N-Acetylation of Biodegradable Supramolecular Peptide Nanofilaments Selectively Enhances their Proteolytic Stability for Targeted Delivery of Gold-Based Anticancer Agents

Abstract

Peptide materials are promising for various biomedical applications; however, a significant concern is their lack of stability and rapid degradation in vivo, due to non-specific proteolysis. For materials specifically designed to respond to disease-specific proteases, it would be desirable to retain high susceptibility to target proteases, while minimizing the impact of non-specific proteolysis. We describe N-terminal acetylation as a simple synthetic modification of amphiphilic self-assembling peptides that contain an MMP-9-cleavable segment and form soluble, nanoscale filaments. We found that the N-terminus capping of these peptides did not impact significantly on their self-assembly behavior, critical aggregation concentration, or their ability to encapsulate hydrophobic payloads. By contrast, their proteolytic stability in human plasma (especially for anionic peptide sequences) was considerably increased while susceptibility to hydrolysis by MMP-9 is retained when compared to non-acetylated peptides, especially during the first 12 hours. We note, however, that due to the longer time scale required in in vitro studies (72 hours), non-specific proteolysis of both anionic acetylated peptides leads to similar activity in vitro, despite differing MMP-9 kinetics during the early stages. Overall, the enhanced stability against non-specific proteases, combined with the ability of these nanofilaments to enhance the effectiveness of gold-based drugs towards cancerous cells compared to healthy cells brings these acetylated peptide filaments a step closer toward clinical translation.

Graphical Abstract

1. Introduction

Peptides and peptide-based materials are increasingly investigated for biomedical applications such as drug delivery vehicles,^1–9 imaging agents and targeting vectors¹⁰ due to their chemical versatility combined with the biocompatibility of their amino acid degradation products that can be metabolized. Despite the advantages of peptide-based therapeutic carriers, one frequently mentioned disadvantage is their rapid degradation in vivo, due to proteolytic recognition of the peptide backbone by endogenous proteases, specifically exopeptidases.^11–13 While unmodified linear peptides can be degraded within several minutes in physiological conditions,¹⁴ there are several FDA-approved protein and peptide-based therapeutics with half-lives ranging from several hours to nearly 8 days.^15–17

Several peptide modifications have been explored and shown to decrease proteolytic degradation kinetics, including PEGylation, cyclization, and the incorporation of D-amino acids, β-amino-acids and sulfonamides.^18–25 In addition, N-, and C-terminal capping has been observed in several native hormones and neuropeptides such as thyrotropin-releasing hormone (TRH), and melanocyte releasing hormone (α-MSH). These biomolecules have N-termini capped by α-pyroglutamyl and acetyl groups, respectively.¹³ The strategy of N-terminal capping peptide termini with various acyl groups has been employed to enhance the proteolytic stability of short peptides in vitro and in vivo. Benuck et al. described the effects of N-terminal acetylation of somatostatin in human serum and the resulting reduction in proteolysis over the course of 4 hours. Other common amino protecting groups are formyl and pyroglutamyl residues that have been used to enhance peptide stability.²⁶

While these modifications are reported to increase proteolytic stability of peptides, some modifications may lead to conformation changes in the peptide backbone and among the residue side chains, resulting in the disruption of secondary structure formation.^27,28 Modifications may result in loss of stability in vitro and in vivo,²⁹ and prevention of target receptor recognition.^30–33

In supramolecular peptide materials, the presentation of peptides within supramolecular context has potential stabilizing effects that may lead to reduced degradation by restricting enzyme access to the cleavable sequences through electrostatic repulsion and/or peptide packing.³⁴ For materials specifically designed to respond to disease-specific proteases, it would be desirable to retain high susceptibility to these target proteases, while minimizing the impact of non-specific proteolysis. We hypothesized that N-terminal acetylation is a promising modification in this respect. It should be noted that several previous supramolecular peptide designs incorporated a protected N-terminus using either aromatic groups (frequently naphthyl) or alkyl tails,^1,2,35–44 though their effect on non-specific proteolytic stability has not been reported. Here, we evaluate the effect of N-terminal acetylation on the self-assembly properties of MMP-9 responsive peptide sequences. We then study the retention of MMP-9 cleavage while reducing general proteolytic breakdown of supramolecular nanostructures (Fig. 1). We also describe their efficacy as drug delivery vehicles compared to non-acetylated counterparts, previously shown to enhance efficacy of encapsulated gold-based drugs on cancer cells by taking advantage of enhanced local proteolytic activity contributed by MMPs and non-specific proteases.⁴⁵

Figure 1.

Cartoon showing the effect of N-terminal acetylation on the assembly, proteolytic stability and cleavage products when incubated with human plasma. N-terminal acetylation is indicated by the gray circle upon the hydrophobic (purple) segment of the peptides.

2. Materials and Methods

Solid Phase Peptide Synthesis.

Preloaded Wang resins and fluorenylmethyloxycarbonyl (Fmoc)-protected amino acids were purchased from Bachem. A CEM Liberty Blue microwave-assisted solid phase peptide synthesizer was employed to synthesize the peptides. Peptides were generated in a ~1:5 resin to amino acid ratio and excess diisopropylcarbodiimide, Oxyma (ethyl(hydroxyamino)cyanoacetate), and a 20% piperidine in dimethylformamide (DMF) solution. Peptides were capped using acetic anhydride (10% in DMF). The resin-bound peptides were washed 3x in dichloromethane, and subsequently washed 3x in diethyl ether using a filtration column. Cleavage of peptides from the resins, and removal of side chain protecting groups was performed by using a TFA cocktail (containing 95% trifluoroacetic acid (TFA), 2.5% triisopropylsilane (TIS), and 2.5% water) for 2 hours. Precipitation in cold diethyl ether, and centrifugation to decant supernatant, afforded the cleaved peptides. Crude peptides were then dissolved in Milli-Q water and lyophilized overnight.

Preparatory High-Performance Liquid Chromatography.

Lyophilized crude peptides were purified on a preparatory C₁₈ column using a Thermo Scientific Dionex Ultimate 3000 system, after dissolving them in a 50% acetonitrile in water solution containing 0.1% formic acid (FA). A A rotary evaporator was used for removing acetonitrile before further lyophilization.

Removal of TFA counterions.

A 10 mM HCl solution of purified peptides was used to prepare 1 mM peptide solutions. Residual TFA removal was performed by lyophilization. Peptides were washed 3x to remove all residua TFA. HCl washed peptides were characterized using ¹⁹F NMR spectroscopy to confirm removal of TFA.

¹⁹F NMR.

5 mm NMR tubes were used to place the peptide solution samples prepared (500 μL solutions of 1 mM peptide in deuterium oxide, D₂O). A Bruker AVANCE HD III 117 700 MHz spectrometer equipped with a 5-mm QCI-F cryoprobe at a frequency of 658.79 MHz was used to measure ¹⁹F NMR spectra (25 °C). ¹⁹F NMR acquisition parameters (1-Dimensional) were the following: 8.0 ms pulse length (60° flip angle), 100 ppm sweep width (65,789 Hz) centered at −80 ppm, 0.50 sec acquisition time, 2 sec relaxation delay time, and 256 scans (11 min). ¹⁹F NMR spectra (1D) were processed with 5 Hz exponential line broadening. TopSpin 3.5 was used for visualization.

Peptide Stability in PBS.

10mM phosphate buffer (pH adjusted to 7.4 using 0.5M HCl or NaOH) was used to generate 1 mM peptide solutions. Peptides were incubated in a stationary heat block (37 °C), and samples were analyzed via LC-MS every 24h. 30 μL of the incubated solution was added to 270 μL of a 50% acetonitrile in water solution containing 0.1% FA. The analysis of samples was performed on an LC-MS system comprised of an Agilent 1200 Liquid Chromatography system coupled to an Agilent 6340 ion trap mass spectrometer. A Phenomenex Luna Omega column (C₁₈, 50 × 2.1 mm) was employed to inject samples (4 μL) using a 2.5– 95% acetonitrile in water (0.1% formic acid) gradient. A flow rate of 0.300 mL/min for 7.5 minutes was set-up, followed by a wash step with 95% acetonitrile for 1 minute.

Coarse Grained Molecular Dynamics Simulations.

GROMACS⁴⁶ (version 2021.5) was used to simulate 950 peptide molecules modelled at the coarse-grained level using MARTINI 2.1^47,48 force field with a similar protocol to our previous work with the non-acetylated peptides ⁴⁵ and references therein.^49–52 To model the N-terminal acetylation, an additional N-terminal Gly BB bead was added and the charge on this bead was switched off (Qd→Nda) to model the neutral nature of the N-terminus. Consequently, the KK and DD peptides had a net charge of +1 and −3 respectively and therefore an appropriate number of Martini ions were added to the solvated peptide boxes for neutralization.

Critical Aggregation Concentration.

1 mM peptide solutions were prepared using PBS (pH adjusted to 7.4 using dilute HCl or NaOH) Solutions were serially diluted in PBS and thoroughly vortexed. Peptide solutions were then incubated at 50 °C for 15 min in a stationary heat block, then 2 μL of a stock pyrene solution (100 μM in methanol) was added to 100 μL of each peptide solution, mixed gently, and incubated for another 5 min at 50 °C. Solution is cooled to room temperature to allow co-assembly the peptides with pyrene. A Jasco FP-8500 spectrofluorometer was used to measure pyrene emission spectra in a microfluorescence cuvette (3 mm path length). Emission was measured from 350 to 450 nm (λ_ex = 310 nm) using the following measurement parameters: 20 nm excitation and 1 nm emission bandwidth, 0.2 s response, medium sensitivity, 0.2 nm data interval, at 200 nm/min). The critical aggregation concentration was determined by plotting the ratio of intensities between the third and first peak of the pyrene emission spectra. Increasing peptide concentrations were measured in increments of 0.1 mM until a change of the slot slope was observed.

Zeta Potential.

An Anton Paar Litesizer 500 Particle Analyzer was used to measure zeta-potential. 1 mM peptide samples were prepared in 10 mM phosphate buffer (pH adjusted to 7.4 using 0.5M HCl or 0.5M NaOH). 65 μL of peptide solution was pipetted into Univette low volume cuvette, and three series of measurements were completed at 25 °C using Smoluchowski approximation.

Atomic Force Microscopy.

A Bruker Dimension FastScan was used to take AFM images using a FASTSCAN-B probe using fast scan mode. 1 mM Peptide solutions were prepared in phosphate buffer (pH adjusted to 7.4 using dilute HCl or NaOH), sonicated for 10 minutes and drop cast on a freshly cleaved mica surface and allowed to dry for 72 hours before imaging.

Fourier Transform- Infrared Spectroscopy.

20 mM peptide solutions were prepared in deuterated phosphate buffer were and pH adjusted to 7.4, then sonicated for ten minutes. 5 μL of peptide solution was drop cast between two CaF₂ cells with PTFE spacers (12 μM thickness × 13 mm diameter). A Bruker Vertex 70 spectrometer was used to measure absorbance spectra from 4000 to 800 cm⁻¹ with 64 scans at 4 cm⁻¹ resolution.

Stability in Human Plasma.

500 μL of 2 mM peptide solutions prepared in phosphate buffer (pH 7.4) was directly added to 500 μL of human plasma (purchased from Sigma-Aldrich, P9523) and incubated at 37 °C in a stationary heat block. At the desired time point, 75 μL of the plasma solution was directly added to 60 μL of a 1:1 acetonitrile:ethanol solution in order to precipitate plasma proteins. Resulting solution was centrifuged at 15000 rpm for 3 minutes (3x) and the supernatant was separated for analysis. 30μL of the supernatant was added to 270 μL of a 50 % acetonitrile in water solution containing 0.1% formic acid. An Agilent 6340 ion trap mass spectrometer was used to analyze samples by injection on a Phenomenex Luna Omega column (C18, 50 × 2.1 mm) using a of 2.5 – 95% acetonitrile (containing 0.1% formic acid) gradient for eight minutes at a flow rate of 0.300 mL/min followed by a wash step with 95% acetonitrile for two minutes.

Synthesis of Gold Compound.

The preparation of gold compound Au-1 was was performed following a reported synthetic method.⁵³ H[AuCl₄] was purchased from Strem Chemicals (Newburyport, MA). Benzyl bromide, ethyl iodide, and silver oxide were purchased from Sigma Aldrich (St. Louis, MO). 4,5-diphenylimizadole and lithium bis(trimethyl silyl)amide were purchased from Alfa Aesar (Haverhill, MA). Reaction solvents were purchased anhydrous from Fisher Scientific (BDH, ACS Grade) and Sigma-Aldrich. Solvents were dried in a SPS machine, and kept over molecular sieves (3 Å, beads, 4–8 mesh); otherwise over sodium, if necessary. Deuterated solvents were purchased from Cambridge Isotope Laboratories, Inc., and were kept over molecular sieves (3 Å, beads, 4–8mesh). Celite (Celite 545, Diatomaceous Earth) was purchased from VWR International and used as received.

Encapsulation Method.

De-ionized water was used to prepare 2 mM of peptide stock solutions. Acetonitrile was used to prepare stock gold-based compound solutions. 100 μL of peptide solution in water was combined with 100 μL of Au-1 solution in acetonitrile in a 1.5 mL centrifugal tube, for a final concentration of 1mM peptide in a 1:1 water:acetonitrile solution. The resulting solution was vortexed thoroughly then sonicated for 20 minutes, followed by concentration under reduced pressure. The resulting dried film was reconstituted in 200 μL of 10 mM phosphate buffer (pH adjusted to 7.4 using dilute HCl or NaOH) to suspend the peptide-gold compound mixture. The mixture was gently vortexed and sonicated for 20 min. Centrifugation (10⁴ rpm for 1 minute) was employed to separate non-encapsulated compound Au-1 from suspended peptide-gold compound nanofilaments. 200 μL of supernatant was collected for further characterization and analysis.

Gold Quantification.

A Perkin Elmer Analyst 800 was employed for quantification. A transversely heated graphite atomizer (THGA) furnace system at 242.8 nm was used. Digestion of sample aliquots was performed in a concentrated nitric acid solution (70 %), refluxing for 2 hours, drying, and redissolving in adequate volumes of 10 % HCl solution to be within the linear calibration curve range (10–100 ppb, r²=0.997). Gold standard solution in 10% HCl were used as calibration solutions.

MMP-9 Kinetics.

100 ng/mL MMP-9 (pre-activated human, catalytic domain (purchased from Sigma, SAE0078) was incubated with drug-loaded peptide solutions (1 mM) in PBS supplemented with CaCl₂, ZnCl₂ in a stationary heat block at 37 °C for 72 hours. 30 μL of peptide solution was directly added to 270 μL of a 50% acetonitrile in water solution containing 0.1% formic acid at each time point. A LC-MS system comprised of an Agilent 1200 LC system coupled to an Agilent 6340 ion trap mass spectrometer was used for sample analysis. 4 μL of sample was injected onto a Phenomenex Luna Omega column (C18, 50 × 2.1 mm) using a gradient of 2.5 – 95% acetonitrile in water (0.1% FA) over 7.5 minutes. A flow rate of 0.300 mL/min was set-up, followed by a wash step with 95% acetonitrile for one minute.

Cell Culture.

IMR90 human lung fibroblast and MDA-MB-231 human breast adenocarcinoma cells were cultured with Dulbecco’s modified Eagle’s medium (DMEM; Fisher Scientific) containing 10 % fetal bovine serum, (FBS; Fisher Scientific), 1 % penicillin–streptomycin (PenStrep; Fisher Scientific), and 1 % and minimum essential media (MEM) nonessential amino acids (NEAA; Fisher Scientific). Human clear cell renal cell carcinoma Caki-1 line was cultured using Roswell Park Memorial Institute (RPMI-1640; Fisher Scientific, Hampton, NH, USA) and supplemented with 10 % FBS, 1 % MEM-NEAA, and 1 % PenStrep. All cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, Virginia, USA) and cultured in a humidified incubator at 37 °C under 5 % CO₂ and 95 % air.

Cell Viability Analysis.

The cytotoxicity of acetylated peptides, Au-1 compound, and Au-1 compound encapsulated in acetylated peptides was determined using a fluorometric-based cellular viability assay with treated Caki-1, MDA-MB-231 and IMR-90 cell lines. Cells were seeded at a concentration of 5.6 × 10³ (cancerous)- 6 × 10³ (non-cancerous) cells per well in 100 μL of appropriate complete media into tissue-culture-grade 96-well flat-bottom microplates (CELLTREAT, Pepperell, MA, USA) for 24 h under 5 % CO₂ and 95 % air at 37 °C in a humidified incubator. Compound Au-1 was dissolved in a 1:1 solution of DMSO and triethylglycol (TEG) (1%), Ac-PD, Ac-PK, Ac-AD, Ac-AK, Ac-PD+1, Ac-AD+Au-1 were provided in buffer solution following AAS calculations to determine the concentration of encapsulated payload. Following a 72-hour incubation period, PrestoBlue cell viability reagent was used to quantitively measure the viability of treated cells. 11 μL per well of 10× PrestoBlue (Invitrogen, Carlsbad, CA, USA) labeling mixture was added to the cells for a final concentration of 1X PrestoBlue and incubated under 5 % CO₂ and 95 % air in a humidified incubator for 1 h at 37 °C. The optical fluorescence of each well was then was quantified at 560 excitation/590 emission using a BioTek Synergy Multi-mode microplate reader (BioTek Instruments, Inc., Winooski, VT, USA). The ratio of absorbance of treated to untreated cells was used to calculate the percentage of surviving cells in least two independent experiments, each with triplicate measurements. All data presented are expressed as the mean ± standard deviation.

3. Results and Discussion

3.1. Synthesis and characterization of acetylated peptides

MMP-9 responsive amphiphilic peptides forming nanofilaments were described previously³⁴ These peptides are modular and comprised of three segments, a hydrophobic, self-assembling segment (FF),⁵⁴ MMP-9 cleavable segment (XLGLAG, X=A/P), and a hydrophilic charged segment (KK/DD) (Fig. 1). We compared four sequences FFX₁LGLAGX₂ where X₁ was either A or P and X₂ was chosen as DD or KK, so varying charge and varying packing density (PD, PK, AD, AK, Fig. 1) for supramolecular assembly of protease-cleavable soluble filamentous structures with hydrophobic cores with variable loading of two gold metallodrugs.^34,45 PD and AD showed sequence-dependent capacity to load these hydrophobic payloads and, remarkably, enhanced the efficacy of these drugs against cancer cells, which was explained in terms of enhanced proteolytic activity (MMPs and other proteases) in the vicinity of these cells. We hypothesized that N-terminal acetylation (Fig. 1) could increase proteolytic stability toward non-specific exopeptidases, thereby holding promise for enhanced half-lives, with retention of specificity for the target MMP leading to nanofilaments with improved pharmacological profile.

We acetylated the N-terminus of the previously reported (first-generation) peptide sequences, which is anticipated to have only minimal impact on self-assembly propensity. Fmoc-based solid phase peptide synthesis (SPPS) was used to synthesize the peptides sequences (Ac-PD, Ac-PK, Ac-AD, and Ac-AK) and acetic anhydride (10 % in DMF) was used to cap the N-terminus in the solid phase. Crude peptides were identified using liquid chromatography-mass spectrometry (LC-MS) (Fig. S1–4) and purified using preparatory HPLC. Residual TFA salts were removed by washing with 10 mM HCl, and removal of TFA was confirmed using ¹⁹F NMR spectroscopy (Fig. S9–12).

3.2. Effects of N-terminal acetylation on self-assembly

For further insights into the self-assembly of the four acetylated peptides, we used a combination of computation and experimental validation, using coarse grained molecular dynamics (MD) simulations, AFM, zeta potential, and FT-IR (Figs. 2–3). Coarse grained molecular dynamics simulations indicate that similar to the non-acetylated peptides,⁴⁵ acetylated peptide self-assembly is driven primarily by the sequestration of the FF (represented by purple beads) and XLGLAG (represented by green beads) into the core of the aggregates (Fig. 2A–B, Fig. S9 A–C). The overall solvent accessible surface areas (SASAs) of all four peptides were measured as an indicator for aggregation, and also separately quantified for the individual segments (purple, green blue) to gain insights into whether these sections remain solvent exposed or become buried in the core of the structure. As expected, upon assembly, the hydrophobic (Ac-FF) and MMP-9 substrate (XLGLAG, X=A/P) segments became buried within the core while the charged segments (di-aspartic acid, DD for Ac-PD, Ac-AD, and di-lysine, KK for Ac-PK, Ac-AK) remained solvent exposed due to favorable interactions (Fig 2C, Fig. S9 D–G). Additional segmental SASA data can be found in Fig. S9 H–J.

Figure 2.

Characterization of self-assembly behavior of acetylated peptides (Ac-PD, Ac-PK, Ac-AD, Ac-AK). A) Snapshots of a 5000 ns Coarse-Grained MD simulation of Ac-PD, B) Total Solvent accessible surface areas (SASAs) and segmental contributions to SASA for Ac-PD in the first 100 ns of the trajectory with inset showing structure of one coarse-grained Ac-PD peptide, C) Equilibrated SASAs of KK residues (top) and DD residues (bottom) for the acetylated peptides in the last 1000ns of the trajectory D) CAC vs. Aggregation Propensity (derived from SASAs) of acetylated and non-acetylated peptides.

Figure 3.

Characterization of self-assembly behavior of acetylated peptides (Ac-PD, Ac-PK, Ac-AD, Ac-AK). A) CAC, cLogP (as predicted by Chemdraw), and zeta-potential analysis of peptide nanostructures, B) AFM images of self-assembled acetylated peptides (Scale bar 500 nm), C) FT-IR of the Amide I region showing the acetylated.

The CAC of the acylated peptides was determined using pyrene as a fluorescent probe and is shown in Fig. 2D, Fig. 3A (Fig. S11). Aggregation propensity of the peptides was determined by comparing initial and final SASAs over the 5000 ns simulation (Fig. 2D, Fig. S9 K). The measured CAC values of the acetylated peptides, while different from the non-acetylated analogues,⁴⁵ remained below 1 mM, and followed the same overall trend where cationic peptide CAC values have lower values compared to the anionic peptides: 0.3 and 0.2 mM for PK and Ac-PK; 0.4 and 0.5 mM for AK and Ac-AK; 0.5 and 0.8 mM for PD, Ac-PD, and 0.6 and 0.9 mM for AD, Ac-AD, respectively. The experimental values aligned with the decreased aggregation propensity observed in the MD simulations for acetylated Asp containing peptides. Overall, Lys containing peptides show enhanced aggregation propensity (lower CAC values) upon acetylation, and Asp containing peptides show enhanced values, due to favorable (Lys)/unfavorable (Asp) electrostatics with the C-terminus respectively (Fig S9 K).

Zeta-potential measurements (Fig. 3A) show the expected positive and negative surface charges (28.5 ± 0.7 mV, 10.6 ± 0.5 mV, for Ac-AK, Ac-PK, respectively, and −24.9 ± 2.2 mV, −24.9 ± 0.6 mV for Ac-PD, Ac-AD, respectively). These trends also agreed with the simulations, observing the solvent accessible surface area (SASA) values of the charged sections (Fig. 2C) (DD for Ac-PD, Ac-AD, KK for Ac-PK, Ac-AK, respectively), we can see similar solvent exposure for Ac-PD and Ac-AD (Fig. 2C bottom). For the Lys containing peptides, (Ac-PK, Ac-AK), an increased solvent exposure is observed for Ac-AK relative to Ac-PK (Fig. 2C top), in agreement with the zeta potential values.

AFM was used to investigate the effect of N-terminal acetylation on nanostructure morphology (Fig. 3B). All four peptides form filamentous structures, demonstrating that N-acetylation does not prevent the formation of 1-dimensional ordered nanostructures. The structures show varying aspect ratios and degrees of bundling, with cationic peptides assembling into longer fibers, while anionic peptides assemble into increasingly bundled filaments. MD simulations correctly predict acetylated peptide assembly into nanofilaments (Fig. S10).

The Amide I region of the FT-IR spectra was investigated to observe any changes in the packing of the modified peptides (Fig. 3C). As expected, the Pro-containing peptides show disordered hydrogen bonds (1640–1650 cm⁻¹). Ac-AK contains a sharp peak around 1620 cm⁻¹, indicating parallel β-sheet formation. The acetylation of AD introduces some disorder as can be seen from the broad shoulder around 1640 cm⁻¹. Overall, these trends are in line with those observed for the non-acetylated peptides.⁴⁵ The combination of scattering, microscopy, spectroscopy, and molecular dynamics experiments show conclusively that the terminal acetylation has only minimal impact on the self-assembly behavior of these peptides.

3.3. Effects of N-terminal acetylation on proteolysis

Next, we moved on to assess the proteolysis kinetics of these systems. Similar to the non-acetylated peptides, the N-capped peptides were also shown to be stable in phosphate-buffered saline (PBS) solution at 37 °C up to 72 hours, with the exception of Ac-PK, of which ~70% remained (Fig. S12). We recently reported on the stability of the non-acetylated peptides in PBS.⁴⁵ We then sought to investigate whether the N-acetylated peptides could retain their MMP-responsiveness. Acetylated and non-acetylated peptides were incubated with MMP-9 (100 ng/mL) for 72 hours and hydrolysis kinetics are shown in Fig. 4A by way of product formation. Similar kinetics were observed, with a slight decrease for acetylated peptides, as evidenced by lower product formation (FFPLG, FFALG, and Ac-FFPLG, Ac-FFALG, for non-acetylated and acetylated peptides, respectively). This is likely due to conformational changes to the backbones leading to sub-optimal enzyme engagement.

Figure 4.

A) Acetylated and non-acetylated peptide incubated with 100 ng/ml MMP-9 at 37 °C. Peptide fragment formation quantified using LC-MS of two separate trials. Peptide fragments shown are as follows: Ac-FFPLG for Ac-PD and AC-PK, Ac-FFALG for Ac-AD and Ac-AK, FFPLG for PD and PK, FFALG for AD and AK, B) Plasma stability of peptides up to 24 hours and acetylated peptides up to 72 hours (37 °C). Peptide remaining quantified using LC-MS.

The overall proteolysis kinetics of peptide sequences, free termini vs. acetylated, were measured in human plasma at 37 °C, up to 24 hours using liquid chromatography-mass spectrometry (LC-MS) (Fig. 4B, Table S1). We observed significant degradation of all four non-acetylated peptides. At 0.5-hour incubation time, parent peptides remaining were ~80%, ~60%, ~22%, and ~19% for PD, AD, PK, and AK, respectively. This rapidly dropped to ~6.5%, <2%, ~0.01%, and ~0.05% after 6 hours. Comparing the proteolysis of the acylated vs non-acylated peptides in human plasma, the acylated anionic peptides showed a significant decrease in proteolysis rate and higher stability over time (Fig. 3B and Table S2), and therefore additional time points were collected, up to 72 hours. The impact of acetylation was found to be strongly dependent on the overall charge of the nanofilaments, with cationic peptides showing marginally increased stability, with a plasma half-life <0.5 hours. N-capped anionic peptides showed significant enhancement in stability, half-lives found for Ac-AD and Ac-PD were 8.64 and 20.7 h respectively, thought to be due reduced electrostatic recruitment of endo and exopeptidases (with negative surface charges at neutral pH), such as aminopeptidase N and dipeptidyl peptidase IV.^34,55,56 The cationic peptides, and the nanofilaments they form are expected to support electrostatic recruitment interaction are therefore still rapidly degraded by proteases. Not only can N-terminal acetylation increase proteolytic stability compared to non-specific proteases, but the stabilized peptides also retain their enzyme specificity and their ability to act as targeted materials.

Peptide fragments observed suggest that hydrolysis of N-terminal amino acid residues is a significant contributor to overall stability. The abundance of most of the longer fragments (3–9 amino acids in length) increases, then decrease over time as they are further hydrolyzed into smaller fragments, and down to dipeptides or individual amino acids. (Fig. S13–16). As expected, an abundance of individual amino acids (we focused on Phe), increases over time as the endpoint of proteolysis. Fig S13 A–B shows peptide fragments observed for PD focusing on N-terminal hydrolysis at F↓FPLGLAGDD. The presence of FPLGLAGDD is greatest after 0.5 hours, then decreases as it is being further degraded. Correspondingly, the presence of F (Phe) increases up to 24 hours, as the peptide is being fully degraded into Phe and other individual amino acids.

As expected, different fragments were observed for acylated and non-acylated peptides (Fig. S17–S20). Similar to the non-acetylated peptides, peptide fragments are formed and degraded into shorter ones over time, while individual amino acids, such as F (Phe), continue to increase over time. Phe fragments, which are an anticipated product of N-terminal exopeptidase action were observed over 24 hours to compare (non-)acetylated peptides show that formation is delayed and gradual in the case of acetylated peptides, suggesting decreased recognition of N-capped peptides by endogenous exopeptidases (Fig. S21). N-terminal acetylated peptides mostly retain MMP-9 responsiveness, while significantly delaying non-specific proteolytic degradation, though by 24 hours, peptides are mostly degraded.

3.4. Performance of N-acetylated peptide filaments in drug loading and release. Effects on cell viability.

Due to the greater stabilization observed for acetylated anionic peptides Ac-PD and Ac-AD in human plasma compared to PD and PK, we investigated the ability of the stabilized peptide assembly to act as delivery vehicles by encapsulating a gold(I)-N-heterocyclic carbene (NHC) compound (Au-1) into the core using the recently reported protocol for PD and AD.⁴⁵ While gold compounds have emerged as attractive potential chemotherapeutics due to their high efficacy against cisplatin resistant cancer cell lines, their clinical translation is mostly hindered by poor water solubility, low bioavailability and short circulating time. Delivery strategies based on nanotechnology are being explored to overcome these pharmaceutical deficiencies.⁵⁷ We demonstrated that the use of peptide nanofilaments (PD and AD) is advantageous over other systems described, due to their facile synthesis and modification potential to include different enzyme-responsive segments. In addition, these peptide nanofilaments are biodegradable, and break down in the tumor area into fragments that subsequently coordinate to the gold species increasing their cytotoxicity.⁴⁵ Compound Au-1 previously described,⁵³ showed high cytotoxicity against multiple human cancer cells.⁵⁸ Importantly, Au-1 also displayed relevant in vivo efficacy (tumor growth inhibition) in a human renal carcinoma Caki-1 xenograft mice model.⁵⁹ The encapsulation efficiency (Fig. S22) followed the same trends as the first-generation non-acetylated peptides, where the increased β-sheet formation of Ac-AD leads to more efficient loading than the disordered assembly of Ac-PD (encapsulation efficiencies of 27.0 % and 14.8% for Ac-AD, and Ac-PD, respectively). Therefore, N-terminal acylation does not prevent its potential drug encapsulation.

The drug-loaded anionic acetylated peptide nanocarriers were investigated in human cells to determine if their increased efficacy (compared to the hydrophobic gold compound Au-1) that was observed previously⁴⁵ would remain intact with the N-terminal modification. We performed cell viability experiments using three human cell lines, two cancerous known to overexpress MMP-9, (clear renal cell carcinoma Caki-1, triple negative breast cancer MDA-MB-231) and one non-cancerous (IMR-90 fibroblasts)^1,45,60,61 (Fig. 5) We note that while the same amount of gold is used at the start of the encapsulation protocol and subsequently, the amount of encapsulated gold does vary slightly. The amount of encapsulated gold (0.82 μM in Ac-PD and 1.0 μM in Ac-AD) is in a similar range and therefore considered appropriate for comparisons. IC₅₀ values for Au-1 in these cells were: 27.7 ± 0.5 μM (Caki-1), 40.8 ± 2.9 μM (MDA-MB-231) and 84.4 ± 6.2 μM (IMR-90).

Figure 5.

Cell Viability of gold anticancer agent Au-1, gold-loaded peptide nanofilaments (Ac-PD+1, Ac-AD+1), and peptides (Ac-PD, Ac-AD) in three human cell lines. Caki-1, MDA-MB-231 and IMR-90 incubated with drug-loaded peptide for 72 h.

Similar to the first-generation peptides,⁴⁵ acylated peptide nanofilaments alone did not display any cytotoxicity at 1 mM. Free gold compound Au-1 (1 μM, so well below IC₅₀) displayed minimal cell viability decrease. A significant decrease in viability for the gold-loaded acetylated peptides was observed relative to free gold compound Au-1 (1 μM) for both peptides in both cancerous cell lines, though it is most pronounced in renal cancer cells (Caki-1). The cytotoxic effect in triple negative breast cancer cells is reduced, particularly for Ac-PD+Au-1 compared to non-acetylated PD+Au-1.⁴⁵ Ac-PD is more likely to be cleaved by MMP-9 than Ac-AD (Fig. 4A), and the decreased activity observed in triple negative breast cancer cells suggests that endopeptidase activity (MMP-9) may contribute less than exopeptidase activity in vitro.

There is minimal activity in non-cancerous lung fibroblasts, indicating the selectivity for cancer cells when loaded into the enzyme responsive peptides for Ac-PD+Au-1. We observe a significant decrease in viability for Ac-AD+Au-1 in healthy cells, though not nearly as much as in the two cancerous cell lines. Most gold(I) compounds are usually not very selective towards cancerous cells (especially renal and breast cancer) which is a problem for clinical applications. This is true for well-known gold(I) compound Auranofin (FDA-approved for the treatment of rheumatoid arthritis) which is being explored for repurposing for different diseases including cancer.⁶² As mentioned before, free compound Au-1 shows a degree of selectivity towards Caki-1 and MDA-MB-231 cells when compared to non-cancerous IMR 90 fibroblast (ca. 3-fold and 2-fold, respectively). However, the concentration of 1 needed to exert a decrease of viability of 50% in the cancerous cells is much higher (27.7 ± 0.5 μM for Caki-1) than that of Au-1 (1 μM) when encapsulated in Ac-PD+Au-1 and Ac-AD+Au-1, for cell viability decreases of 78% and 90% respectively. Importantly, the increased activity of acylated peptides loaded with Au-1 in cancerous cells observed is similar to non-acylated peptides, as reported.⁴⁵ While there is evidence of decreased exopeptidase activity in acetylated peptides up to 24 hours, they may not be reflected in the cell viability experiments that were run for 72 hours. The expected effect of MMP-9 cleavage may be overwhelmed by the magnitude of endogenous protease activity. Additionally, while Ac-AD is less responsive to MMP-9 than Ac-PD, it is degraded at a faster rate in plasma due to non-specific protease activity (Fig. 4), which may contribute to the observed similar increased drug effectiveness observed in vitro, as also observed in our previous work.⁴⁵

The possible effects of upregulated non-specific proteases on cleavage of peptide nanocarriers in vitro, and the role of post-hydrolysis drug-bound peptide fragments in activity in cancerous cells was discussed previously.⁴⁵ The acetylated peptide carriers have higher proteolytic stability with a significant delay in breakdown over the first 24 hours, but in the time scale relevant to the cell experiments they can be expected to be fully degraded, and consequently they display a similar increase in efficacy in cancerous cells. We note that the effects of the observed increase in stability may be more pronounced when investigated in vivo, where the stability may lead to an increased circulation time. Peptides are hydrolyzed into individual amino acids (that have GRAS status), which is advantageous as it prevents the formation of possibly cytotoxic peptide fragments. After 72-hour incubation with plasma, <20 % of the parent peptide remains. We do, however, observe different peptide fragments for the acetylated and non-acetylated peptides, which may indicate that the sequence of peptide fragments is less important, and the enhanced drug efficacy effect is dominated by the presence of short peptide fragments, as proposed previously.

4. Conclusions

In conclusion, we have developed N-acylated supramolecular peptide nanofilaments that can be used as protease-responsive nanocarriers for hydrophobic metal-based anticancer agents. The self-assembly properties and drug loading capacity is comparable to that of non-acylated analogues. We found that capping the N-terminus via simple acetylation of anionic peptides can drastically decrease non-specific proteolytic breakdown in serum during the first 24 hours. The effect was observed to be strongly dependent on the surface charge of the nanofilaments, with non-specific hydrolysis of cationic peptides is only marginally reduced, suggesting a role for electrostatic repulsions with anionic proteases. Moreover, we found that acylated peptides retained their enzyme-responsive nature and can be hydrolyzed by MMP-9 with kinetics overall comparable to non-acylated analogues. As a proof-of-concept, peptide filaments were loaded with Au(I)-NHC complex, and the protocol developed can theoretically be used to load other hydrophobic drugs. In line with observations for non-acylated peptides discussed previously, the drug-loaded, N-terminal acetylated peptides showed a significant increase on cancer cell survival relative to free drugs. Despite the stabilizing effect of N-acetylation towards non-specific proteolysis during the first day, in the 72-hour time course of the cell culture experiments it is likely that MMPs and other proteases collectively degrade the peptide nanostructures, hence these structures target the general enhanced proteolytic environment near cancer cells, rather than specific proteases.

Despite these similar results observed when comparing with non-acetylated peptides in vitro, the work shows promise for clinical applications, where N-acetylation is expected to be essential to maintain stability and achieve desired targeting by reducing competing non-specific proteolysis away from the target site, however this hypothesis remains to be tested. In depth mechanistic studies will follow in the future to better understand cell specific effects of the fate of the peptide filaments with and without acetylation.