In vivo Serum Enabled Production of Ultrafine Nanotherapeutics for Cancer Treatment

Abstract

Systemic delivery of hydrophobic anti-cancer drugs with nanocarriers, particularly for drug-resistant and metastatic cancer, remain a challenge because of the difficulty to achieve high drug loading, while maintaining a small hydrodynamic size and colloid stability in blood to ensure delivery of an efficacious amount of drug to tumor cells. Here we introduce a new approach to address this challenge. In this approach, nanofibers of larger size with good drug loading capacity are first constructed by a self-assembly process, and upon intravascular injection and interacting with serum proteins in vivo, these nanofibers break down into ultra-fine nanoparticles of smaller size that inherit the drug loading property from their parent nanofibers. We demonstrate the efficacy of this approach with a clinically available anti-cancer drug: paclitaxel (PTX). In vitro, the PTX-loaded nanoparticles enter cancer cells and induce cellular apoptosis. In vivo, they demonstrate prolonged circulation in blood, induce no systemic toxicity, and show high potency in inhibiting tumor growth and metastasis in both mouse models of aggressive, drug-resistant breast cancer and melanoma. This study points to a new strategy toward improved anti-cancer drug delivery and therapy.

Graphical Abstract

Q.M. and M. Z conceived various aspects of the project. Q.M. performed all experiments on synthesis and characterize of nanofibers and nanoparticles, in vitro and in vivo experiments, data collection and analysis. G. L. and V. K. P. assisted on materials synthesis and characterization. G.L performed in vitro and in vivo experiments. Z. R. S., S. C, H. W., and R. N. G. performed experiments on chemical and morphological characterizations of materials and worked on data analysis. Experiments were designed and manuscripts were written by Q. M. and M. Z. with input from other authors.

1. Introduction

Nanoparticles (NPs) have shown great promise to improve the safety and efficacy of existing therapeutic agents in cancer therapy.^1–4 A carrier with a small hydrodynamic size, high drug loading capacity, and good stability in blood is essential in nanomedicine to the delivery of an efficacious amount of drug to target cells.^5–8 These favorable properties facilitate in vivo navigation to target sites, prolong circulation in blood, and help overcome extra-and intra-cellular barriers for enhanced bioavailability. The NP size and its stability in blood are two of critical properties that dictate drug trafficking in vivo and biological fate.^9,10 Generally, the chance of the recognition and removal of NPs by human mononuclear phagocyte system significantly increase when NPs are greater than 100 nm, and the basal lamina of the kidneys has pores of approximately 10 nm to filter out small matters, which sets the size constraint for NPs to the range of 10–100 nm to ensure prolonged circulation in blood.^11,12 However, due to the strong hydrophobicity nature of majority anti-cancer drugs, it is a great challenge to achieve high drug loading while maintaining NPs sufficiently small and colloidally stable in blood.^13–15 As the particle size is made smaller, its surface area to volume ratio gets larger, and more of its hydrophobic drugs is closer to the surface of the particle to cause serum instability, which in turn hinders the ability to create nanoparticles of smaller sizes.

Paclitaxel (PTX) is one of the most effective therapeutics among all anti-cancer drugs and can be used against a variety of solid tumors, including breast cancer and melanoma among others.^16–19 Because of its low solubility in water, PTX is dissolved in Cremophor EL/dehydrated ethanol (i.e., commercial Taxol) to create an injectable solution used in the clinic, but the presence of Cremophor EL incurs severe side effects including hypersensitivity, neurotoxicity, and nephrotoxicity.^20,21 As such, NPs made from a variety of materials (such as polymers, metals, and ceramics) and in various forms (lipids, liposomes, and nanoparticles) have been investigated to deliver PTX without use of toxic solvents; however, these PTX-loaded NPs are mostly greater than 100 nm in size,^18,22,23 and yet show limited a drug loading capacity, largely due to the hydrophobicity of the drug that makes NP incompatible with aqueous blood. Abraxane, a PTX protein-bound NP formulation has been used clinically for breast cancer therapy, but has a size of 130 nm.²⁴ Despite the improved patient outcome made with Abraxane, its use is limited to the advanced breast cancer due to the severe toxic effects to healthy tissues.¹⁶ This severe toxicity is believed to be primarily caused by the low drug accummulation in tumor due to large size, which necessitates the an excessive amount of Abraxane to be administered. Furthermore, NPs loaded with hydrophobic drugs readily interact with serum proteins and tend to grow in size due to the protein adsorption, which alters their physicochemical property and pharmacological profile that were originally designed to facilitate in vivo navigation and tumor cell targeting.

Unstable NPs could quickly aggregate and exhibit physicochemical properties far different from those of their un-aggregated counterparts. NPs for drug delivery are commonly used in physiological conditions (such as serum) that are very different from the conditions in which they were synthesized. The change in environmental conditions often results in an elevated surface energy and thus impairs their colloidal stability. Despite the utilization of various stabilizers, many NPs are still not stable enough in storage and/or in biological relevant media to serve as effective drug carriers.²⁵

Here, we present a new approach to address the limitations discussed above. In this approach, PTX-loaded chitosan-polyethylene glycol (CHI-PEG-PTX) nanofibers are synthesized via a self-assembly process (Scheme 1a), and upon systemic administration, the nanofibers break down in vivo into smaller NPs by interaction with serum proteins (Scheme 1b). These NPs retain the good drug loading characteristic and intended functionality, both inherited from their parent nanofibers. The CHI-PEG-PTX nanofibers are 10 nm in diameter and 0.1–2 μm in length and the produced CHI-PEG-PTX NPs, with serum proteins adsorbed on their surfaces, have a diameter of 20.6 ± 2.8 nm, which falls well into the range of the size constraint for in vivo nanocarriers. This strategy addresses the issues of limited drug loading capacity and poor colloidal stability of small-size NPs (< 100 nm) and delivers improved therapeutic effects in cancer treatment. The nanoparticles demonstrate a prolonged blood circulation time, minimal systemic toxicity, and high potency in inhibiting tumor growth in both mouse models of aggressive and drug-resistant breast cancer and melanoma. CHI-PEG-PTX nanoparticles significantly outperformed the clinically-used drug formulation of paclitaxel in terms of minimized toxicity to health tissues and potency in treating breast cancer. This is the first demonstration that intravascular injected NP with anticancer drug alone can inhibit metastasis of highly aggressive 4T1 breast cancer in an orthotopic model that mimic stage IV metastatic breast cancer in human.

Scheme 1.

Schematic illustration of a, CHI-PEG-PTX nanofiber (NF) synthesis; and b, in situ fragmentation of NFs into NPs by interaction with serum proteins.

2. Results and Discussion

2.1. Preparation and characterization of CHI-PEG-PTX nanofibers.

We designed our nanocarrier with all components to be as biocompatible as possible. Chitosan (CHI), a natural polymer, is the primary material of our nanocarrier (CHI-PEG-PTX). Chitosan is a versatile pharmaceutical ingredient that is both biocompatible and biodegradable. Chitosan has been widely used as scaffolding materials in tissue engineering and as drug delivery vehicles.^26–31 Its abundant primary amine groups allow for easy chemical conjugation with other molecules such as PEG and various therapeutic agents.

Short chain CHI (Mw = 3900) was modified with PEG to produce a stable and soluble CHIPEG carrier for delivery of hydrophobic drug PTX. PTX was carboxylated to form 2’-succinyl PTX (PTX-COOH) by succinic anhydride under basic conditions (Fig. 1a). Because the C2’ ester bond is susceptible to hydrolysis, PTX is expected to be released only in tumor cells where pH is low.³² The formation of PTX-COOH was characterized with mass spectrometry (m/z 954.7, positive mode), and the result confirmed the successful modification of PTX with COOH (Supplementary Information,Fig. S1). PTX-COOH was then covalently bonded to CHI-PEG to form CHI-PEG-PTX through an EDC/NHS reaction in a sodium bicarbonate (pH 8.5)/dimethyl sulfoxide (DMSO) solution (Fig. 1a). EDC/NHS activated PTX-COOH reacted with primary amines on CHI-PEG to form amide bonds.³³

Figure 1.

Synthesis and chemical characterization of CHI-PEG-PTX. a, Schematic representation of synthesis of CHI-PEG-PTX. b, FT-IR spectra of CHI-PEG, PTX and CHI-PEG-PTX. c, ¹H NMR spectra of CHI-PEG, PTX and CHI-PEG-PTX.

After purification, the covalent bonding between PTX and chitosan-PEG (CHI-PEG-PTX) was characterized by FT-IR and NMR. The FT-IR spectrum of CHI-PEG-PTX shows an additional peak at 1732 cm⁻¹ (-C=O) and increased signal intensity at 1653 cm⁻¹ (-C-C- and -CO-NH-) as compared to the spectrum of CHI-PEG, indicating the successful conjugation of PTX (Fig. 1b). The NMR spectrum of CHI-PEG-PTX shows addition of aromatic protons (7.1–8.1 ppm) which are not present in CHI-PEG (Fig. 1c), further confirming the successful conjugation of PTX.³⁴ The PTX loading capacity on CHI-PEG-PTX was determined to be 8.45 wt% utilizing a quantitative NMR method (Supplementary Information, Fig. S2). The drug loading information obtained here was used to determine the drug dose in subsequent biological evaluations. The strong covalent bonding between PTX and chitosan-PEG cannot be cleaved by interaction with serum in the body at room temperatures.

After the removal of DMSO from the CHI-PEG-PTX product mixture, the solution appeared as a turbid suspension. The CHI-PEG-PTX suspension was ultrasonicated for 10 min (10 seconds on, 5 seconds off, 40% power amplitude) until the suspension turned clear, indicating the reassembly of CHI-PEG-PTX molecules. The CHI-PEG-PTX solution was then stored at 4°C for 12 h to form CHI-PEG-PTX nanofibers (Scheme 1a). The morphology of CHI-PEG-PTX nanofibers were characterized by TEM. The TEM image revealed a nanofibrous structure with an average fiber diameter of 10 nm and length of 0.1–2 μm (Fig. 2a). We further imaged the samples at higher magnification with negative staining using uranyl acetate. The TEM images with negative staining revealed more details of a needle-shaped structure of nanofibers (Fig. 2c).

Figure 2.

Physicochemical properties of CHI-PEG-PTX nanofibers and nanoparticles. TEM micrographs of (a and c) CHI-PEG-PTX nanofibers, acquired (a) without and (c) with negative staining, and (b and d) CHI-PEG-PTX NPs formed in 10% serum, acquired (b) without and (d) with negative staining. Insets: enlarged images. (e) Hydrodynamic size and (f) zeta-potential of (i) 10% serum, (ii) CHI-PEG-PTX nanofibers (CHI-PEG-PTX NFs), and (iii) CHI-PEG-PTX NPs in 10% serum. (g) Z-average size of CHI-PEG-PTX NPs incubated in 10% serum over a 30-min period.

2.2. Serum protein induced fragmentation of CHI-PEG-PTX nanofibers into CHI-PEG-PTX nanoparticles.

To study protein adsorption onto CHI-PEG-PTX nanofibers and determine the size and morphology of these nanomaterials upon interaction with serum proteins, CHI-PEG-PTX nanofibers were examined by TEM after incubation with fetal bovine serum at 37°C for 10 min. TEM images, acquired with and without negative staining, of serum-incubated CHI-PEG-PTX nanofibers were compared with the images obtained from untreated CHI-PEG-PTX nanofibers. Notably, CHI-PEG-PTX nanofibers completely disappeared after 10-min incubation in serum and, instead, small particles were observed (Fig. 2b, d). To understand the mechanism underlying this fiber-to-particle transformation, we measured the hydrodynamic sizes and zeta-potentials of serum protein, and CHI-PEG-PTX nanofibers and CHI-PEG-PTX NPs in water. The serum proteins had a size of 15.8 ± 4.4 nm which accounts for all types of serum proteins and their small aggregates. CHI-PEG-PTX nanofibers (NFs) had a size of 565 ± 85.6 nm (Fig. 2e). Notably, CHI-PEG-PTX nanoparticles (NPs) had a size of 34 ± 13.2 nm in the first minute of the incubation in serum, which was reduced to 20.6 ± 2.8 nm in the next 25 min (Fig. 2g). The rapid drop in size in the first minute of the incubation indicates a quick nanofiber destruction by serum protein molecules, which was followed by a slow adjustment process to yield ultra-small, stable NPs. The zeta potential of CHIPEG-PTX nanofibers was slightly positive, presumably resulted from the residue free amines from the serum adsorbed on the CHI backbone. After the nanofiber breakdown in serum and subsequent protein adsorption, the zeta potential of resultant CHI-PEG-PTX NPs is −5.84 mV (Fig. 2f), close to the zeta potential of serum proteins (−5.71 mV), indicating adsorbed proteins dominate the surface charge. As a comparison, we synthesized large CHI-PEG-PTX nanoparticles (denoted thereafter by CHI-PEG-PTX LNPs) by a conventional method by which CHI-PEG-PTX was dispersed in water and subjected to short ultrasonication (without sufficient time to self-assemble into nanofibers) and then purified by centrifugation. CHI-PEG-PTX LNPs thus produced had an average hydrodynamic diameter of 112 nm (Supplementary Information, Fig. S3a) and were unstable in serum with diameter increased to more than 200 nm in 10 min (Fig. S3c). Although the CHI-PEG-PTX LNPs were stable after sonication and centrifugal purification, these LNPs had a strong tendency to interact with serum proteins upon introducing into media containing serum due to the presence of the positively charged chitosan groups and hydrophobic paclitaxel molecules on chitosan. This interaction resulted in an increase in their hydrodynamic size. Similar observation is reported by a recent study with chitosan nanoparticles of different sizes.³⁵ The CHI-PEG-PTX nanofibers have a zeta potential higher than CHI-PEG-PTX NPs (Fig. S3b) and their size in serum decrease over time (Fig. S3c).

Nanomaterials can interact with proteins through a number of mechanisms including electrostatic, hydrophobic, and hydrogen bonding interactions.³⁶ We believe that serum proteins bind onto CHI-PEG-PTX nanofibers primarily by both electrostatic interaction between negatively charged protein and positively charged CHI backbone and the hydrophobic interaction between PTX and serum albumin.³⁷ These interactions may disrupt the intermolecular hydrogen bonding that maintains the fibrous structure of CHI-PEG-PTX nanofibers, resulting in the decomposition of the nanofibers. To examine the role of hydrogen bonding in the CHI-PEG-PTX nanofibers, we freeze-dried and then thawed CHI-PEG-PTX nanofibers in water to dissolve them in water (which disrupts the hydrogen bonding), or freeze and then thawed them without water (non-hydrogen bonding disruption). As shown in Fig. S4a (Supplementary Information), CHI-PEG-PTX nanofibers lost their fibrous structure after freeze-drying and re-dissolving in water. Conversely, the frozen and thawed CHI-PEG-PTX nanofibers retained their fibrous structure (Fig. S4b). These results suggest the involvement of water molecules provides the hydrogen bonding in the formation of the nanofibrous structure. The freeze-drying removed water molecules and thus disrupts water and nanofiber interaction while freezing did not remove water, i.e., did not break the hydrogen bonds.

The long-term stability of the nanofibers in aqueous solution was confirmed by TEM examination of the fiber structure in water over time. The nanofibers retained their fibrous structure after they were stored in water for at least 2 months. The nanofibers were then immersed in serum and examined by TEM, and as shown in Fig. S5 (Supplementary Information) the serum-triggerable fragmentation property of the nanofibers were retained over this period.

2.3. Therapeutic effect of CHI-PEG-PTX in vitro.

We modified CHI-PEG-PTX nanofibers with NHS-Cy5 to form CHI-PEG-PTX-Cy5 nanofibers which allow for fluorescence imaging. The morphology of the nanofibers remained unchanged after NHS-Cy5 conjugation as examined by TEM (Supplementary Information, Fig. S6). 4T1 cancer cells were incubated with CHI-PEG-PTX-Cy5 in cell culture medium for 1 h. Because these NFs were converted to NPs immediately after dispersion in serum protein-rich cell culture media (<1 min), it was NPs, rather than NFs, that interacted with cells. The confocal microscopic images of the cells incubated with CHI-PEG-PTX NPs for 1h clearly show the uptake of NPs into the cells and their localization in the cytoplasm or endosomes of the cells, as evidenced by the strong fluorescence signal observed at the boundary between the nuclear envelope and plasma membrane (Fig. 3a(iv)). We have shown earlier that CHI-PEG-PTX nanofibers had a slightly positive charge (0.8 mV) at pH 7.4, and CHI-PEG-PTX NPs had a zeta-potential of −5.8 mV (Fig. 2f). A negatively-charged NP complex likely enters cells through endocytosis.^38,39 As endocytosis is an energy-dependent process,⁴⁰ we examined cellular uptake of CHI-PEG-PTX NPs at 4°C and 37°C by confocal microscopy and flow cytometry. The fluorescent images acquired showed that the signal from CHI-PEG-PTX NPs was significantly lower at 4°C (Fig. 3a(vii) than at 37°C (Fig. 3a(iii)). The quantification by flow cytometry further revealed that the cellular uptake of the NPs at 4°C was nearly 80% lower of that at 37°C (Fig. 3b and c). These results suggest that CHI-PEGPTX NPs entered cells primarily through an energy-dependent endocytosis process.

Figure 3.

Cellular uptake and cell kill of CHI-PEG-PTX NPs. a, Confocal microscopic images of 4T1 cells incubated for 1 h with Cy5-labeled CHI-PEG-PTX at 4°C (middle row) or 37°C (top row), with un-treated cells as reference (bottom row). b, Flow cytometric analysis of CHI-PEGPTX uptake into 4T1 cells after 1 h of incubation at 37°C or 4°C. c, Mean fluorescence intensity from flow cytometry analysis shown in b. ****P < 0.0001 by one-way ANOVA with Turkey’s post-hoc test between each pair of columns. d, Flow cytometric analysis of apoptosis of 4T1 cells after incubation with (i) cell culture medium, (ii) CHI-PEG, (iii) free PTX, and (vi) CHI-PEG-PTX.

To investigate the therapeutic efficacy of CHI-PEG-PTX on tumor cells, we assessed the apoptosis of 4T1 cells after they were incubated with CHI-PEG-PTX nanofibers for 48 h (PTX concentration in cell culture medium: 1 μM). Another two cell samples, as controls, were treated either with free PTX having a concentration equivalent to the PTX concentration of CHI-PEG-PTX or with CHI-PEG (i.e., without PTX). 4T1 control cells had a slight apoptosis of 6.9%, which is likely due to aggressive growth and rapid metabolism, and is consistent with the typical apoptosis range of 4–10% for 4T1 cells without any treatments reported in the literature.^41–43 CHI-PEG induced slightly apoptosis of 13.5%. Free PTX induced significant apoptosis of 4T1 cells (63.4% late apoptotic cells). CHI-PEG-PTX NPs also induced significant apoptosis (52.6% late apoptotic cells) but slightly less than that induced by free PTX (Fig. 3d). This was expected, since additional steps are involved to release the PTX on CHI-PEG-PTX NPs to induce cytotoxicity as compared to free PTX. Specifically, after the cellular internalization of CHI-PEG-PTX NPs, esterases hydrolyze the ester bond between PTX and CHI-PEG to release the PTX to the cytoplasm.⁴⁴ Once released from CHI-PEG-PTX, PTX binds to microtubules, triggers mitosis arrest (G2/M phase) and induces cellular apoptosis.⁴⁵ By way of contrast, free PTX enters cancer cells and binds microtubules directly. It is worth noting that despite the slightly higher therapeutic potency exhibited by free PTX, PTX is a highly hydrophobic molecule. The in vivo administration of PTX requires to dissolve PTX in a solution with both a surfactant and an organic solvent (e.g., Cremophor EL and ethanol, respectively), which both are toxic to healthy tissues and not well-tolerated by patients.⁴⁶ Conversely, CHI-PEGPTX NFs readily dissolve in PBS and cell culture media, and the resultant solution can be safely administered in vivo.

2.4. Blood circulation, biodistribution, and systemic toxicity of CHI-PEG-PTX assessed in wild-type mice

CHI-PEG-PTX was conjugated with Cy5 for fluorescence imaging. The CHI-PEG-PTX-Cy5 nanofibers was injected intravascularly into wild-type mice, blood was drawn at 2, 6, 18, 24, and 48 h post-injection, and the fluorescence intensity of Cy5-conjugated CHI-PEG-PTX was measured. Another two groups of mice, as controls, were injected with either PTX dissolved in Cremophor EL+ethanol (PTX-crem) or in saline. Cremophor-EL+ethanol is commonly used to dissolve hydrophibic PTX to create a mixture solution for injection (e.g., a commercial version is Taxol).¹⁸ The use of Taxol is limited due to its severe toxicity, short-term physical stability, and tendency to precipitate out from the aqueous media.²² The blood half-life of CHI-PEG-PTX was determined to be 24.9 h by measurements of the fluorescence intensity over time (Fig. 4a), which was significantly longer than the half-life (1.31 h) of PTX-crem reported previously).⁴⁷ Organs were collected at 6 and 48 h, respectively, after administration of CHI-PEG-PTX, and imaged with an IVIS 200 imaging system (Fig. 4b and c). At 6 h, CHI-PEG-PTX was mostly present in liver and kidneys with minimal accumulation in lungs, spleen and brain. At 48 h, the signal in all organs largely decreased, due apparently to the elimination by the reticuloendothelial system. The CHI-PEG-PTX NPs showed a robust ability to cross endothelial barriers as indicated by signals in major organs. This favorable property may be attributed to its ultrafine size, transcytotic properties of chitosan groups,⁴⁸ or nanoparticle induced endothelial leakiness (“nanoEL”) effect observed recently.^49,50

Figure 4.

Pharmacokinetics, biodistribution and toxicology evaluation of Cy5-labeled CHI-PEGPTX in wild-type mice. a, CHI-PEG-PTX concentration in blood determined by fluorescence intensity measurements over time to yield the blood clearance profile (n = 3 mice per time point). b, Ex vivo measurements of the fluorescence intensity from various organs at 6 h and 48 h post-injection, measured using an IVIS 200 imaging system. c, Fluorescence images of organs collected from mice at 6 and 48 h post-injection of CHI-PEG-PTX. The scale bar displays counts (or relative light units). d, Toxicity evaluation of CHI-PEG-PTX from blood samples 24 h post-injection, with PTX-crem and Saline as controls. *** P < 0.0005, *P < 0.05 by one-way ANOVA with Turkey’s post-hoc test.

To assess potential toxic effects of CHI-PEG-PTX, we measured levels of red blood cells (RBC), hemoglobin (HGB), aspartate transaminase (AST), alanine transaminase (ALT) and blood urea nitrogen (BUN) collected from healthy mice 24 h after administration of either CHI-PEG-PTX nanofibers or PTX-crem. Another group of mice injected with saline were also included in the study as control. As shown in Fig. 4d, no significant differences in these levels were observed between mice treated with saline control and with CHI-PEG-PTX. However, mice treated with PTX-crem exhibited notable decreased levels of ALT and AST compared to mice treated with saline (p < 0.05), indicating that PTX-crem caused acute hepatic toxicity to mice, which agrees with the result reported previously.⁵¹ A slight decrease in blood urea nitrogen (BUN) was also observed in the PTX-crem treated group, indicating acute renal toxicity of PTX-crem.⁵² These results indicate that, in contrast to PTX-crem formulation, CHI-PEG-PTX does not cause acute toxicity and may serve as a safer clinical option.

2.5. Inhibition of tumor growth by CHI-PEG-PTX in a xenograft mouse model of breast cancer.

The efficacy of CHI-PEG-PTX on inhibiting tumor growth was evaluated in a breast cancer mouse model. 4T1 breast adenocarcinoma cells were used to mimic stage IV metastatic breast cancer in human. 4T1 cells are highly aggressive and can spontaneously metastasize to various organs when injected into BALB/c mice.⁵³ The cells also have a “triple negative” phenotype (lack of ER, PR and HER2 receptors on cell surface) and are insensitive to chemotherapeutics, such as PTX, making it difficult to target and treat.^54–56 Mice were inoculated subcutaneously with 4T1-luc cells at the flank.⁵⁷ On day 4 after inoculation, 5 groups of mice were administered with CHI-PEGPTX nanofibers, PTX-crem, CHI-PEG-PTX LNPs, CHI-PEG, and saline, respectively. Each mouse received 5 sequential injections, each every 3 days (10 mg/kg PTX equivalent for each injection, a total dose of 50 mg/kg).The PTX dose was chosen based on our previous studies in which a similar PTX dose provided appreciable strong tumor growth inhibition.^58,59 The bioluminescence signal intensity of the tumor, mouse body weight, and mouse behavior were recorded every 3 days during a 3-week period. All mice were euthanized 3 weeks after the initial treatment when significant metastasis was observed. In untreated control groups bearing 4T1-luc tumors, the bioluminescence intensity acquired was normalized to the intensity measured at the time point immediately before the first treatment.

As shown in Fig. 5a, the tumors from the saline and CHI-PEG treated control groups (Columns i and ii) grew faster and become considerably larger than those from other groups. Quantitatively, a near 90% inhibition of tumor growth was observed in the group treated with CHI-PEG-PTX nanofibers 21 days after initial treatment as compared to the control group treated with saline (p < 0.05) (Fig. 5b). The PTX-crem treatment resulted in 66% inhibition of tumor growth. No apparent body weight change and abnormal behavior were observed in any of these treatment groups (Fig. 5c). The current prevailing clinical formulation of PTX is PTX+crem, which can give rise to short-term and long-term systemic toxicity due to oxidative stress.⁴⁶ CHI-PEG-PTX NFs developed in this study circumvent the need for toxic ingredients and yet outperform PTX-crem in treating tumors as demonstrated here.

Figure 5.

Tumor growth inhibition in a mouse breast cancer model. a, IVIS images of live mice bearing 4T1-luc tumors, treated with various agents: i. saline; ii. CHI-PEG; iii. PTX-crem; iv. CHIPEG-PTX LNPs; and v. CHI-PEG-PTX nanofibers (CHI-PEG-PTX NFs). b, Bioluminescence intensities acquired from differently treated mouse groups (n = 4 per group). **** P < 0.0001, * P < 0.05 by one-way ANOVA with Turkey’s post-hoc test. c, Body weights of mice of various treatment groups (n = 4 per group), measured over 21 days. d, Optical images of H&E stained tumor tissue sections harvested from mice treated with (i) PBS and (ii) CHI-PEG-PTX NFs. The scale bar represents 75 μm.

Histological images of H&E stained tumor sections (Fig. 5d) show that with CHI-PEGPTX NF treatment, the tumor cell morphology was altered and the tumor cell nuclei became irregular-shaped. Several “bubbled” regions were observed in tumor tissues, similar to those observed in the studies using PTX as a therapeutic agent.^60,61

We further quantified the accumulation of treatment agents in tumors 24 h after the systemic injections. Cy5.5 was used to label PTX, CHI-PEG-PTX LNPs, and CHI-PEGPTX NFs for fluorescence imaging. Tumors were collected and near IR fluorescence intensities from these tumors were measured. The results showed that CHI-PEG-PTX NFs accumulated the most in tumors (Fig. 6a, b), which may explain the best tumor inhibition efficiency exhibited by the NFs as shown in Fig. 5b.

Figure 6.

Assessment of therapeutic agent accumulations in tumors and tumor metastasis inhibition in an orthotopic breast cancer model. a, IVIS fluorescence images of tumors collected from mice 24 h after receiving systemic injection of Cy5.5-labeled agents: i: Saline; ii: PTX-crem; iii: CHIPEG-PTX LNPs; or iv: CHI-PEG-PTX NFs. The injections were made 10 days after tumor inoculation. b, Fluorescence intensities acquired from tumor areas shown in A. * P < 0.05, **P < 0.005, *** P < 0.0005 by one-way ANOVA with Turkey’s post-hoc test. c, Bioluminescence images of primary tumors and metastasis in mice 3 weeks after tumor inoculation. Arrows point to the metastasis sites. Each mouse received four sequential intravascular injections (each every 3 days) one week after tumor inoculation with one of the following agents: i: Saline; ii: PTX-crem; iii: CHI-PEG-PTX LNPs, iv: CHI-PEG-PTX NFs. PTX dose in all agents except saline was equivalent to 10 mg/kg per injection.

2.6. Inhibition of tumor metastasis by CHI-PEG-PTX in an orthotopic mouse model of breast cancer.

An orthotopic mouse model of spontaneous breast cancer metastasis was created by direct injection of 4T1-luc cancer cells into mammary glands in wild type BALB/c mice. 4T1 cells in the orthotopic model have more tendency to metastasize than in the flank model and mimic late stage breast cancer. One week after the inoculation, each mouse received four sequential intravascular injections with one of the following agents (10 mg/kg in each injection and every 3 days): saline, PTX-crem, CHI-PEG-PTX LNPs and CHI-PEG-PTX NFs. Three weeks after the inoculation, mice receiving saline, PTX-crem and CHI-PEGPTX LNPs started showing bioluminescence signals from cancer cells at distant sites (Fig. 6c) from the original tumors. In contrast, the tumors in the CHI-PEG-PTX NF-treated mice grew slightly but no sign of metastasis was observed. Apparently, the treatment with CHIPEG-PTX NFs suppressed or inhibited the tumor growth and metastasis.

2.7. Inhibition of tumor growth by CHI-PEG-PTX in a mouse model of xenograft melanoma.

We expected that our approach would target cancer types that PTX is designed to treat. As a demonstration of its wide applicability, the efficacy of CHI-PEG-PTX to treat another cancer type was evaluated: melanoma in a mouse xenograft melanoma model. B16-F10 cells were inoculated into C57BL/6 mice; B16-F10 tumors are strongly resistant to chemotherapy, mimicking human melanoma.⁶² Mice bearing melanoma were randomly divided into five groups with each group intravascularly injected with one of the following agents: saline (control), PTX-crem, CHI-PEG-PTX-LNPs and CHI-PEG-PTX nanofibers 1 week after the tumor inoculation. A total of four sequential injections were given to each mouse with each in every three days. The equivalent PTX dose in all the agents except saline is 10 mg/kg per injection. By measuring tumor sizes in treated mice over time, we found that PTX-crem and CHI-PEG-PTX-LNPs exhibited slight tumor growth inhibition similar to each other, whereas CHI-PEG-PTX nanofibers inhibited tumor growth by more than 75% at 22 days after tumor inoculations (Fig. 7a). None of the mice showed observable weight loss (Fig. 7b). H&E staining showed tumor cells appeared abnormal in shape and there were bubble-like features in tumor tissue of mice treated with CHI-PEG-PTX nanofibers (Fig. 7c).

Figure 7.

Tumor growth inhibition by CHI-PEG-PTX demonstrated in a melanoma mouse model. a, Tumor growth curves for mice bearing B16/F10 tumors and treated with various agents (4 treatment groups and n = 4 per group). **** P < 0.0001 by one-way ANOVA with Turkey’s post-hoc test. b, Body weights of mice in four treatment groups, measured over a period of 15 days. c, Images of H&E stained tumor sections from mice treated with (i) PBS and (ii) CHI-PEG-PTX NFs. The scale bar represents 50 μm.

3. Conclusion

We have introduced a drug delivery system that overcomes a number of challenges encountered by many current nanocarriers, such as low drug loading capacity, incompatibility with aqueous solutions, and colloidal instability in serum. These challenges have been the primary sources that lead to low drug accumulation in tumor. The ultra-small and stable PTX-loaded NPs were obtained in vivo by serum-interaction with nanofibers that are self-assembled in vitro. The PTX-loaded NPs demonstrated a prolonged blood circulation time, the enhanced accumulation in tumors, and effective inhibition of tumor growth in both a xenograft mouse model of breast cancer and an orthotopic mouse model of aggressive breast cancer metastasis as compared to free PTX and larger PTX nanoparticles. Unlike clinically-used PTX drug, the administration of this nanotherapeutics does not involves toxic organic surfactants or solvents, which eliminates the side effects that severely limit the use of these hydrophobic drugs. The applicability of this nanotherapeutics was further demosntrated in an aggressive and drug-resistant xenograft melanoma mouse model, where it showed inhibition of tumor growth significantly higher than clinically used PTX formulation as well as large-size NPs of the same material loaded with PTX. In summary, our study demonstrated a new strategy that holds technological promise to advance the treatment of cancer with nanomedicine.

4. Materials and methods

4.1. Materials.

Chitosan (MW3900) was purchased from Acmey Industrial Co., Ltd. (Shanghai, China). PTX was purchased from LC Laboratories (Woburn, MA). The wheat germ agglutinin-Alexa Fluor 555 conjugate was purchased from Life technologies (Grand Island, NY). The FITCAnnexin V reagent was purchased from BD Biosciences (San Jose, CA). Xenolight D-luciferin was purchased from PerkinElmer, Inc. (Waltham, MA). All other chemical reagents were purchased from Sigma-Aldrich (St. Louis, MO).

4.2. Synthesis of PTX-COOH.

Synthesis of PTX-COOH was adapted from previously published method³³ with modifications. PTX (50 mg) and succinic anhydride (11.8 mg) were dissolved in 6 mL chloroform containing 56.9 μL pyridine. The solution was stirred for 24 h at room temperature and dried in a vacuum oven overnight. The dried sample was washed 3 times with deionized water under centrifugation (4000 g, 2 min). The sample was then dissolved in 2 ml acetone and transferred into a 15-ml conical tube. Deionized water was slowly added into the acetone solution until the solution turned turbid (10 ml water was added). The suspension was centrifuged for 10 min at 1000 g. The supernatant was removed and 2 mL of deionized water was added to the resuspended PTXCOOH. The suspension was then freeze-dried.

4.3. Synthesis of CHI-PEG-PTX and CHI-PEG-PTX-Cy5.

CHI-PEG was synthesized following an established procedure.⁶³ PTX-COOH (4.3 mg), EDC (3.2 mg) and NHS (1.6 mg) were dissolved in 0.5 mL DMSO in a microtube. The mixture was incubated for 3 h on a rocker. CHI-PEG (20 mg) was dissolved in 0.8 mL sodium bicarb buffer (pH 8.5) in a 20-mL glass vial and diluted with 1.6 mL DMSO. The PTX-COOH/EDC/NHS DMSO solution was then added into the CHI-PEG solution under stirring. The mixture was stirred overnight. Deionized water (8 mL) was slowly added into the mixture under stirring to reduce DMSO percentage. The solution was then dialyzed against deionized water (1 L) for 24 h using 1 kDa RC dialysis tubing. Water was changed 3 times at 1, 3, and 7 h, respectively, after the dialysis was started. The solution turned turbid after the dialysis. All samples (including precipitates) were transferred into a 20-ml glass vial and sonicated for 10 min using a Sonic Dismembrator Model 500 (Fisher Scientific, Pittsburgh, PA) (40% amplitude, 10 seconds on and 5 seconds off). The solution was then centrifuged at 20000 g for 10 min. The supernatant was collected and stored in a fridge. For conjugation of Cy5, 0.3 mL 10× PBS buffer (pH 7.4) was added into 2.7 mL CHI-PEG-PTX solution. 10 μL NHS-Cy5 (5 mg mL⁻¹) was then added into the mixture solution of PBS and CHI-PEG-PTX and the resultant solution was incubated for 2 h on a rocker at room temperature. The solution was then dialyzed against DI water for 24 h with change of water for 3× using 1 kDa RC dialysis tubing. For animal study, the CHIPEG-PTX solution stored in the fridge overnight was concentrated 3.5 times using 30 kDa spin filters.

4.4. Preparation of Cy5.5 labeled PTX-crem.

0.1 mg PTX-COOH, 71 μg Cy5.5-hydrazide (2 μL out of 50 mM solution in DMSO) and 0.1 mg EDC were reacted in 25 μL ethanol for 2 h. 0.9 mg PTX was then dissolved in the mixture solution followed by addition of 25 μL Cremophor EL. Finally, the solution was diluted by 950 μL of PBS to reach 1 ml of final volume. The Cy5.5/PTX ratio (1/9) was equal to those in CHI-PEG-PTX LNPs and NFs.

4.5. Characterizations:

4.5.1. FT-IR.

FT-IR spectra were acquired on a Nicolet 6700 spectrometer (Thermo Scientific Inc., Waltham, MA). Absorbance spectra were acquired at 4 cm⁻¹ resolution and signals were averaged over 32 scans. Samples were mixed with KBr and pressed into a pellet for analysis.

4.5.2. Mass spectroscopy.

Mass spectroscopy spectra of PTX-COOH were acquired on a Bruker Esquire ion trap mass spectrometer (Bruker Daltonics, Billerica, MA) using positive ion mode. The nitrogen pressure and flow rate on the nebulizer were 30 psi and 10 Lmin⁻¹, respectively, and the drying gas temperature was 350°C. The capillary voltage was 3.5 kV. The scan range was set at m/z 50–1750.

4.5.3. ¹H NMR.

All experiments were performed on a 500 MHz Bruker AV Series spectrometer (Karlsruhe, Germany) operating at room temperature and ¹H frequency of 499.956 MHz. CHI-PEG and CHI-PEG-PTX were dissolved in D₂O, respectively, and PTX was dissolved in DMSO-d₆ for NMR measurements. Scanning parameters were NS = 96, AQ = 2.34 s and TD = 32,678. For quantification of PTX in CHI-PEG-PTX, samples were mixed with a known amount of 1,4-Bis[(trimethylsilyl)ethynyl]benzene (BTMSB) and diluted to 10 mg/mL in DMSO before spectra were captured (NS = 128, AQ = 0.66 s, TD = 60,056) using selective irradiation to negate the contribution of aliphatic protons from the PEG polymer to the spectra.^64–66 Peaks known to only show contributions from BTMSB (δ = 7.20, 4H) and peaks known to show only contributions from PTX (δ = 7.75, 2H) were integrated.³⁴ From these values, the absolute concentration of PTX in solution was calculated with the following equation:

$$m_D=\frac{m_{IS}{P}_{IS}{I}_D{M}_D}{M_{IS}{P}_D{I}_{IS}}$$ (1)

where m_D is the mass of PTX, m_IS is the mass of the internal standard (BTMSB), P_D and P_IS represent the numbers of protons responsible for the resulting peak in the drug and the internal standard, respectively; I_D and I_IS represent the integration of the peaks of the drug and the internal standard, respectively; M_D and M_IS represent the molar mass of the drug and the internal standard, respectively.

4.5.4. TEM.

For negative stain TEM imaging, the sample solution (4 μL) was transferred onto a TEM grid (copper grid, 300-mesh, coated with carbon and Formvar film). The solution was sucked by a filter paper 5 min later and the grid was stained with 4 μL 5% uranyl acetate. After one minute, the staining solution was sucked by a filter paper and the grid was air-dried. Regular non-stained images were similarly obtained without uranyl acetate staining. All images were acquired on a Tecnai G2 F20 electron microscope (FEI, Hillsboro, OR) operating at 200 kV.

4.5.5. DLS.

The hydrodynamic size and ζ-potential of samples were measured using a Zetasizer Nano-ZS (Malvern Instruments, Worcestershire, UK) at room temperature. For hydrodynamic size characterization of CHI-PEG-PTX NPs, CHI-PEG-PTX nanofibers were incubated with 10% serum in DI water and the measurements were made immediately at 25°C.

4.6. Cell culture.

4T1 and 4T1-luc cells were provided by Dr. Lingfen Liu in Fred Hutchinson Cancer Research Center. Cells were grown in RPMI-1640 media supplemented with 10% fetal bovine serum and 1% antibiotic-antimycotic (Life technologies, Grand Island, NY). Cells were cultured in an incubator maintained at 37°C, 5% CO₂ and 95% humidity.

4.6.1. Characterization of cellular uptake by confocal laser scanning microscopy.

4T1 cells were seeded onto glass-bottom petri dishes (Mattech). After overnight incubation, cells were incubated with CHI-PEG-PTX nanofibers (50 μg mL⁻¹) for 1 h at either 37°C or 4°C. Cells were washed 3 times with cold PBS, fixed with 4% paraformaldehyde for 15 min at 37°C, stained with 5 μg mL⁻¹ WGA-Alexa Fluor 555 (Invitrogen, Carlsbad, CA) for 5 min at 37°C, and washed with PBS 3 times (5 min each). Cells were then incubated with DAPI for 5 min at 37°C, followed by washing with PBS. After PBS washing, cells were mounted with VECTASHIELD mounting medium (Vector Laboratories, Inc. Burlingame, CA). The images of cells were acquired using a Leica SP8 confocal laser scanning microscope (Leica, Germany).

4.6.2. Cellular uptake assay by flow cytometry.

4T1 cells were incubated with CHI-PEG-PTX nanofibers for 1 hr (50 μg mL⁻¹) at either 37°C or 4°C, followed by washing with cold PBS for 3 times. Cells were then trypsinized and resuspended in cold PBS and analyzed by flow cytometry (FACSCanto II, BD Biosciences).

4.6.3. Cell apoptosis assay by flow cytometry.

4T1 cells were seeded into 6-well plates and incubated overnight. CHI-PEG-PTX nanofiber stock solutions were added into cell cultures at a final PTX concentration of 1 μM. Cells were incubated with the CHI-PEG-PTX nanofibers for 48 h. Cells were then trypsinized, aspirated, and washed with PBS. Cells were counted and suspended in 0.1 mL Annexin V binding buffer containing 50 μg/mL propidium iodide and 5 μL FITCAnnexin V reagent. Cells were further incubated for 15 min at room temperature in dark. 0.4 mL Annexin V binding buffer was then added prior to analysis by flow cytometry. Data were acquired by a FACSCanto II system and analyzed by the FlowJo software (Treestar, Inc., San Carlos, CA).

4.7. Animal study:

All animal studies were conducted in accordance with University of Washington Institute of Animal Care and Use Committee (IACUC) approved protocols as well as with federal guidelines.

4.7.1. Blood half-life of CHI-PEG-PTX-Cy5 in mice.

Five-week-old female BALB/c mice were purchased from The Jackson Laboratory (Bar Harbor, Maine) and housed in the animal research facility at UW. One week later, mice were treated with 500 μL of 2 mg mL⁻¹ CHI-PEG-PTX-Cy5 in PBS (n = 3) by i.v. injection. At 1, 3, 6, 18, 24 and 48 h post-injection, blood (100 μL per mouse) was collected through retro-orbital bleeding or terminal cardiac puncture. The total amount of blood drawn from each mouse never exceeded one percent of the body weight of the animal in this experiment. Whole blood was spun for 2 min at 5000 g using a benchtop centrifuge to separate the plasma. Forty μL of plasma was diluted into 100 μL PBS and placed in a 96-well black plate. The plate was scanned on a SpectraMax i3 plate reader (fluorescence mode) (ex = 646 nm, em = 676 nm) to measure Cy5 fluorescence signal intensities.

4.7.2. Biodistribution of CHI-PEG-PTX-Cy5 in mice.

The same mice used for the blood half-life study were used in this study. Mice were euthanized 6 or 48 h after CHI-PEG-PTX-Cy5 injections. Organs (heart, lungs, liver, spleen, kidneys, muscle and brain) were collected and fluorescence intensities from these organs were measured by a XENOGEN IVIS 200 imaging system (PerkinElmer Inc.). Imaging parameters were set to: excitation wavelength: 675 nm; emission filter: 720–750 nm; exposure time: 1 second; binning factor: 2; f/stop: 4.

4.7.3. In vivo toxicity study.

Wild-type BALB/c mice were treated with PTX-crem or CHI-PEGPTX solution via i.v. injection at a PTX-equivalent dose of 10 mg kg⁻¹. Mice injected with saline were used as controls. Twenty-four hours post-injection, blood was collected by retro-orbital bleeding. Blood samples were delivered to Phoenix Central Laboratory for hematology analysis.

4.7.4. Tumor growth inhibition in flank xenograft mouse model of breast cancer.

The 4T1-luc cells were transfected to stably express luciferase so that bioluminescence could be used to monitor tumor growth.⁵³ 4T1-luc cells were trypsinized, suspended in PBS (10⁶ cells/mL), and injected subcutaneously into the right flanks of 6-week-old female BALB/c wide-type mice (0.1 mL per mouse). Four days after tumor inoculation, the mice were administered with PTX or CHIPEG-PTX via i.v. injection every 3 days for a total of 5 times. For each injection, the equivalent does of PTX was 10 mg kg⁻¹ per mouse. Other groups of mice were treated with various control media (saline, CHI-PEG, PTX-crem, and CHI-PEG-PTX LNPs). The images of mice with 4T1-luc cells was taken first right before the first treatment and then every 3 days for 21 days, a time frame similar to those of published studies using 4T1 model.^67–69 For bioluminescence imaging of tumors, each mouse was injected with 150 mg kg^–1 luciferin intravascularly and imaged with a XENOGEN IVIS 200 system (PerkinElmer, Inc.). Imaging parameters were set to emission filter: open; exposure time: 30 seconds; binning factor: 2; f/stop: 4.

4.7.5. NP accumulation in tumor and metastasis inhibition in an orthotopic mouse model of breast cancer.

4T1-luc cells (10⁶ cells per mouse) were injected subcutaneously into the #9 mammary gland to create orthotopic metastatic breast tumors in 6-week-old female BALB/c mice. For assessment of NP accumulation in tumors, each mouse received one intraperitoneal (ip) injection of one of the four specified agents (see Fig. 6 caption) 10 days after tumor inoculation. Mice were euthanized 24 h post-injection and tumors were collected and measured by an IVIS imager (excitation filter: 745 nm; emission filter: indocyanine green (ICG)). For assessment of metastasis inhibition, one week after tumor inoculation, each mouse received successive four ip injections (3 days apart) of one of the four specified agents. The specified agents include i: saline; ii: Cy5.5-labeled PTX-crem; iii: Cy5.5-labeled CHI-PEG-PTX LNPs; iv: Cy5.5-labeled CHI-PEG-PTX NFs. For all the agents except saline, the equivalent PTX dose was 10 mg/kg per injection. Mice were injected with 150 mg kg^–1 luciferin intraperitoneally and imaged by an IVIS 200 imaging system.

4.7.6. Tumor growth inhibition in a flank xenograft mouse model of melanoma.

6-week-old female C57BL/6 mice were used in this study. B16-F10 cells were trypsinized and suspended in PBS (10⁶ cells/mL) and injected subcutaneously into the right flank of each mouse (80 μL per mouse). One week after tumor inoculation, mice were administered with either PTX-crem, CHI-PEG-PTX LNPs or CHI-PEG-PTX NFs by intravascular injection in every 3 days for a total of 4 times. The equivalent PTX dose for all three agents was 10 mg kg⁻¹ per mouse per injection. Mice treated with saline served as controls.

4.7.7. H & E staining and imaging for histopathological evaluation.

The deparaffinized samples were stained with haemotoxylin and eosin, and mounted with Prolong Gold mounting medium. Microscopic images of tissues were acquired using a Nikon ECLIPSE TE 2000-S microscope.

4.8. Statistical analysis.

Student’s unpaired t-tests were performed for comparisons of individual groups. One-way and analyses of variance (ANOVA) followed by Turkey’s post-hoc multiple comparison tests were used for comparisons of multiple groups (calculated by GraphPad Prism software).