈₁ and E₁₁₄ at all intermediate concentrations. Yet the difference between molecular weights of these PEO homopolymers is substantial. When converted to a mass basis, the protection of PEO homopolymer collapses onto a nearly universal curve (Figure 6B); E₄₅₅ exhibits nominally less LDH leakage at the higher concentrations. This result indicates that protection potency is independent of molecular weight and is dictated by the total mass-based concentration of the solution.

3.6. Weight Percent EO.

Given the value of hydrophobic anchors (PPO blocks and t-butyl end group) and the efficacy of PEO, we postulated it would be beneficial to assess how these factors combine in individual molecules. Previous work showed the importance of having a sufficiently long enough PPO segment to engage the membrane (> 6 units) and longer PEO segments aided protection at 14 μM.³² To investigate this further, we measured protective efficacy from diblock copolymers with increasing PPO block size (P₀, P_~15, and P_~30) in the context of P_xE_~115 and P_xE_~185. In general, protective efficacy increased with longer PPO block length (Figure 7A for P_xE_~115 and Figure 7B for P_xE_~185) (‡ E₁₁₄ vs. P₃₂E₁₂₀, 104% vs. 59% (p <0.01, f=5.1), 41% vs. 3.5% (p <0.001, f=7.9), and 29% vs. 1% (p <0.001, f=11.4) at 1.4, 4.4, and 14μM respectively) (_ E₁₁₄ vs. P₁₇E₁₁₂, 41% vs. 25% (p <0.001, f=7.9) and 29% vs. 4% (p <0.001, f=11.4) at 4.4 and 14μM respectively) († E₁₈₁ vs. P₂₇E₁₈₉, 40% vs. 22% (p <0.05, f=7.9) and 24% vs. 4% (p <0.001, f=11.4) at 4.4 and 14μM respectively) (* E₁₈₁ vs. P₁₄E₁₈₉, 24% vs. 16% (p <0.05, f=11.4) at 14 μM). We also performed the inverse modification by changing PEO length at constant PPO length – minimal impact is observed for P₁₅E_x with x = 46, 112, 189 (Figure 7C) (at 14 μM, P₁₄E₁₈₉ (16%) vs. P₁₇E₁₁₂ (4%) and P₁₆E₄₆ (3%) with β p <0.001 and λ p <0.001 respectively, f=10.9), whereas increasing the percent of PPO at relatively constant block size by reducing the length of PEO improves performance (P₃₂E₁₂₀ vs. P₂₇E₁₈₉; Figure 7D) (at 1.4 μM, 94% vs 59% (φ p <0.05, f=2.8); at 4.4 μM, 22% vs 6% (φ p <0.001, f=6.9); at 5.5 μM, 14% vs 3% (φ p <0.001, f=7)).

3.7. Overall Threshold Effect.

As detailed in the preceding sections, polymer composition and architecture significantly impact cell protection potency. Nevertheless, it is striking that over a suite of 12 reasonably diverse polymers there is a transition from ineffective to achieving 50% reduction in LDH leakage within a relatively narrow window of 0.8 – 4 μM. (Figure 8A). At 0.7 μM, all block copolymers show LDH release comparable to the non-treated, buffer-only control, i.e. no protection. At 3 μM, all PEO-PPO block copolymers showed at least 50% reduction in LDH release with protection increasing to under 10% LDH release at 14 μM. PEO homopolymers exhibit a similar threshold, where 50% reduction in LDH leakage is achieved by 4 μM; however, PEO homopolymer, especially those of smaller molecular weight, take higher doses of up to 50 μM to show protection efficacies lower than 20% LDH release compared to the non-treated, buffer-only control. In combining the analysis of architecture and end group chemistry, it is observed that P₁₅E₂₀₀P₁₅ and t-butyl-P₁₆E₁₈₀ exhibit the strongest performance, both superior to E₇₅P₃₀E₇₅ and P₁₄E₁₈₉, which are comparable (Figures 4 and 5 and summarized in Figure S8). This demonstrates that comparable benefit, in this context, is gained from PPO valency or a hydrophobic ‘anchor’. Moreover, a reduction in PEO bulk (from t-butyl-P₁₆E₁₈₀ to t-butyl-P₁₅E₇₅) decreases efficacy (Figure S8). The clinically relevant polymer concentration at the membrane and amount of LDH release which indicates in vivo protection is unknown. However, we can compare the results from the threshold effect to doses used in in vivo studies based on a crude estimation of the polymer concentration in animal and clinical trials by dividing the dose of P188 in grams by the estimated blood volume of the subject. In studies where E₇₅P₃₀E₇₅ was protective in vivo, the calculated blood concentration was above the protective threshold seen in our in vitro osmotic stress/LDH leakage assay as illustrated in Figure 8B. (The lowest concentration shown, ~8 μM, corresponds to a Duchenne Muscular Dystrophy study on male subjects ages 12–25).

Figure 8.

A threshold of polymer protection was found between 1 and 4 μM. A) All polymers show a reduction in LDH to 50% compared to non-treated control within the window shown in green. B) In previous in vivo studies where E₇₅P₃₀E₇₅ was protective, the polymer concentrations are above the threshold. Protective blood concentrations of E₇₅P₃₀E₇₅ are estimated by the polymer dose per the blood volume of the model. Referenced above: *⁴⁷, †¹⁰, ‡⁵, _¹², and θ³². Several polymers are presented with data also shown in previous figures to facilitate focused comparisons.

4. DISCUSSION

The range of polymer concentrations dosed in the in vitro osmotic stress-LDH assay revealed that all polymers tested share a relatively narrow window where protection, defined here as 50% reduction of LDH leakage, is reached. This protection threshold shows that all polymers tested follow the same trend of increased protection with increased concentration. Despite architectural differences PEO-PPO-PEO and PPO-PEO-PPO triblocks, PPO-PEO diblocks, and PEO homopolymers have a degree of protection against enzyme leakage at high concentrations, which indicates membrane-polymer interaction for all polymers tested. In previous work by Kim et al. using surface plasmon resonance and neutron reflectivity experiments, both PEO and E₇₅P₃₀E₇₅ have been found to interact with model lipid bilayer membranes at high concentrations (4.5 mM).^34,35 Although the current study has focused on much lower concentrations, i.e. 0.7 to 1200 μM, the trend observed in the LDH release assay indicates that PEO and block copolymers have protective properties at the high concentrations; therefore, the results herein compliment the previous work showing homopolymer and copolymer-membrane interactions.

Although all the homopolymers and block copolymers show a similar concentration threshold of protection, differences in performance at lower concentration allow for distinction between different architectures and molecular variations. Notably, PEO homopolymer was able to provide protection from enzyme leakage in this study compared to the non-treated control at higher doses and molecular weights. When increasing the molecular weight from 8000 to 20,000 Da, the PEO homopolymer had significantly better performance on a molar basis. This may be a consequence of a lower entropy of dissolution at higher molecular weights, which would favor membrane adsorption assuming a favorable interaction energy with the membrane relative to lower molecular weights. However, the total mass in the solution appears to be important as the curves of LDH release vs. mg/ml collapse onto one curve. Adsorption of larger macromolecules on a membrane should be more favorable than smaller polymers due to a lower entropy of dissolution and a greater number of favorable interactions with the surface. However, this is not observed when analyzing the polymer performance as a function of the mass-based concentration, where all PEO homopolymers appear to perform equally in LDH leakage (Figure 6). Although we do not fully understand this result, one possibility is that LDH release protection is attained upon achieving a particular limit of surface coverage of the PEO, which would occur at roughly the same mass loading provided the polymer chains flatten out on the membrane surface. Clearly, this effect warrants further investigation. As the mass in solution increases, there is an increase in protection. The added protection could be caused by a higher fraction of membrane surface coverage. In a study with Pluronic F127 and model liposomes, the liposome surface coverage increases with increasing concentration.³¹ As F127 concentration increased, the percentage of polymer in solution bound to the liposome decreased, indicating there is an upper limit to the amount of polymer that can bind.³¹ These results suggest that as the PEO dosage increases from 0.003–1 mg/ml, the protection increases due to enhanced membrane surface coverage and then levels off because the membrane becomes saturated with PEO at high doses. Although not as protective as the block copolymers, PEO homopolymer does interact with the cell membranes in this and previous studies^20–26, which supports neutron reflectivity and surface plasmon resonance results showing PEO interacting with the model membranes at higher concentrations.^34,35

Addition of a hydrophobic PPO block in the architecture of a triblock or diblock reduces LDH leakage compared to homopolymer PEO. The importance of hydrophobicity is evident when comparing E_xP_y diblocks, where increasing hydrophobicity up to 31 mol % PO decreased LDH leakage. When comparing PEO homopolymer to PEO-PPO copolymers, the addition of a hydrophobic block allows for much lower LDH release. The hydrophobic block acts as an anchor because it has an increased affinity to the inner membrane to avoid unfavorable interactions of the PPO with the aqueous phase, in agreement with the findings of Kim et al.³² This increase in polymer-membrane interaction is correlated with an increase in protection. However, if the molecule becomes too hydrophobic it will solubilize (lyse) rather than stabilize the membrane. Wang et al. found a composition window where poly(2-methacryloyloxyethyl phosphorylcholine)-PPO-poly(2-methacryloyloxyethyl phosphorylcholine) polymers go from protective to toxic between 0.37 and 0.70 mole fraction PPO when examining neuronal survival after oxygen-glucose deprivation;³⁰ this is consistent with polymers in this study being protective up to 0.25 mole fraction PPO. Although hydrophobicity is an important factor in protection, it needs to be balanced with a hydrophilic block to retain protection.

Inverting the triblock to P₁₅E₂₀₀P₁₅ allowed for better protection against enzyme leakage at lower concentrations than E₇₅P₃₀E₇₅ and E_xP_y diblocks. The added potency is attributed to the two hydrophobic PPO ends inserting into the membrane, which increases the number of segments that have favorable interactions with the hydrocarbon portion of the membrane. Previous work with PS-PEP-PS triblock copolymer micelles dispersed in a selective solvent for the PEP blocks showed an extraordinarily longer exchange time than for PEP-PS-PEP based micelles due to the necessity to simultaneously remove two PS blocks from the micelle core.³⁸ By analogy, we attribute the increased protection against enzyme leakage with P₁₅E₂₀₀P₁₅ versus E₇₅P₃₀E₇₅ to the preferential association of the two PPO end block with the lipid bilayer core. Conversely, the E₇₅P₃₀E₇₅ did not show an added benefit of two PEO ends as it exhibited LDH leakage similar to the E_xP_y diblocks.

Increasing the hydrophobicity of the PPO end group in E_xP_y diblocks also led to a decrease in LDH release across the concentration threshold. Results found in this study, and previous works by Kim et al. and Houang et al., show that t-butyl-P_xE_y outperforms the analogous hydroxyl-P_xE_y. Increasing the hydrophobicity of the PPO end group increases the favorable interactions of the hydrophobic block with the non-polar core of the membrane, allowing it to be a more potent anchor.^1,32 Based on the fact that PEO homopolymer provides some protection, the results are consistent with a two-fold overall mechanism. First, the hydrophobic segments are responsible for enhancing anchoring of the molecule into the membrane. Second, the hydrophilic segments provide a therapeutic action, possibly both at and below the membrane surface. The nature of this stabilizing interaction remains a mystery.

Combining these mechanisms for membrane interaction, a balance of overall molecule hydrophobicity, hydrophobic blocks to anchor into the membrane, and hydrophilic blocks to stabilize the membrane association provides optimal protection from enzyme leakage. Increasing the number of hydrophobic blocks, the overall molecule hydrophobicity, or PPO end group hydrophobicity allows us to take advantage of these design parameters to achieve the lowest amount of LDH release. However, the clinically relevant polymer concentration or amount of LDH release is unknown. Although blood concentrations are used to estimate and compare polymer performance in vitro and in vivo, local concentration in the tissue could be much different; therefore, we cannot assert which part of the in vitro protection curve is most indicative of in vivo performance. Therefore, the whole curve is under consideration for screening polymers in vitro as the small difference in polymer performance at low concentration may be significant to in vivo application. Here we note that in a previous report dealing with muscle response in Duchenne diseased mice,¹ a t-butyl-PPO-PEO diblock copolymer performed equivalent to P188, while a HO-PPO-PEO analogue was not effective in restoring muscle activity. While this result is qualitatively consistent with the assay results shown in Figure 5A, it highlights the subtle nature of the interactions that discriminate between physiologically effective versus ineffective compounds.

5. CONCLUSION

We have sampled the membrane protective properties of a variety of block copolymers and homopolymers in an in vitro assay in which murine myoblasts are subjected to an osmotic stress and LDH enzyme release is quantified to assess membrane damage. The assay developed by Kim et al.³² and utilized herein was able to consistently apply stress to the cells and measure enzyme leakage as a metric of membrane damage across a wide range of polymer dosages to reveal the conditions associated with protection. Comparison of FDA approved triblock copolymer P188 (E₇₅P₃₀E₇₅) with inverted triblock P₁₅E₂₀₀P₁₅, various E_xP_y diblocks, and several PEO homopolymers at concentrations between 0.7 to 1200 μM lead to the discovery of a universal threshold in concentration, between 0.8 and 4 μM, where all polymers transition from being unprotective to protective against enzyme leakage. At lower concentrations there is differentiation between molecular architectures, where optimal performance was obtained with P₁₅E₂₀₀P₁₅ and end-functionalized t-butyl-PEO-PEO diblock copolymers. These results, combined with previous studies,^{1,27–32,34,35} including molecular dynamics simulations,^1,33 lead to an optimal molecular design that includes two ingredients: (1) increased molecule hydrophobicity by one or more hydrophobic blocks, possibly augmented with a non-polar end group, and (2) a PEO block of sufficient length to render the molecule soluble in an aqueous medium and to protect against lysing the cell membrane.

ABBREVIATIONS

PEO

poly(ethylene oxide)

PPO

poly(propylene oxide)

LDH

lactate dehydrogenase

t-butyl

tertiary-butyl

PS

poly(styrene)

PEP

poly(ethylene-alt-propylene)

Me

methyl

THF

Tetrahydrofuran

SEC

size exclusion chromatography

NMR

nuclear magnetic resonance

MALDI TOF-TOF

matrix assisted laser desorption/ionization time-of-flight mass spectrometry

CMC

critical micelle concentration