Electrohydrodynamic Jetting of Mucoadhesive Protein Nanoparticles as a Chemopreventive Strategy for Oral Squamous Cell Carcinoma

ABSTRACT

Intraepithelial delivery of cancer preventive therapies for oral squamous cell carcinoma (OSCC) has been limited by factors such as rapid mucus turnover, enzymatic degradation, and salivary clearance. These challenges, coupled with poor solubility and associated low bioavailability, have hampered clinical progress. To address these challenges, we present an effective method for encapsulation and sustained release of poorly soluble, apolar therapeutics via mucoadhesive protein nanoparticles (PNPs). To demonstrate feasibility, PNPs loaded with N‐(4‐hydroxyphenyl) retinamide (4HPR, fenretinide), a hydrophobic chemopreventive compound with low solubility were produced via a three‐step process: 1) high‐pressure homogenization to solubilize and stabilize 4HPR via association with human serum albumin (HSA), 2) electrohydrodynamic (EHD) jetting of albumin‐bound 4HPR to formulate 4HPR‐HSA PNPs, and 3) collection of the nanoparticles in the presence of a polycationic compound to infer mucoadhesive properties and physiological stability. This methodology resulted in the effective synthesis of environmentally stable 4HPR‐loaded PNPs, which featured an average size of 192 ± 21 nm, a surface zeta potential of +31 ± 6 mV in water, and overall 4HPR loadings of up to 7.1 wt.%. In vitro activation of the apoptosis execution phase enzyme, caspase‐3, confirmed sustained release and biological activity of 4HPR. Enhanced binding capacity with mucin (K_D = 6.1*10⁻¹¹ M) was confirmed through surface plasmon resonance (SPR) spectroscopy. This generalizable nanoparticle technology addresses a critical challenge in chemopreventive and targeted drug delivery, where clinical efficacy is limited by limited bioavailability and low drug concentrations of highly apolar agents.

Electrohydrodynamic jetting is used for encapsulation of highly apolar agents in protein nanoparticles (PNPs). Interfacial complexation of the protein nanoparticles with polycations introduces mucoadhesive properties to address challenges in chemopreventive and targeted drug delivery. In vitro activation of apoptosis results from sustained therapeutic release from mucoadhesive PNPs.

1. Introduction

Oral squamous cell carcinoma (OSCC) is one of the most prevalent malignancies in the world, making it the sixth most common human cancer [1]. Despite efforts at early diagnosis and prevention, the aggressiveness of tumors at early stages has capped the survival rate at approximately 40%–50% [2]. While immunotherapy and targeted therapy have shown promising results in OSCC management, surgery combined with adjuvant radiotherapy remains the primary treatment modality. Nonetheless, the inconsistency in treatment efficacy and a wide range of life‐long persisting oral dysfunctions experienced by the patients during and after treatments necessitate better options to combat this disease [2].

Chemoprevention entails the employment of natural or synthetic compounds to reverse, suppress, or inhibit the progression of premalignant lesions to cancer [3, 4, 5]. Systemic and local delivery of chemopreventive agents before OSCC development provides an alternative to avert extensive morbidity and mortality. However, most OSCC chemoprevention trials have failed to produce satisfying results due to challenges, namely: (i) compounds—often vitamins or polyphenols—with variable polarity and uniform low solubility that lead to poor bioavailability, (ii) the effective concentrations of chemopreventive compounds tend to be higher than those of pharmaceutical drugs, causing a need for high doses to reach activity, and (iii) a need for targeted or local delivery strategies to mitigate the enhanced toxicity concerns due to the use in relatively healthy individuals [3]. These challenges are more pronounced with systemic delivery, as they tend to cause significant side effects with extensive delivery considerations, i.e., immunogenicity, toxicity, excipients, etc. [4]. In the context of OSCC, premalignant lesions that arise in the oral cavity are visible and amenable to direct agent application and treatment monitoring. In addition, local delivery provides a fundamental pharmacologic advantage, i.e., delivery of chemopreventive‐relevant levels to treatment sites without deleterious drug‐related systemic effects [3, 6, 7]. These considerations prompted our interest in developing local nanoparticle‐based delivery formulations for chemopreventive compounds.

The synthetic vitamin A derivative, N‐(4‐hydroxyphenyl) retinamide (4HPR, fenretinide), possesses a wide range of chemopreventive effects that include growth regulatory effects (induction of apoptosis and terminal differentiation), functions as a small molecule kinase inhibitor (suppression of gratuitous signaling), and suppression of matrix metalloproteinases and invasion (essential for OIN transformation to OSCC) [4, 8, 9, 10, 11]. Despite its cancer‐preventive profile, chemoprevention with 4HPR has proven extremely challenging [12, 13, 14]. Major challenges relate to the physico‐chemical properties of 4HPR, i.e., low bioavailability due to its pronounced hydrophobicity and limited stability under physiological conditions due to autocrystallization and rapid metabolism [15, 16]. Previous OSCC chemoprevention trials that employed systemically administered 4HPR encountered bioavailability challenges due to first‐pass metabolism that inactivated 4HPR by methylation [13, 17]. Attempts to overcome this issue with higher systemic 4HPR doses induced deleterious effects and were similarly ineffective [3, 13]. Various strategies have been used to increase 4HPR bioavailability [16, 18, 19, 20]. While these formulations successfully increased 4HPR solubility in water, they featured a burst release that rendered them inapt for prolonged release in local delivery systems. Furthermore, these formulations relied upon a large number of potentially mucosa‐irritating excipients to increase 4HPR loading [16, 19, 20]. In addition to fundamental pharmaceutical issues, the challenging nature of negotiating the mucosal barrier, i.e., the protective layer of mucus and surface keratin [21, 22, 23] requires novel, tailor‐made delivery systems.

Nanocarriers offer multiple advantages for mucosal delivery owing to their versatility and tunability. Various materials for matrix design and device configuration can enhance therapeutic efficacy [22]. Apart from shielding the encapsulated drug from environmental factors and enzymatic degradation, nanoparticles can facilitate the transport of drugs through epithelial cells or tight junctions and promote cell internalization via receptor‐mediated endocytosis [24]. Furthermore, the large surface area‐to‐volume ratio of the nanocarrier provides increased sites for surface decoration, resulting in better carrier adhesion or penetration [22, 23, 24] In addition, balanced mucoadhesive and mucopenetrating properties of nanoparticles can prolong contact time and improve the local delivery of therapeutics uniformly across the oral cavity [22, 25]. Most importantly, high loading efficiency and controllable drug release from nanocarrier formulation can reduce drug dosage, thereby decreasing the toxicity and adverse side effects associated with high‐dosage conventional treatments [26].

Electrohydrodynamic (EHD) jetting is a versatile technique that utilizes high voltage to atomize macromolecular solutions to form fibers and particles ranging from micro to nanometers in diameter, offering excellent engineering capabilities. While our lab has demonstrated the ability to generate nanoparticles from various materials (e.g., polymers, proteins) to deliver therapeutics effectively, 4HPR nanoparticle formulation presents unique challenges, as outlined above [23, 27, 28, 29, 30, 31, 32, 33]. The low solubility of 4HPR in aqueous solutions can result in an inhomogeneous solution that may lead to unstable jetting, undesired drug crystallization, and inhomogeneous drug distributions, thereby producing inadequate particle sizes and shapes, and undesirable drug loadings and release kinetics. However, our previous studies have shown that albumin loaded with another apolar drug, paclitaxel, can be EHD jetted through a stable Taylor cone to produce a homogenous distribution of nanoparticles and high drug loading capacities [32].

2. Results and Discussion

2.1. Fabrication and Characterization of 4HPR‐HSA PNPs

The 4HPR‐encapsulated protein nanoparticles were fabricated through a three‐step methodology [32]. First, 4HPR was associated with human serum albumin (HSA) through high‐pressure homogenization. Briefly, a pre‐homogenized crude mixture of 4HPR (1 mg mL⁻¹, 10% w/w_protein) in HSA‐chloroform‐ethanol‐water solution was high‐pressure homogenized at 1.6547*10⁵ kPa for 7 cycles, followed by lyophilization with a process yield of 80% [32, 34]. The resulting product, 4HPR‐HSA, was determined to have a loading efficiency of 71% and a 4HPR‐to‐HSA mass ratio of around 1/14. HSA‐bound 4HPR displayed enhanced solubility in aqueous solutions compared to the free drug. 4HPR‐HSA was dissolved in a 20% (v/v) methanol‐water solution to formulate the jetting solution. A flow rate of 0.08 mL h⁻¹, an electric voltage between 5.8 and 6.2 kV, and a jetting distance of 7 cm between the needle and collection plate were found necessary to obtain a Taylor cone and generate homogenous spraying (Scheme 1). During the jetting process, an electric field was used to induce jet formation and rapid evaporation of the solute components, resulting in solid protein nanoparticles.

SCHEME 1.

Three‐step workflow used for the production of 4HPR‐HSA PNPs comprised high‐pressure homogenization, EHD jetting, and interfacial complexation. The association of 4HPR in HSA is achieved first through mixing of 4HPR dissolved in a water‐immiscible solvent with an aqueous solution, followed by a pre‐homogenization process by tip sonication. High‐pressure homogenization followed by electrohydrodynamic jetting and interfacial complexation with a polycation resulted in stable 4HPR‐HSA PNPs that were positively charged in aqueous environments.

The geometric properties of jetted nanoparticles are essential for evaluating the jetting process and act as a critical checkpoint to assess nanoparticle characteristics before and after suspension. ImageJ‐based analysis of scanning electron microscopy (SEM) images showed that the jetted 4HPR‐HSA nanoparticles had an average diameter of 181.9 ± 34.6 nm (Q1/median/Q3 = 166.2/180.8/195.3 nm), a circularity of 0.89 ± 0.09 (Q1/median/Q3 = 0.89/0.92/0.94), and a roundness of 0.86 ± 0.15 (Q1/median/Q3 = 0.84/0.93/0.96) (Figure 1A,B). To elucidate the impact of high‐pressure homogenization in EHD jetting, statistical analyses to compare the geometric properties of the 4HPR‐HSA and a control, blank HSA nanoparticles (i.e., jetted HSA), prepared under identical conditions, are provided in Figure 1C,D, and Table S1. The comparison found 4HPR inclusion conveyed no significant difference in the diameter and circularity between these nanoparticle groups. However, a slight decrease in roundness for the 4HPR‐HSA nanoparticles was observed (Figure S1).

FIGURE 1.

Characterization of as‐jetted HSA and 4HPR‐HSA PNPs. SEM micrographs of jetted (A) HSA PNPs and (B) 4HPR‐HSA PNPs, with size distribution shown in insets, n = 1400 (scale bar = 2 µm). Statistical values, (C) average diameter and (D) roundness, were obtained through ImageJ and analyzed using a one‐way ANOVA followed by a Tukey's post hoc test, where *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

Surface modification and stabilization of the jetted protein nanoparticles were achieved through interfacial complexation with a polycationic compound added to the collection process (Figure 2) [35]. This approach is based on the hypothesis that a suspending solution with a higher isoelectric point relative to the protein's isoelectric point (HSA = 4.7) would promote electrostatic interactions and complexation of a thin layer of a positively charged polycation (here, polyethyleneimine) onto the negatively‐charged nanoparticle surface [36]. Zeta potential measurements of the nanoparticles confirmed this premise (Figure 3B). To minimize cytotoxicity, all excess polycations were removed by centrifugation.

FIGURE 2.

Interfacial complexation of protein nanoparticles. The as‐jetted 4HPR‐HSA PNPs are stabilized by interfacial complexation through the simple addition of a polycationic compound into the aqueous collection solution.

FIGURE 3.

Characterization of 4HPR‐HSA PNPs in suspension. The (A) average diameter and (B) zeta potential in water obtained by DLS/ELS and analyzed using a one‐way ANOVA followed by a Tukey's post hoc test, where *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001, n = 4. (C) Overlay DLS size distribution profile of the suspended HSA PNPs and 4HPR‐HSA PNPs in water. (D) Overlayed absorbance spectrum of free 4HPR, 4HPR‐HSA, and 4HPR‐HSA PNPs in water. 4HPR is dissolved in 5% (v/v) ethanol‐water mixture while 4HPR‐HSA and 4HPR‐HSA PNPs were suspended in water.

Subsequently, 4HPR‐HSA PNPs were resuspended and characterized with DLS/ELS to determine their size, size distribution, and surface potential. 4HPR‐HSA PNPs possessed a diameter of 191.5 ± 20.8 nm, a polydispersity index (PDI) of 0.135 ± 0.04, and a positive surface charge of +31.3 ± 5.7 mV in water (Figure 3A–C). The fact that the hydrodynamic diameters of 4HPR‐HSA PNPs were larger than their respective diameters obtained by SEM analysis can be attributed to (slight) swelling of the nanoparticle matrix in aqueous media. The corresponding low PDI indicates good homogeneity, while their positive zeta potential confirms interfacial complexation with the polycationic compound. In addition, a side‐by‐side comparison between the 4HPR‐HSA PNPs and HSA PNPs found no significant difference in size and zeta potential (Figure 3A,B, Table S2). Intensity plots obtained by dynamic light scattering confirmed the results of the SEM analysis (Figure 3C).

Next, successful 4HPR encapsulation was confirmed via absorbance measurements by UV–vis spectroscopy. 4HPR, a retinoic acid derivative, has a distinctive yellow color and a characteristic absorbance band in the mid‐300 nm range. The absorbance spectra for free 4HPR, 4HPR‐HSA, and 4HPR‐HSA PNPs displayed matching intensity maxima at 365 nm (Figure 3D), indicating the presence of chemically intact 4HPR in both 4HPR‐HSA and 4HPR‐HSA PNPs. Differences in the spectrum of 4HPR‐HSA PNPs relative to free 4HPR can be attributed to nanoparticle scattering. The scattering signal disappeared after acidification of the nanoparticle suspension, suggesting rapid dissolution of the particles at pH values below 5 (Figure S2).

2.2. Stability of 4HPR‐HSA PNPs in Aqueous Suspension

The stability of nanoparticles in suspension is crucial to achieving therapeutic goals in drug delivery. Nanoparticle stability can ensure controlled release, extend shelf life, enhance the bioavailability of poorly soluble drugs, and improve scalability [22]. Dynamic light scattering measurements provide size measurements and a polydispersity index (PDI) over an extended period of time, allowing for monitoring of 4HPR‐HSA PNPs' stability in suspension. Our results show that 4HPR‐HSA PNPs display excellent size stability at room temperature in water, with less than 3% deviation from the average diameter over 6 days (Figure 4A; Figure S3).

FIGURE 4.

Storage stability of 4HPR‐HSA PNPs in water suspension. (A) The average diameter and PDI of 4HPR‐HSA PNPs in water over time. The diameter is labeled blue, and the PDI is labeled red. (B) The relative intensity of the absorbance reading at 365 nm for 4HPR, 4HPR‐HSA, and 4HPR‐HSA PNPs suspended in water stored in the dark. Bars represented mean ± SD (n = 4).

It is well‐documented that a loss in the absorbance of the 4HPR derivatives over time can be attributed to the recrystallization of 4HPR [16]. This effect will negatively impact the bioavailability of 4HPR. For poorly soluble molecules, such as 4HPR, amorphous drug states are generally preferred as they are associated with higher solubility and retention of bioactivity relative to their crystalline forms [37, 38, 39]. To be able to adequately monitor the absorbance readings of amorphous 4HPR available, 4HPR‐HSA and 4HPR‐HSA PNPs were suspended in water at room temperature to minimize the release of the encapsulated molecule and the interference of released 4HPR in solution. Comparing the relative intensity of three different 4HPR derivatives to their initial values, we found that the albumin‐bound 4HPR, either in the form of 4HPR‐HSA or 4HPR‐HSA PNPs, was less susceptible to recrystallization than free 4HPR under identicalconditions (Figure 4B). Our results show significant differences between groups and indicate that more than 70% of the free 4HPR molecules recrystallize within 48 hours, while more than 80% of the albumin‐bound 4HPR remains in an amorphous, bioavailable form (Figure S4). Further imaging of the drop‐cast suspensions after 2‐day incubation confirmed that the loss of the absorbance signal is due to 4HPR's recrystallization in aqueous solution. Conversely, minimal crystalline structures were found in the drop‐cast of 4HPR‐HSA PNPs suspension (Figure S5B–D), reaffirming the design of the particle to retain 4HPR and its bioactivity.

In addition, differential scanning calorimetry (DSC), a thermoanalytical technique that allows the differentiation of thermal events, can provide further evaluation of the crystallinity state of the encapsulated 4HPR in the nanoparticles [16, 40]. The absence of the melting peak of 4HPR (173.3 °C) in the 4HPR‐HSA PNPs thermogram indicates that 4HPR is more likely to be in its amorphous, bioactive state when encapsulated in the particle (Figure S5A).

2.3. Mucin Binding Capacity of 4HPR‐HSA PNPs

As mucin‐rich saliva naturally coats the oral cavity, mucin interactions are critical factors to consider in OSCC prevention. Effective OSCC prevention requires saturation of and penetration through mucin‐based barriers. It has been reported that electrostatic interaction, hydrogen bonding, and hydrophobic interaction can enhance interaction with mucin, while a smaller size can promote better penetration [21]. These properties should be considered when designing mucoadhesive PNPs with improved penetration and prolonged retention. It is also important to elucidate 4HPR‐HSA PNPs ‐mucin interaction in vitro. The mucin binding capacity of PNPs was initially assessed by monitoring the mucin binding to PNPs. After incubation of a defined amount of PNPs with 75 µg mL⁻¹ of mucin for 30 min and removal of the nanoparticles via centrifugation, UV–vis spectroscopy was used to determine the remaining amount of free mucin in the supernatant [41, 42]. Because we expected the mucin binding to be dominated by electrostatic interactions, cationic PS and anionic PS particles were included as control groups representing opposite surface charges (Table S3). Absorbance measurements at 255 nm after centrifugation support significant mucin binding for all samples (Figure S6A). In this experiment, we evaluated mucin binding after centrifugation relative to the non‐particle‐containing mucin suspension. The results shown in Figure 5 indicate that both cationic PS and anionic PS particles bind mucin, albeit at statistically different levels (p < 0.001) than 4HPR‐HSA PNPs. Quantitatively, 4HPR‐HSA PNPs had a binding capacity that is 3‐fold higher than cationic PS particles. Further, considering the mass of mucin removal per unit surface area of the nanoparticle revealed that 4HPR‐HSA PNPs are 6‐ and 11‐fold more pronounced than cationic PS and anionic PS particles, respectively (Figure S6B). These results also imply that nanoparticle size and density are of lesser relevance for mucin binding capacity. Considering that the 4HPR‐HSA PNPs had a similar zeta potential than cationic PS particles, electrostatic interactions with mucin are likely not the only factor contributing to 4HPR‐HSA PNPs’ enhanced mucoadhesion. We suspect that hydrogen bonding between the carboxyl and hydroxyl groups of mucin molecules with interfacial amine groups of 4HPR‐HSA PNPs, Van der Waals interactions, and entanglement of the polycation with adsorbed mucin all contribute to their enhanced binding capacity [43].

FIGURE 5.

Mucin binding capacity of nanoparticles. The data represent the differences in the average absorbance measurements of mucin/nanoparticles mixture relative to mucin‐only suspension post‐centrifugation. Bars represented mean ± SD, and the data were analyzed using a one‐way ANOVA followed by a Tukey's post hoc test, where *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001, n = 4.

2.4. Mucin Binding Kinetics of 4HPR‐HSA PNPs

To further elucidate the interfacial interaction and binding kinetics of the 4HPR‐HSA PNPs with mucin, surface plasmon resonance (SPR) spectroscopy was employed. First, mucin was immobilized onto a gold‐coated chip surface via covalent conjugation to emulate the surface of the mucosa. Mucin immobilization was achieved through EDC/NHS chemistry onto pre‐modified carboxylate gold sensor chips provided by the vendor (Figure 6A). Subsequent mucin injections show that the mucin‐immobilized surface was saturated, and a long wash (>1800 s) revealed little signal dissipation, evidence of effective binding (Figure 6B) to the surface. We proceed further with the injections of the 4HPR‐HSA PNPs to observe the impact of surface properties on adhesion capabilities. Although adhesion occurred to both channels, our results show superior binding of 4HPR‐HSA PNPs to the mucin‐immobilized surface than its non‐immobilized counterpart (Figure S7). For consistency, we employed the same commercial polystyrene nanoparticles used in the previous study, cationic PS and anionic PS particles, as controls. In addition, a reference channel that did not contain mucin was included to account for non‐specific binding (NSB). Subtraction of the mucin‐immobilized channel and the reference channel effectively accounted for the non‐specific binding and was done for all experiments to yield the SPR tracings that reflect the nanoparticles' interaction with mucin (Figure 6C).

FIGURE 6.

Assessment of 4HPR‐HSA PNPs binding affinity for mucin through SPR spectroscopy. (A) Scheme of SPR sensor preparation. A carboxyl‐functionalized gold sensor chip is immobilized with mucin through EDC/NHS chemistry, followed by nanoparticle injection. SPR sensorgrams of (B) immobilization of mucin to carboxyl sensor chips, (C) dose‐dependent response of 4HPR‐HSA PNPs to the mucin‐immobilized surface, (D) overlay of normalized PS anionic, PS Cationic, and 4HPR‐HSA PNPs at 1.33 pm, and (E) the average dissociation constant K_D of nanoparticles to the mucin‐immobilized surface. The data was analyzed using a one‐way ANOVA followed by a Tukey's post hoc test, where *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001, n = 3.

As expected, 4HPR‐HSA PNPs exhibited the highest level of binding to the mucin‐immobilized surface. Binding to the non‐mucin immobilized reference channel was anticipated as the NHS‐ester can generate negatively charged carboxyl groups upon hydrolysis, capable of promoting electrostatic interactions. For kinetic analysis, all sensorgrams were fitted to a Langmuir 1:1 model. This allowed us to derive rate constants for the association and dissociation of the nanoparticles to the mucin substrate. Due to the difference in size and density, we expected that the particles injected would exhibit different RU_max values. Consequently, all readouts are expressed relative to their RU_max for better side‐by‐side comparison between different particles (Figure 6D). While strong binding affinities were observed for all particles, the normalized signal of 4HPR‐HSA PNPs confirmed better mucin adhesion as indicated by their slower egress rate. Our kinetic model produced average K_D values of 6.1 ± 2.40 × 10⁻¹¹ m for 4HPR‐HSA PNPs, 2.8 ± 4.6 × 10⁻¹¹ m for cationic PS, and 4.0 ± 3.0 × 10⁻¹⁰  m for anionic PS particles, respectively (Figure 6E; Figures S8 and S9). While the K_D values for 4HPR‐HSA PNPs and the cationic PS particles varied only by an order of magnitude, their respective dissociation rates differed substantially. 4HPR‐HSA PNPs had a significantly slower egress rate than the cationic nanoparticles (Figure S9). Overall, our SPR spectroscopic studies support the data of the mucin binding capacity experiment: while positively charged particles featured increased avidity to the negatively charged mucin, the involvement of non‐electrostatic contributions reinforced the association of 4HPR‐HSA PNPs to mucin, thus resulting in a stronger retention as would be required for prolonged therapeutic release.

2.5. Fenretinide Loading, Encapsulation, and Release

Quantification of the loading efficiency of the 4HPR derivatives can be achieved by UV‐vis spectroscopy. To quantify the loading of fenretinide in 4HPR‐HSA PNPs, nanoparticles were incubated in a 50% (v/v) ethanol‐water mixture containing 0.1% (v/v) acetic acid (pH 5). Under these conditions, 4HPR‐HSA PNPs disassembled into their individual components, and interference of the UV–vis readings due to scattering can be avoided (Figure S2). Additionally, ethanol increased the solubility of 4HPR and promoted its dissociation from albumin. Based on the calibration curve shown in Figure S10, the loading efficiency of 4HPR‐HSA was calculated to be 71% based on Equation (2). Similarly, the overall production yield of 4HPR‐HSA PNPs was 31%, according to Equation (3). To determine the in vitro release of 4HPR from HSA PNPs, nanoparticles were incubated at 37 °C in distilled water and centrifuged. Next, the supernatant was analyzed via UV‐vis spectroscopy at designated time points. Ultrapure distilled water was used, and 5% (v/v) methanol was added to aid the dissolution of 4HPR and ensure that the release was not limited by poor 4HPR solubility. In addition, 0.1% (v/v) of acetic acid was added to the extracted supernatant to stabilize 4HPR. The Weibull cumulative distribution function (CDF) was used to fit the release curve because it resulted in the highest R² value relative to alternative release models. Based on the Weibull CDF, a β value of 0.63 was derived, indicating that under these release conditions, Fick's diffusion dominated the fenretinide release mechanism from 4HPR‐HSA PNPs. Overall, only 55.0% of the total payload was released at the 8.6‐hour mark (τ = 8.6) (Figure 7) [44, 45].

FIGURE 7.

Release profile of fenretinide from 4HPR‐HSA PNPs. The release was conducted at 37 °C in distilled water supplemented with 5% (v/v) ethanol. The release profile is depicted as the % cumulative release relative to the total mass released over the release duration. The release curve followed the Weibull cumulative distribution function (CDF) as shown in the red dotted line. Bars represented mean ± SD (n = 3).

2.6. Determination of 4HPR‐HSA PNPs Bioactivity via Activation of the Apoptosis Execution Phase Enzyme Caspase 3

The induction of programmed cell death (apoptosis) and removal of target cells from the proliferative pool is a well‐recognized mechanism of chemoprevention for fenretinide [45]. Consequently, the capacity of bolus‐delivered 4HPR and 4HPR‐HSA PNPs‐released fenretinide to induce the functional activity of the execution phase enzyme caspase‐3 was assessed. As depicted in Figure 8A [46], the 4HPR's mode of action is as follows: 1) 4HPR‐HSA PNPs undergo endocytosis, where the particles ultimately accumulate in lysosomes that possess inherently lower pH (pH = 4.5–5.0) that will destabilize the particle's interfacial complexation and allow 4HPR to diffuse into the cytoplasm. 2) Reactive oxygen species (ROS) generation triggered by 4HPR disrupts the electron transport chain (ETC) at the mitochondrial membrane, leading to permeabilization of the mitochondrial outer membrane. 3) The ETC disruption induces mitochondrial damage and permeabilization that results in the release of cytochrome c into the cytoplasm. 4) Cytochrome c binds to the protease‐activating factor (Apaf‐1) and triggers Apaf‐1 oligomerization into a heptameric structure, forming the apoptosome. 5) The apoptosome recruits procaspase‐9 and activates the initiator phase caspase‐9, and 6) activated caspase‐9 cleaves and activates the execution phase enzyme, caspase‐3. Once caspase 3 is activated, the cell is destined for apoptosis [46, 47].

FIGURE 8.

4HPR‐HSA PNPs bioactivity via activation of the apoptosis execution phase enzyme, caspase‐ 3. (A) A schematic depiction of 4HPR‐HSA PNPs uptake and 4HPR release via endocytosis to trigger caspase‐3. (B) Caspase‐3 activity after dosing with groups of control (CTR) + blank PNPs (n = 5), 5 µM 4HPR + blank PNP (n = 5), and 4HPR‐HSA PNPs (5 µm 4HPR) (n = 9). Each group was applied with the same number of particles (6.18 × 10⁹ particles/mL⁻¹), ****p < 0.0001, ***p < 0.001, and *p < 0.05, respectively. Figure A is constructed and modified from BioRender.com.

This study compared the ability of bolus‐delivered 5 µm 4HPR and 4HPR‐HSA PNPs (5 µm 4HPR released 24 h) to induce caspase‐3 activity in human oral premalignant epithelial cells (EPI) and an OSCC cell line (JSCC1) (Figure 8B). Blank sPNPs particles at the same dosing concentration of 4HPR‐HSA PNPs were applied to the control and bolus‐delivered 4HPR group to account for the effect of nanoparticle addition. Our results showed that activation of caspase‐3 was successfully achieved for all 4HPR‐containing groups, evident through the increased units of activity/µg protein value to control (0.047 ± 0.004) and that the addition of nanoparticles is not a limiting factor. We further confirmed the in vitro bioavailability and bioactivity of 4HPR delivered from the 4HPR‐HSA PNPs nanoparticles, displaying cellular activity of 0.067 ± 0.004 units of activity/µg protein. In addition, we observed the efficacy of bolus‐delivered 5 µm 4HPR as reported in the literature [4]. However, 4HPR‐HSA PNPs demonstrated a lower activity than the bolus‐delivered 4HPR. We attribute this to differences in the bioavailability of 4HPR in vitro, as the bolus‐delivered 4HPR causes an abrupt 5 µm challenge to the cells upon addition. In contrast, 4HPR‐HSA PNPs presented a slower, sustained release over time in this artificial in vitro setting [48, 49]. Nevertheless, these findings are generally consistent with our previous work, which has confirmed the chemopreventive benefits of sustained‐release, local 4HPR delivery relative to bolus delivery in vivo [16]. From a translational perspective, sustained‐release formulations would also reduce the need for daily injections or intravenous infusions.

3. Conclusions

This work demonstrates a novel strategy to address major challenges of local chemopreventive strategies, i.e., mucoadhesive, non‐toxic nanoparticles with high loading capacity and sustained release for poorly soluble, highly apolar chemopreventive agents. Surface modifications that promote either electrostatic interaction, hydrogen bonding, or physical entanglement with mucin are preferred to enhance the mucoadhesion of nanoparticles. Here, we demonstrate the capability to stabilize protein nanoparticles through a polycationic interfacial complexation that also endows mucoadhesive properties. The resulting HSA PNPs containing albumin‐bound 4HPR were sub‐200 nm in diameter with low polydispersity, demonstrating a low dissociation constant, a sustained release of 55% payload release at 8.6 h, and induced caspase‐3 activity to activate apoptosis for chemoprevention. For future clinical applications, we envision nanoparticle‐based co‐formulations with 4HPR and the IL‐6 receptor antagonist tocilizumab (TCZ) to further suppress the development of OSCCs. This could take the form of a formulated Janus nanoparticle to enable field coverage throughout the oral cavity [23]. In conclusion, these studies show the feasibility of encapsulating poorly soluble therapeutics into protein nanoparticles with enhanced mucoadhesion as a local delivery system.

4. Materials and Methods

4.1. Materials

Human serum albumin (A1653), mucin from porcine stomach, type II (M2378), branched polyethylenimine (408727), Tween20 (P2287), chloroform (C2432), HPLC‐grade methanol (34860), and ethanol (459828), were purchased from Sigma–Aldrich (St. Louis, MO). Albumin from bovine serum albumin, Alexa Fluor 488, 647 conjugate (A13100/A34785), LC‐MS grade water (51140), FluoSpheres Amine‐Modified Microspheres (F8764), Ultrapure distilled water (10977), distilled water (15230‐162), and Dulbecco's Phosphate Buffered Saline (14190‐144) were purchased from Thermo Fisher Scientific (Waltham, MA). Fluoresbrite Multifluorescent Microspheres (24050‐5) were bought from Polysciences, Inc. (Warrington, PA). Fenretinide (1396) was purchased from Tocris Bioscience (Bristol, United Kingdom).

4.2. High Pressure of Homogenization for Nanoparticle Albumin‐Bound 4HPR (4HPR‐HSA)

A high‐pressure homogenizer (HPH) was employed as described in the literature with slight modifications [32, 50]. Chloroform was added to ultrapure water with vigorous shaking to make a chloroform‐saturated water mixture. The chloroform‐saturated aqueous solution was removed and added to a 50 mL conical tube holding pre‐weighted (294 mg) human serum albumin (HSA). 4HPR (30 mg) was dissolved in a chloroform/ethanol mixture and added to the HSA‐dissolved aqueous phase, which underwent pre‐homogenization with tip sonication. High‐pressure homogenization with EmulsiFlex‐B15 was done at 1.6547 × 10⁵ kPa for seven cycles. The resulting product was flash‐frozen with liquid nitrogen and lyophilized overnight to obtain 4HPR‐encapsulated human serum albumin, 4HPR‐HSA.

4.3. Electrohydrodynamic (EHD) Jetting of 4HPR‐HSA and HSA

4HPR‐HSA and human serum albumin (HSA) were electrohydrodynamically jetted to produce nanoparticles following methods previously published [30, 31, 32, 51, 52, 53]. Briefly, a jetting solution comprised of 3 mg mL⁻¹ human serum albumin or 4HPR‐HSA was prepared in a 20% methanol‐in‐water mixture. A positive voltage was applied to a pumping syringe (0.08 mL h⁻¹) and a stable Taylor cone was formed. The applied voltage and jetting distance for the most stable and consistent jetting were determined to be between 5.8 and 6.2 kV and 7 cm. A flat pan was used to collect the jetted nanoparticles, jetted 4HPR‐HSA PNPs, and jetted HSA PNP. The nanoparticles were stored in the dark until further processing.

4.4. Characterization of the Jetted 4HPR‐HSA PNPs and HSA PNPs

The jetted 4HPR‐HSA PNPs and HSA PNPs were characterized using the FEI Nova 200 Nanolab SEM/FIB at the Michigan Center for Materials Characterization (MC)² [. The nanoparticles were imaged using a voltage of 5 kV, a current of 0.40 nA, and a dwell time of 3 µs. Statistical analyses were made with ImageJ using the methods previously reported [53]. Geometric properties of the jetted nanoparticles, i.e., diameter, circularity, and roundness, were obtained.

4.5. Interfacial Complexation and Stabilization of 4HPR‐HSA PNPs and HSA PNPs

A concentration (8 µg mL⁻¹) of polyethylenimine (PEI) was prepared. On the surface of the collection pan, PEI solution was added to stabilize and collect the nanoparticles. The collected particles then underwent centrifugation for 50 min at 21500 RCF to remove excess PEI and to obtain the desired particle size. The resulting pellet was resuspended with horn sonication to obtain interfacial complexed nanoparticles (HSA PNPs and 4HPR‐HSA PNPs) in suspension.

The size and zeta potential of the nanoparticles in suspension were measured via dynamic light scattering (DLS) using the Malvern Instruments Zen 3600 Zetasizer. Nanoparticle Tracking Analysis (NTA) by ZetaView Twin was used as a secondary analytical source to confirm the size and distribution and obtain concentration under fluidic flow, n = 3.

4.6. Stability of 4HPR‐HSA PNPs in Suspension

The nanoparticles were suspended in water and kept in the dark throughout the experiment. The size and polydispersity index (PDI) measurements were made with DLS. No additional processing (i.e., sonication) was made before measurement, and the results were reported as measured, n = 4.

4.7. Characterization of 4HPR‐Derivatives via UV–vis Spectroscopy

The samples were suspended in water and loaded into a quartz cuvette for absorbance measurements by a Molecular Devices SpectraMax5e Multi‐Mode Microplate Reader. For 4HPR, 5% ethanol was used to ensure solubilization of the hydrophobic drug. A spectrum read from 250 to 550 nm was done to obtain the absorbance spectrum of the samples and to generate the calibration curve.

4.8. Bioavailability of 4HPR of 4HPR‐HSA PNPs in Suspension

4HPR, 4HPR‐HSA, and 4HPR‐HSA PNPs were suspended in water and loaded into a quartz cuvette for absorbance measurements. 4HPR, 4HPR‐HSA, and 4HPR‐HSA PNPs were kept from light at room temperature (n = 4). For 4HPR, 5% ethanol was used to ensure solubilization of the hydrophobic drug. On subsequent days, endpoint readings of the suspensions were made at the wavelength of 365 nm to determine the intensity of 4HPR. SEM images were taken via drop casting to assess the recrystallization of 4HPR in suspension.

4.9. Differential Scanning Calorimetry of 4HPR‐HSA PNPs

Glass transition temperature and melting point of 4HPR‐HSA PNPs and 4HPR were determined with Differential Scanning Calorimetry (DSC) (Discovery, TA instruments, New Castle, DE). The instrument was routinely calibrated with an indium standard under nitrogen during the preventative maintenance schedule. The samples were kept in a well‐capped container to avoid moisture buildup, and the nanoparticles were lyophilized to remove excess water prior to measurements. 2 mg of the samples were loaded onto a low‐mass aluminum pan, capped with an aluminum lid, and heated with the modulating temperature program from 20 to 250 °C under the ramping temperature of 2 °C min⁻¹. All measurements were done in triplicate, and the average was reported.

4.10. Mucin Binding Capacity With 4HPR‐HSA PNPs

The mucoadhesion properties of 4HPR‐HSA PNPs, FluoSpheres^TM Amine‐Modified Microspheres (cationic PS particles), and Fluoresbrite® Multifluorescent Microspheres (anionic PS particles) were studied by mixing 1 × 10¹⁰ particles with 75 ug mL⁻¹ of mucin for an incubation period of 30 min. The samples were centrifuged for 45 min at 21300 RCF, and the remaining supernatant was extracted for absorbance measurement at 255 nm using a UV spectrophotometer. A mucin solution prepared in identical conditions was used as a reference to evaluate adhesion activities. Statistics were obtained through one‐way ANOVA via the Tukey post hoc method, where *p ≤ 0.05, ** p ≤ 0.01, and *** p ≤ 0.001. Calculation of the average mucin mass (µg) per surface area is calculated via Equation (1):

$$\frac{m_{Mucin}}{4\times \pi \times r^2\times N_{particles}}$$ (1)

where m _Mucin is the mass of mucin removed from the supernatant and N_particles is the number of particles in the solution.

4.11. Immobilization of Mucin for Surface Plasmon Resonance

Mucin immobilization onto the high‐capacity carboxyl sensors proceeded following the protocol provided by the vendor (Nicoya Life Science) with slight modifications. A two‐channel system from OpenSPR‐XT was used to provide adhesion measurements and analysis. Briefly, 10 mM of hydrochloric acid was added at a flow rate of 150 µL min⁻¹ for surface cleaning. Surface activation was made through the injection of a mixture of 0.1 m 1‐ethyl‐3‐(3‐dimethylaminopropyl)‐carbodiimide (EDC) and N‐hydroxysuccinimide (NHS) at a flow rate of 20 µL min⁻¹ with 5 min interaction time. Mucin sourced from the porcine stomach, type II (100 µg mL⁻¹) was injected at a flow rate of 20 µL min⁻¹ to channel 2 with a 5 min interaction time and 1 h dissociation time to stabilize the baseline and ensure full binding of mucin to the surface. Channel 1 followed the same procedure simultaneously without the injection of mucin.

4.12. Mucoadhesion Studies Between 4HPR‐HSA PNPs and Mucin via Surface Plasmon Resonance

The adhesion properties of 4HPR‐HSA PNPs to mucin were accessed with surface plasmon resonance via OpenSPR‐XT. To convert the number of particles into molarity, we assume that each particle is one singular entity, and the conversion of particles mL⁻¹ can be calculated to molarity. Briefly, suspended nanoparticles were injected into the mucin‐conjugated surface at a flow rate of 20 µL min⁻¹ with a 5 min interaction and a 15 min dissociation time. The obtained data were analyzed and plotted with TraceDrawer (Uppsala, Sweden).

4.13. Quantification of 4HPR in 4HPR‐PNPs

Quantification of the loading efficiency of 4HPR‐HSA and the jetting yield of 4HPR‐HSA PNPs was achieved through absorbance measurements. Briefly, samples were suspended in a 50% (v/v) ethanol‐water mixture, followed by an addition of 0.1% (v/v) acetic acid. The samples were loaded into a 96‐well plate with standard solutions of 4HPR prepared in identical conditions. Absorbance reading was done at 365 nm (λ_max). The loading efficiency (LE) was calculated via the generated calibration curve and Equation (2):

$$\frac{m_{4HPR-measured}}{m_T}\times 100\%$$ (2)

where m _{4HPR − measured} is the measured mass of 4HPR calculated through the calibration curve and m_T the theoretical mass of 4HPR loaded. The jetting yield is calculated via Equation (3):

$$\frac{m_{4HPR-measured}}{m_J\times LE}\times 100\%$$ (3)

where m _J is the mass loaded into the jetting solution.

4.14. In Vitro Release Studies of 4HPR‐HSA PNPs

The nanoparticles were suspended in a glass vial containing 5% (v/v) ethanol in water at 37 °C. The vial was shaken continuously for 10 days. At predetermined time points, the sample was extracted and centrifuged for 50 min at 21500 RCF to spin down all the particles. The supernatant was extracted for measurement after flash‐frozen with liquid nitrogen and lyophilization. The freeze‐dried samples were dissolved in a 50% (v/v) ethanol‐water mixture, followed by 0.1% (v/v) acetic acid, and analyzed with UV–vis spectroscopy.

4.15. Determination of 4HPR‐HSA PNPs Bioactivity via Activation of the Apoptosis Execution Phase Enzyme, Caspase 3

Human oral squamous cell carcinoma (OSCC) SCC2095sc (ATCC CRL‐2095, derived from a 56‐year‐old male patient with a tongue OSCC) was cultured in Advanced DMEM supplemented with 1x Glutamax and 5% heat‐inactivated FBS (GIBCO; Life Technologies). SCC2095sc were seeded at 1 × 10⁶ cells well⁻¹ and treated with 1) Control (CTR) + drug‐free surface‐capped nanoparticle (scPNP), 2) bolus 5 µm 4HPR+ drug‐free HSA PNPs, 3) 6.18 × 10⁹ particles mL⁻¹ of 4HPR‐HSA PNPs (5 µM 4HPR released 24 h) at 37 °C, 5% CO₂ in 6‐well plate. After treatment, cell lysates were collected, and caspase‐3 enzyme activity was determined using the Caspase‐3 Activity Assay Kit (Cell Signaling Technologies, Danvers, MA) following the manufacturer's instructions. Caspase‐3 Activity Assay Kit detects fluorescent AMC dye produced from cleavage of Ac‐DEVD‐AMC by activated caspase‐3 in apoptotic cells. Recombinant human Cleaved Caspase‐3 protein (Active) (Abcam, Specific activity: ≥ 15 000 units mg⁻¹) was used as caspase‐3 assay stand curve. One unit of the recombinant caspase‐3 was the enzyme activity that cleaves 1 nmol of the caspase substrate DEVD‐AMC per hour at 37 °C in a reaction solution. Cellular activity was expressed in Units of activity per microgram of protein.

These caspase‐3 induction data were evaluated by a Kruskal‐Wallis ANOVA followed by a Dunn's multiple comparison post hoc test. Mean + s.e.m., CTR+ blank HSA PNPs, n = 5; 5 µM 4HPR+ blank HSA PNPs, n = 5; 4HPR‐HSA PNPs (5 µM/24 h), n = 9. Relative to the drug‐free control cells, both the 4HPR bolus and 4HPR‐HSA PNPs‐treated cells showed a statistically significant increase in caspase‐3 activity, ****p < 0.0001, ***p < 0.001, and *p < 0.05, respectively. No significant difference was detected in the comparison of the 4 HPR bolus delivery relative to the 4HPR‐HSA PNPs‐treated cells.

4.16. Statistical Analyses

Data analysis was used to determine whether there is a significant difference in the mean values across the groups (Origin Pro, Northampton, MA). A one‐way ANOVA followed by Tukey's post hoc test, where *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001, was used to assess the difference among geometric properties of the as‐jetted HSA and 4HPR‐HSA PNPs, size and zeta potential in suspension, bioavailability of 4HPR on day 2, mucin binding capacity of the nanoparticles, and binding parameters across different particles. Tukey's post hoc test was selected for pairwise comparison of selected groups within a set group.

For in vitro assessment of caspase‐3 induction, the ordinal data mandate a nonparametric approach. Hence, evaluations were done by a Kruskal‐Wallis ANOVA followed by a Dunn's multiple comparison post hoc test. For all statistical analyses, a p‐value of < 0.05 was considered statistically significant unless otherwise stated.

In the event that a statistical analysis demonstrated an even higher level of significance, e.g., p<0.001, the higher level of statistical significance was reported.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information: mabi70036‐sup‐0001‐SuppMat.docx

Chang A., Mae B., Pei P., et al. “Electrohydrodynamic Jetting of Mucoadhesive Protein Nanoparticles as a Chemopreventive Strategy for Oral Squamous Cell Carcinoma.” Macromol. Biosci. 25, no. 10 (2025): e00661. 10.1002/mabi.202400661

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.