Amine Functionalized Trimetallic Nitride Endohedral Fullerenes: A Class of Nanoparticle to Tackle Low Back/Leg Pain

Abstract

Low back pain is the most common health problem with a prevalence of over 80% worldwide and an estimated annual cost of $100 billion in the United States. Intervertebral disc degeneration accounts for a major cause of low back pain. However, there is still a lack of safe and effective treatment to tackle this devastating condition. In this study, we synthesized four functionalized trimetallic nitride endohedral metallofullerenes (carboxyl-f-Sc₃N@C₈₀, carboxyl-f-Gd₃N@C₈₀, amino-f-Sc₃N@C₈₀, and amino-f-Gd₃N@C₈₀) and characterized them with X-ray photoelectron spectroscopy, matrix-assisted laser desorption/ionization-time of flight mass spectrometry, and UV–vis. Via electron paramagnetic resonance, all four metallofullerene derivatives possessed dose-dependent radical scavenging capabilities (hydroxyl radicals and superoxide anions), with the most promising radical scavenging properties shown in the amine functionalized C₈₀ metallofullerenes. Both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ at 1 μM significantly reduced lipopolysaccharide induced reactive oxygen species production and mRNA expressions of pro-inflammatory mediators (inos, tnf-α, il-1, and cox-2) in macrophages without apparent cytotoxicity through regulating activity of p38 MAPK, p65, and nuclear translocation of NF-κB. Furthermore, in an established mouse model of lumbar radiculopathy, amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ effectively alleviated ipsilateral mechanical hyperalgesia for up to 2 weeks. In dorsal root ganglia explant culture, we also showed that amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ ameliorated TNF-α elicited neuroinflammation. In summary, we presented results for a potent radical scavenging, anti-inflammatory and analgesic nanoparticle, amino-functionalized eighty-carbon metallofullerenes in vitro and in vivo. Our study provides important assets for developing pleiotropic treatment strategies to tackle the inflammation, a significant pathological hallmark in the intervertebral disc degeneration and associated pain.

Graphical Abstract

INTRODUCTION

Low back pain is a common reason for lost work days,^1,2 with a prevalence of over 80%^3,4 and an estimated annual cost of more than $100 billion in the U.S.⁵ Intervertebral disc degeneration accounts for a major cause of low back pain.^6,7 Currently, there are no disease-modifying therapies to manage disc degeneration and associated pain given the efficacy and safety of existing anti-inflammatory treatments for discogenic pain, such as non-steroidal anti-inflammatory drug and epidural spinal injection.⁸ Therefore, new therapeutic strategies are urgently desired to tackle this clinical condition.

Inflammation and oxidative stress have been increasingly recognized as hallmarks for development of disc degeneration and associated pain.^9,10 Various inflammatory mediators, such as cytokines and chemokines, are elevated in patient degenerative disc tissues.¹¹ In the prior research, others and us have revealed an abundance of macrophage infiltration in both herniated disc tissue of mouse¹² and human,^13,14 which may promote inflammation through release of pro-inflammatory cytokines and exacerbate the progression of disc diseases.¹⁵ Oxidative stress is observed in the course of many diseases since the presence of excessive reactive oxygen species (ROS) can cause cellular oxidative stress, which may lead to subcellular damage such as DNA, RNA, protein, mitochondrial, and membrane or other lipid degradation.¹⁶ Furthermore, high concentrations of ROS can promote the expression of inflammatory cytokines in the degenerated tissue.¹⁷ Accumulating studies have reported the existence of oxidative stress in degenerated discs that promoted catabolic metabolisms. Nerlich et al. suggested that “brown degeneration” of discs possibly resulted from non-enzymatic glycoxidative reactions of lysyl residues from long-living matrix proteins.¹⁸ Later studies confirmed that excessive ROS played an important role in the progression of intervertebral disc disease.^19-21 Thus, targeting inflammation and oxidative stress may be an effective strategy to retard disc degenerative changes.

Fullerenes are polyhedral closed carbon cages made exclusively of carbon atoms.²² They exhibit various activities given their unique structures and properties.²³ Fullerenes are powerful antioxidants due to their delocalization of the π-electrons over the carbon cage, which can readily react with free radicals. Endohedral metallofullerenes and trimetallic nitride endohedral metallofullerenes were discovered in the 1990s and soon recognized for their multifunctional capabilities in biomedical applications.²⁴ The large surface area of fullerenes and metallofullerenes allows for a variety of functional groups to be attached to the surface and provides the potential for a variety of modalities including magnetic resonance imaging (MRI) and antioxidative properties. For instance, as potential MRI contrast agents, functionalized gadolinium encapsulated metallofullerenes, Gd@C₈₂ and Gd₃N@C₈₀, attracted much attention since they exhibit higher ¹H relativity than most commercial contrast agents.^25-27 In addition, with a core of Gd atom, nanoparticle Gd@C₈₂(OH)₂₂ and derivatives have been reported with excellent anticancer activities and potential immunomodulatory effects.²⁸ Moreover, the carbon cage of Sc₃N@C₈₀ can be iodinated with ¹²⁴I to form a single modality positron emission tomography (PET) or a dual MRI and PET imaging agent for Gd₃N@C₈₀.²⁹ We also developed an interleukin-13 peptide conjugated trimetallic nitride endohedral fullerene Gd₃N@C₈₀ derivative that effectively targeted glioblastoma multiforme (GBM) cells given its ability as a great MRI contrast agent.^30,31 So far, the therapeutic potentials of Gd₃N@C₈₀ nanoplatform have not been fully explored in inflammatory disorders, especially in the domain of intervertebral disc degeneration and associated pain.

Over the past years, we have focused on identifying optimal fullerene and metallofullerene derivatives for treating disc degeneration induced back pain.^32-34 This is because fullerene derivatives possess unique properties such as potent ROS scavenging compared to other antioxidants via enhanced absorption, bioavailability, and readily functionalization for targeted delivery. We discovered that both local administration of water-soluble fullerol (C₆₀–OH)^33,35 and systemic delivery of a targeting peptide conjugated C₆₀ (FT-C₆₀)³² could effectively quench the pro-inflammatory responses and oxidative stress in various in vitro and in vivo models associated with disc degeneration.^36,33,37 Recently, we further demonstrated that carboxylic acid functionalized carboxyl-Gd₃N@C₈₀ exhibited potent antioxidative and anti-inflammatory potential in macrophages, similar to their nonmetal or monometal counterparts.³⁴ Compared to C₆₀, the eighty-carbon cage of M₃N@C₈₀ (M = metal) offers a larger surface area for functionalization. The delocalization of higher numbers of π-electrons over a larger surface area, compared to C₆₀, makes M₃N@C₈₀ capable of reacting with more free radicals by accepting their unpaired electrons. Also, the encapsulated metal ions can bring additional properties to M₃N@C₈₀, for instance, the MRI imaging capability of gadolinium (Gd) ions. Given their radical scavenging properties and the additional properties provided by encapsulated ions, f-M₃N@C₈₀ are promising to become multimodality molecules.

Here, we hypothesize that surface modification alters the radical scavenging and anti-inflammatory effects of f-M₃N@C₈₀. We synthesized and characterized four functionalized trimetallic nitride endohedral metallofullerenes (carboxyl-f-Sc₃N@C₈₀, carboxyl-f-Gd₃N@C₈₀, amino-f-Sc₃N@C₈₀, and amino-f-Gd₃N@C₈₀), evaluated their radical scavenging capability, in vitro antioxidative and anti-inflammatory effects in macrophages, and investigated the analgesic effects in a mouse model of lumbar radiculopathy. This study innovatively reveals the relationship between structural modifications of trimetallic nitride endohedral metallofullerene and their performance in biological activities such as radical scavenging and anti-inflammation. This study will also guide future preclinical development of functionalized trimetallic nitride endohedral metallofullerenes in becoming a class of nanomedicine to tackle degenerative disc diseases and associated pain.

EXPERIMENTAL METHODS

Chemicals and Reagents.

Gd₃N@C₈₀ and Sc₃N@C₈₀ were purchased from LUNA Innovations (Danville, VA). Hydrogen peroxide (H₂O₂), ferrous sulfate heptahydrate (FeSO₄), xanthine, xanthine oxidase from bovine milk (XOD), diethylenetriamine-pentaacetic acid (DTPA), α-cyano-4-hydroxycinnamic acid (CHCA), 2,5-dihydroxybenzoic acid (DHB), and lipopolysaccharide (LPS) were purchased from Sigma-Aldrich. Spin traps, 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide (DEPMPO) was purchased from Focus Biomolecules (Plymouth Meeting, PA), and 5-tert-butoxycarbonyl-5-methyl-1-pyrroline-N-oxide (BMPO) was supplied by Cayman Chemical (Ann Arbor, MI). 1,2-Dichlorobenzene (ODCB) was purchased from Beantown Chemical (Hudson, NH). Sodium hydroxide, deionized water, and toluene were purchased from Fisher Scientific (Pittsburgh, PA). The clear EPR quartz capillary tubes (ID 1.0 mm, OD 1.2 mm) were obtained from Wilmad-LabGlass (Vineland, NJ). Gibco Dulbecco’s modified Eagle medium (high glucose 4.5 g/L) (DMEM), Ham’s F-12 Nutrient Mixture (F12), fetal bovine serum (FBS), penicillin/streptomycin (Pen/Strep), and Prolong Gold antifade reagent with 4′,6-diamidino-2-phenylindole (DAPI) were purchased from Life Technologies (Carlsbad, CA). Other chemicals were obtained from Sigma-Aldrich (St. Louis, MO) unless mentioned.

Preparation of Functionalized Metallofullerene Derivatives.

Synthesis of four metallofullerene derivatives was performed following previously reported procedures.^27,38 Succinic acid acyl peroxide was prepared first. Succinic anhydride (4 g) as fine powder was added to 10 mL of ice cold hydrogen peroxide (8%) and stirred for 30 min in an ice bath (Figure S1A). The resulting gel-like solution was filtered through a 0.45 μm-pore-size PTFE membrane using vacuum filtration and washed by a small amount of cold water. ¹H NMR (400 MHz, acetone-d₆): 2.72 ppm (m); ¹³C NMR (100 MHz, acetone-d₆): 24.70, 27.90, 168.39, 172.01 ppm. Gd₃N@C₈₀ (6 mg) and succinic acid acyl peroxide (5 mg, 5 equiv) were dissolved in 12 mL of o-dichlorobenzene (ODCB). The resulting solution was deaerated by flowing argon and heated at 84 °C for 120 h. Additional succinic acid acyl peroxide (5 equiv) was added every 12 h. Then NaOH (8 mL 0.2 M) was added to the resulting brown sludge to extract the water-soluble product (Figure S1B). Two layers were obtained. The top layer, containing carboxyl groups functionalized Gd₃N@C₈₀, was concentrated and purified via a Sephadex G-25 size-exclusion gel column. The Sc₃N@C₈₀ was functionalized in a similar way (Figure S1C). To prepare the amine group functionalized Gd₃N@C₈₀, Gd₃N@C₈₀ (2.5 mg) was added to 1 mL of H₂O₂ (20 wt %) and 0.4 mL of NH4OH (28 wt %) and stirred vigorously at 50 °C for 2 h. The black suspension turned to a yellow solution during this process (Figure S1D). Ethanol (10 mL) was used to precipitate the water-soluble product. The suspension was centrifuged at 10 000 rpm for 10 min, and the pellet formed at the bottom of the centrifuge tube was washed with an excess amount of ethanol and air-dried. The amine functionalized Sc₃N@C₈₀ was also obtained in a similar fashion (Figure S1E).

Characterization of Four Metallofullerene Derivatives.

CHCA or DHB matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS) was carried out to confirm the attachment of the functional groups on the metallofullerenes and to determine the molecular weight and formula of the four metallofullerene derivatives. The 0.1% formic acid was added to facilitate the ionization of the metallofullerene derivatives in MALDI-TOF MS. To obtain the binding energy of C_1s and N_1s electrons to assist in determining the elemental composition and the empirical formula, four synthesized metallofullerene derivatives were also characterized by X-ray photoelectron spectroscopy (XPS) (PHI VersaProbe III scanning XPS microscope, Spectra Research Corporation, Mississauga, ON). The molecular formulas of the four metallofullerene derivatives were tabulated by exhaustively listed all the possible molecular formulas that fit the mass spectra. The molecular formulas were corroborated by the XPS data. Meanwhile, the UV–vis absorption spectra of four metallofullerene derivatives were collected in deionized water using an Agilent Cary 60 UV–vis spectrophotometer (Agilent, Santa Clara, CA).

EPR Detection of Radical Scavenging Capability.

Following our previously reported protocols,³⁴ the EPR method was coupled with a spin-trapping agent DEPMPO. Hydroxyl radicals were generated by the classical Fenton reaction, which involved the reaction of FeSO₄ (5 mM) and H₂O₂ (25 mM). DEPMPO (2.5 mM) was used to bind short-lived ^•OH to form a more stable adduct and allow the detection of EPR signals. To this Fenton reaction solution, water and various concentrations of metallofullerene derivatives were added before the addition of DEPMPO. To test the superoxide radical anion (O₂^•−) scavenging activity, BMPO was used to trap and detect O₂^•− by EPR spectroscopy. The superoxide radical scavenger reaction was initiated by the addition of xanthine oxidase solution (XOD), containing 20 mM xanthine (1 M NaOH in PBS), 20 mM BMPO, 20 mM DTPA, and 4 U/mL XOD, in the presence or absence of various concentrations of metallofullerene derivatives. The EPR spectra were recorded at 2 min after the addition of the Fenton reagent and xanthine oxidase solution for the generation of ^•OH and O₂^•−, respectively. The EPR assays were carried out at ambient temperature using a Bruker ELEXSYS-II EPR spectrometer (Bruker, Billerica, MA). Clear quartz capillary tubes (ID 1.0 mm, OD 1.2 mm) were used as the sample container. The following instrument settings were used for collecting EPR spectra: microwave frequency of 9.88 GHz, microwave power of 20.02 mW, field modulation frequency of 100 kHz, and modulation amplitude of 2 G.

In Vitro Culture of Raw 264.7 Macrophages.

Macrophages Raw 264.7 cells were purchased from American Type Culture Collection (ATCC) (Manassas, VA). Cells were cultured with Dulbecco’s modified Eagle’s medium (DMEM, high glucose, Gibco, Grand Island, NY) supplied with 10% fetal bovine serum (Gibco), 1% penicillin (100 μg/mL), and 1% streptomycin (100 μg/mL) (Gibco) and maintained in 5% CO₂ at 37 °C.

MTS Cytotoxicity Assay.

Raw macrophages 264.7 cells were seeded into 96-well plate at a density of 10⁵/mL (0.1 mL cell suspension/well) overnight before incubating with complete growth media with or without 0.1, 1, and 5 μM amino-f-Gd₃N@C₈ or amino-f-Gd₃N@C₈₀ for 3 days. Images of cells were taken at a magnification of 200 and then subjected to MTS assay according to the manufacturing protocol of CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay kit (Promega, Madison, WI). Then OD₄₉₀ were obtained with SpectraMax ABS Plus microplate reader (Molecular Devices, San Jose, CA).

Detection of Intracellular Reactive Oxygen Species (ROS).

To evaluate protective effects of metallofullerenes derivatives against oxidative stress, intracellular staining of ROS was performed.^34,39 Raw 264.7 cells were seeded onto Nunc Lab-Tek Chamber Slide 8 well system (ThermoFisher Scientific, Waltham, MA) and incubated overnight until ~70% confluence. Cells were incubated with aqueous metallofullerenes solutions (amino-f-Sc₃N@C₈₀ or amino-f-Gd₃N@C₈₀) at 1 μM in serum-free media (DMEM with 1% penicillin and streptomycin) for 20 h and then stimulated with lipopolysaccharide (LPS) (100 ng/mL) for an additional 5 h at 37 °C. The dose and time of treatments were selected based on our prior studies. Cells were then incubated with 5 μM 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) (Life Technologies, Carlsbad, CA) in serum-free DMEM for 30 min at 37 °C in dark, rinsed with warm phosphate buffered saline (PBS, Invitrogen, Waltham, MA) (150 μL × 2), and mounted with Prolong Gold antifade reagent with DAPI. Fluorescence images were captured with a fluorescence microscope (LSM 510-UV, Carl Zeiss, Germany) and processed using NIS Element (Nikon Instruments, Melville, NY) with identical settings. ROS signal was capture in FITC green channel, and DAPI signal was visualized in blue channel. For each condition, triplicated experiments were performed. For each chamber, at least three regions of interest (ROIs) were analyzed. The ROS level was estimated based on the green fluorescence intensity in each condition.

Cell Treatment Protocols for Real-Time RT-PCR, Griess Assay, and Western Blotting.

In a 24-well plate, raw 264.7 cells were preseeded in complete growth media at a density of 2 × 10⁵ per mL (0.5 mL per well) and were incubated with either amino-f-Sc₃N@C₈₀ or amino-f-Gd₃N@C₈₀ at 1 μM overnight and then stimulated with or without LPS (100 ng/mL) for various time points including LPS treatment for 5 h for RNA isolation, 20 h for Griess assay, and 30 min for Western blots.

After LPS stimulation for 20 h, culture media were collected and subjected to Griess Assay following the manufacturer’s instruction (Promega, Madison, WI).³⁴

After LPS stimulation for 5 h, total RNA of Raw 264.7 cells were isolated with Trizol Reagent (Invitrogen) and quantified with a NanoDropTM 1000 spectrophotometer (ThermoFisher Scientific, Waltham, MA). One microgram of total RNA was used for cDNA synthesis with iScript cDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA) following the manufacture’s instruction. Real-time reverse transcription polymerase chain reaction (RT-PCR) was performed with RT^2 SYBR Green Fluo FAST Mastermix (Qiagen Sciences, Germantown, MA) on a QuantStudio3 Real-time PCR system (Applied Biosystems, Waltham, MA). Each qPCR reaction solution was prepared by 10 ng of cDNA, 5 μM of desired primer, and SYBRPremix Ex Taq (1×) in a total volume of 12.5 μL with 35 cycles. The mRNA expressions of il-6, il-1, cox-2, inos, and tnf-α were evaluated, respectively. The 18S rRNA were used as internal controls. Sequences of primers are listed in Table S1.

For Western blot, since phosphorylation of p38 and p65 typically happens within a short period after LPS stimulation, 30 min post LPS treatment Raw 264.7 cells were lysed in 0.1 mL of lysis buffer (RIPA buffer supplemented with 1× proteinase inhibitor cocktail and PMSF). Lysates were subjected to SDS-PAGE and transferred to a nitro-cellulose membrane. After incubation with Odyssey blocking buffer (LI-COR Biosciences, Lincoln, NE) for 1 h at room temperature, membranes were incubated with mouse p-p38 (1:1000, Cell Signaling Technology, Danvers, MA), mouse p38 (1:1000, Cell Signaling Technology), rabbit p-p65 (1:1000, Cell Signaling Technology), rabbit p65 (1:1000, Cell Signaling Technology), and mouse β-actin (1:5000, R&D System, Minneapolis, MN) antibodies overnight at 4 °C followed by incubation with goat antirabbit Alexa Fluo680 (1:3000) or goat antimouse Alexa Fluo680 (1:3000) (ThermoFisher Scientific, Waltham, MA) for 1 h at room temperature. Membranes were scanned using ChemiDoc Imaging System (Bio-Rad Laboratories, Hercules, CA), and signal intensity was quantified with ImageJ.

Animals.

The use of animals was approved by the Institutional Animal Care and Use Committee. C57BL/6 mice (8–12 weeks, 20–25 g, both male and female) (Envigo, Indianapolis, IN) were socially housed at the centralized animal facility of the University of Virginia. The rooms were fitted with a 12-h light/dark cycle and temperature of 25 °C. Food and water were provided ad libitum.

DRG Explant Culture for Real-Time RT-PCR and ELISA.

Mice were euthanized in a CO₂ chamber followed by cervical dislocation. The bilateral DRGs were immediately collected from the spinal column following our previously published protocols.³³ Dissected DRGs were cultured in growth medium containing F-12 medium, 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 10 ng/mL nerve growth factors (NGF) (BD Biosciences, San Jose, CA) in a 48-well plate with 8–10 DRGs per well. After DRG explant culture for 2 days, fresh growth media composed of F-12 medium, 100 U/mL penicillin, and 100 μg/mL streptomycin was supplemented with tumor necrosis factor alpha (TNF-α) (25 ng/mL, Cell Signaling Technology) with either amino-f-Sc₃N@C₈₀ or amino-f-Gd₃N@C₈₀ at 1 μM for another 24 h in serum-free medium. DRGs were homogenized in Trizol and processed for real-time RT-PCR as described above. Meanwhile, the supernatant of DRG explant culture was collected to analyze the IL-6 by a ELISA kit (eBioscience, Inc., San Diego, CA) according to the manufacturer’s instructions.

Immunofluorescence Staining of p65 Nuclear Translocation.

Raw 264.7 cells preseeded onto a 8-well Nunc Lab-Tek Chamber Slide system and incubated overnight until ~70% confluence were incubated with either amino-f-Sc₃N@C₈₀ or amino-f-Gd₃N@C₈₀ at 1 μM overnight and then stimulated with or without LPS (100 ng/mL) for 30 min. Cells were gently washed with warm PBS and fixed in 4% paraformaldehyde (PFA) for 10 min at room temperature, washed with PBS and then permeabilized with 0.3% Triton X-100 in PBS for 10 min, blocked with 3% bovine serum albumin (BSA), and incubated with rabbit p65 (1:500, Cell Signaling Technology) at 4 °C overnight. Then cells were incubated with goat antirabbit AlexaFluo594 (1:1000, Life Technologies, Carlsbad, CA) at room temperature for 1 h and mounted with Prolong Gold antifade reagent (with DAPI). IgG control staining was performed with identical procedures excluding the primary antibodies. Fluorescence images were taken with a Nikon ECLIPSE E600 microscope with Zeiss software (Nikon Instruments Inc., Melville, NY) as we established.¹² At a magnification of 200×, all cells were imaged with identical settings. Total cells (DAPI) and p65 nuclear location were counted from at least four images per chamber per condition in triplicated experiments with ImageJ. The number of total cells was enumerated by a customized ImageJ Macro (see Supporting Information).

Animal Surgical Procedure.

Following our established animal protocol, nucleus pulposus (NP) implant surgery was performed using aseptic technique and a surgical microscope.³⁵ General anesthesia was induced with intraperitoneal injection of ketamine/xylazine (60–80/5–10 mg/kg). Animals (male) were randomly divided into three groups: NP implant + saline, NP implant + amino-f-Gd₃N@C₈₀, and NP implant + amino-f-Sc₃N@C₈₀ groups (n = 6/group). In brief, the left L4/5 interlaminar space was exposed, followed by dissection of the coccygeal disc tissue (mainly NP with partial inner annulus fibrosus) from the same animal. Immediately after dissection from mouse tail and prior to implantation, coccygeal disc tissues were bathed in saline (0.9% NaCl), 1 μM amino-f-Gd₃N@C₈₀, and amino-f-Sc₃N@C₈₀ for 30 s, respectively, and then implanted over the exposed interlaminar space. Both amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ were diluted in saline (0.9% NaCl) for this experiment.

Electronic von Frey Test.

To assess the mechanical sensitivity of animals subjected to NP implantation, the electronic von Frey test was performed on both hind paws for three consecutive days prior to surgery and every other day until post operational day (POD) 12. Mice were acclimated on an elevated mesh grid, and mechanical sensitivity was determined using an electronic von Frey apparatus. Five trials were conducted on each paw, with at least 5 min rest time between trials. For each paw, mechanical thresholds from five trials were averaged after the maximum and minimum readings were excluded. Mean ipsilateral and contralateral mechanical thresholds were averaged at each time point.

Safranin-O Staining of Mouse Spine Sections.

The spines (n = 6 mice per group) were harvested at 12 days after surgery and fixed in 10% formalin, decalcified in 0.25 M ethylenediaminetetraacetic acid (EDTA) for 5 days, embedded in paraffin, and sectioned trans-axially. Axial sections (5 μm thickness) were then stained with Safranin-O and Fast Green to visualize implanted disc materials and inflammatory response.^40-42

Statistical Analysis.

Statistical analysis was performed using Prism software (GraphPad Software, La Jolla, CA). In vitro experiments were carried out in triplicate and independently repeated three times. Quantitative data were presented as mean ± SEM. Data comparing two groups were analyzed with a student’s t-test. Data comparing three or more groups were analyzed with one-way ANOVA followed by a Tukey multiple comparison test. For behavior assay, statistical significance was determined using two-way ANOVA followed by the Bonferroni’s multiple comparisons test. A p-value of less than 0.05 was considered statistically significant.

RESULTS AND DISCUSSION

Design, Synthesis, and Characterization of Metallofullerene Derivatives.

Chemical characterization of four synthesized metallofullerene derivatives, carboxyl-f-Sc₃N@C₈₀, carboxyl-f-Gd₃N@C₈₀, amino-f-Sc₃N@C₈₀, and amino-f-Gd₃N@C₈₀ (Figure 1), were performed with MALDI-TOF MS, XPS, and UV–vis spectroscopies.

Figure 1.

Chemical structures and molecular formulas of carboxyl-f-Sc₃N@C₈₀, carboxyl-f-Gd₃N@C₈₀, amino-f-Sc₃N@C₈₀, and amino-f-Gd₃N@C₈₀. The molecular formulas were calculated based on the MS data, shown as Sc₃N@C₈₀(OH)₂₉(CH₂CH₂COOH)₁₆ (top left), Gd₃N@C₈₀(OH)₁₉(CH₂CH₂COOH)₁₃ (top right), Sc₃N@C₈₀O₂₀(OH)₈(NH₂)₁₂(NO₂)₁ (bottom left), and Gd₃N@C₈₀O₂₀(OH)₄(NH₂)₁₂ (bottom right), respectively.

The MALDI-TOF MS spectra of the four metallofullerene derivatives were shown in Figure S2. The peaks at 2771 m/z, 2719 m/z, 1804 m/z, and 2027 m/z represent the molecular peaks of carboxyl-f-Sc₃N@C₈₀, carboxyl-f-Gd₃N@C₈₀, amino-f-Sc₃N@C₈₀, and amino-f-Gd₃N@C₈₀, respectively. The molecular formulas of the four derivatives were determined by the method of exhaustion based on the molecular peaks. The molecular formulas were calculated as Sc₃N@C₈₀(OH)_29⁻, (CH²CH₂COOH)₁₆, Gd₃N@C₈₀(OH)₁₉(CH₂CH₂COOH)₁₃, Sc₃N@C₈₀O₂₀(OH)₈(NH₂)₁₂(NO₂)₁, and Gd₃N@C₈₀O_20⁻(OH)₄(NH₂)₁₂. It is well recognized that Gd has several isotopes. The presence of three Gd atoms in each Gd₃N@C₈₀ significantly broadened, complicated, and made it difficult to acquire the mass spectra of these derivatives. Therefore, the mass spectra obtained for the Gd₃N@C₈₀ derivatives are significant.

As shown in Figure S3, the C_1s XPS multiplex spectrum of the four metallofullerene derivatives and the N_1s spectrum of the amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ were presented. The C_1s peaks, centered at binding energy values of 284.8, 286.6, and 288.8 eV, were assigned to C–C, C–O (or C–N for amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀), and C=O, respectively, indicating the introduction of carboxyl groups and hydroxyl groups (or amine groups) to the carbon cage. The N_1s peaks, centered at binding energy values of 399.69, 401.62, and 407.52 eV, were assigned to −NH₂, protonated −NH₂, and NO₂ groups, indicating the introduction of amine groups and nitro groups to the carbon cage. In XPS, the area of a peak is proportional to the amount of the functional group that is presented in the sample. Therefore, the ratio of the peak areas provides information about the ratio of different functional groups. In the C_1s XPS spectrum of carboxyl-f-Sc₃N@C₈₀ and carboxyl-f-Gd₃N@C₈₀, the ratios of C–O to C=O were 1.8:1 and 1.5:1, respectively, similar to the ratios of 29:16 and 19:13, supporting the molecular formulas of Sc₃N@C₈₀(OH)_29⁻(CH₂CH₂COOH)₁₆ and Gd₃N@C₈₀(OH)_19⁻ (CH₂CH₂COOH)₁₃. In the C_1s XPS of the amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀, C=O to C–O/C–N ratios were 1:1 and 1.2:1, supporting the ratios of oxygen to −OH/−NH₂ groups calculated from the MS data for the amino metallofullerene derivatives. The N_1s XPS spectra indicated the −NO₂ to −NH₂/−NH₃⁺ ratio in the amino-f-Sc₃N@C₈₀ was 1:12 and the amino-f-Gd₃N@C₈₀ had no trace of −NO₂ groups, in accordance with the molecular formula of Sc₃N@C₈₀O₂₀(OH)_8⁻ (NH₂)₁₂(NO₂)₁, and Gd₃N@C₈₀O₂₀(OH)₄(NH₂)₁₂ calculated from the MS data.

The UV–vis absorbance spectra (Figure S4) revealed the changes in the conjugate system, indicated by the disappearance of the characteristic peaks of Gd₃N@C₈₀ at 407 and 555 nm, and the characteristic peaks of Sc₃N@C₈₀ at 377, 424, and 622 nm. This loss of characteristic absorbance of Gd₃N@C₈₀ and Sc₃N@C₈₀ indicated the decreased conjugation features by the removal of p-orbitals from the π system of the cage. The color of the Gd₃N@C₈₀ and Sc₃N@C₈₀ solution was dark brown. The color of the amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ solution was light yellow. The carboxyl-f-Sc₃N@C₈₀ and carboxyl-f-Gd₃N@C₈₀ solutions were light brown. The lighter color of the derivatives also indicated the decrease in the absorption of visible light by the π system of the cage.

Amino-f-Sc₃N@C₈₀ and Amino-f-Gd₃N@C₈₀ Demonstrated Strong Radical Scavenging Capability.

Hydroxyl (^•OH) and superoxide anion (O₂^•−) radicals are among the most common ROS in the biological system by inducing oxidative stress and mediating inflammation. We adopted EPR to assess the radical scavenging properties of hydroxyl (^•OH) and superoxide anion (O₂^•−) radicals in the early phase of our evaluation to determine potential antioxidative effects in the biological system. As carboxyl-f-Sc₃N@C₈₀, carboxyl-f-Gd₃N@C₈₀, amino-f-Sc₃N@C₈₀, and amino-f-Gd₃N@C₈₀ derivatives were added, the peak intensity decreased dramatically in a dose-dependent manner and showed a varied degree of ^•OH radical scavenging (Figure 2). Similarly, all four functionalized metallofullerenes effectively scavenged the O₂^•− radical in a dose-dependent manner (Figure 3). As shown in Figures 2E,F and 3E,F, the percentages of signal illustrated the dose-dependent radical scavenging capabilities in all four functionalized metallofullerenes with the highest scavenging capabilities in amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ for both ^•OH and O₂^•−. Specifically, amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ at 1.42 μM showed ~59% and ~49% signal reduction for ^•OH, and ~71% and ~58% signal reduction for O₂^•−, respectively. This result indicated that both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ possessed relatively stronger radical scavenging capabilities compared to the carboxyl group functionalized ones regardless of encapsulated metal and therefore were used in the following biological tests.

Figure 2.

Hydroxyl radical scavenging properties of four functionalized metallofullerenes. EPR spectra of hydroxyl radicals captured by DEPMPO with and without (A) carboxyl-f-Sc₃N@C₈₀, (B) carboxyl-f-Gd₃N@C₈₀, (C) amino-f-Sc₃N@C₈₀, and (D) amino-f-Gd₃N@C₈₀. (E) Table summary and (F) plot on percentage of signal exhibited dose-dependent scavenging capabilities for hydroxyl radical in all four functionalized metallofullerenes. Amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ had similar hydroxyl radical scavenging profiles. Ultrapure water was used as a control.

Figure 3.

Superoxide radical scavenging properties of four functionalized metallofullerenes. EPR spectra of superoxide radicals captured by BMPO with and without (A) carboxyl-f-Sc₃N@C₈₀, (B) carboxyl-f-Gd₃N@C₈₀, (C) amino-f-Sc₃N@C₈₀, and (D) amino-f-Gd₃N@C₈₀ derivatives. (E) Table summary and (F) plot on percentage of signal exhibited dose-dependent scavenging capabilities of superoxide radical in all four functionalized C₈₀ metallofullerenes with the highest scavenging capabilities in amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀. Ultrapure water was used as a control.

The number and the accessibility of the π bonds and functional groups both impact the radical scavenging capabilities of the metallofullerene derivatives. Modification of carbon cages has been highly associated with alteration of their radical scavenging capabilities that correspondingly affect their performance in biological systems. These adaptations occur because chemical modification, in which a cage carbon is replaced with another unit, plays a significant role in altering the π-system of the fullerene framework. In our previously published work, we showed that two hydrochalarones, hydroxylated and PEGylated Gd₃N@C₈₀ molecules (HyC-1-Gd₃N@C₈₀, and HyC-3-Gd₃N@C₈₀, originally developed as enhanced MRI contrast agents⁴³), possessed much lower ^•OH radical scavenging capability than carboxyl functionalized Gd₃N@C₈₀(OH)₃₀(CH₂CH₂COOH)₂₀ (also referred to as “carboxyl-Gd₃N@C₈₀”). It is of note that both carboxyl and amine functionalized nanoplatform are highly compatible for chemical conjugation with biomedical significant proteins, peptides, and molecules. The EPR results here for both ^•OH and O₂^•− exhibited dose-dependent scavenging capabilities with better performance in amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ compared to carboxyl-f-Gd₃N@C₈₀ and carboxyl-f-Sc₃N@C₈₀. According to the structures solved from the MS and XPS data, for both amine-f-metallofullerene and carboxyl-f-metallofullerenes, there are fewer functional groups attached to the Gd₃N@C₈₀ derivatives than Sc₃N@C₈₀ derivatives, possibly making the conjugation system on the Gd₃N@C₈₀ cages more intact than the Sc₃N@C₈₀ cages. Therefore, regardless of amine- or carboxyl-functional group, the higher radical scavenging ability of the metallofullerene derivatives might be partially contributed by a greater number of double bonds remaining on the cages of these metallofullerene derivatives. Also, amino-f-metallofullerenes generally showed higher radical scavenging capabilities compared to the carboxyl-f-metallofullerenes, and this might be related to the basic nitrogen atoms in the amine groups. In addition, the exposure of the functional groups and the carbon cage may also play important roles in determining the radical scavenging capability. The possible aggregation of the derivatives shall also be considered, since aggregation could occur even as they appear soluble in water, as demonstrated by prior studies.^44-46 We can speculate that the mechanism of the varied radical scavenging of difference derivatives may be multifactorial and worth further investigation in future studies.

Amino-f-Sc₃N@C₈₀ and Amino-f-Gd₃N@C₈₀ Did Not Possess Cytotoxicity in Macrophages.

To assess the impacts of amine functionalized metallofullerene on cell viability, MTS assay was performed in Raw 264.7 macrophages for up to 3 days. Both amino-f-Gd₃N@C₈₀ and amino-f-Gd₃N@C₈₀ did not alter cell morphology at 1 μM compared to control cells after culture for 3 days (Figure 4A). Accordingly, they did not show any toxic effect to cells at all doses tested (0.1, 1, and 5 μM) after 3 days (Figure 4B,C). Similar to other nanoparticles, physicochemical properties of the fullerenes and their derivatives are likely to influence their toxicity, which may be associated with difference in chemical structure, surface modifications, preparation procedures, and dosages. Increased variation in the number of carboxyl, amine, and hydroxyl groups may be incorporated by intentionally varying reaction time at corresponding steps during chemical synthesis. According to literature, water-soluble derivatives are eliminated from the exposed animals within weeks with generally low acute oral, dermal, and airway toxicity. In general, the acute oral, dermal, and airway toxicity is low.⁴⁷ For example, fullerol 10 mg/kg was well tolerated in rats after intravenous administration for up to 2 days after dosing.⁴⁸ On the basis of the cytotoxicity study and EPR results, in vitro doses of 1 μM were selected for the following evaluations.

Figure 4.

Amine functionalized metallofullerenes did not show any toxicity in macrophages in vitro for up to 3 days. Over a 3-day culture, Raw 264.7 macrophages were cultured in complete growth media and treated with amino-f-Gd₃N@C₈₀ and amino-f-Gd₃N@C₈₀ at 0.1, 1, and 5 μM. MTS assay of (A) amino-f-Gd₃N@C₈₀ and (B) amino-f-Gd₃N@C₈₀ suggested that neither metallofullerene affected cell proliferation nor metabolic activities compared to control groups (without C₈₀ metallofullerenes). Experiments were performed in triplicates. (C) Representative microscopic images depicted similar cell morphology in control, amino-f-Gd₃N@C₈₀ (1 μM), and amino-f-Gd₃N@C₈₀ (1 μM) groups after 3-day culture. Images were taken at 200× magnification. Scale bar = 100 μm.

Amino-f-Sc₃N@C₈₀ and Amino-f-Gd₃N@C₈₀ Reduced LPS Elicited ROS and inos in Macrophages.

To confirm our cell-free radical scavenging results, an in vitro macrophage cell model was adopted for biological characterization of antioxidative and anti-inflammatory effects. Upon LPS treatment, elevated ROS (green fluorescence) were observed by intracellular staining with H₂DCFDA, by ~9-fold compared to no treatment control (*p < 0.05 vs ctrl), while both amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ at 1 μM showed a significant inhibitory effect on ROS production by more than ~60% compared to LPS stimulated group (^#p < 0.05 vs LPS) (Figure 5A,B). Activation of inducible nitrite oxide synthase (inos) is a hallmark of proinflammatory phenotype of macrophages. Both amino-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ at 1 μM reduced LPS induced inos mRNA by more than 75% (*p < 0.05 vs control, ^#p < 0.05 vs LPS) (Figure 5C). Both compounds did not activate pro-inflammatory phenotype of macrophages (Figure 5D). Similarly, the production of nitrite (oxidized product of nitric oxide) declined in both amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ groups compared to the LPS group, although it did not reach statistical significance (Figure S5). The prominent antioxidative effects of both amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ in macrophages correlate with their desirable radical scavenging capabilities in cell-free systems and suggested that both amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ possessed strong antioxidative effects.

Figure 5.

Amine functionalized metallofullerenes reduced LPS elicited reactive ROS and mRNA expression of inos in macrophages. Raw 264.7 cells were treated with amino-f-Gd₃N@C₈₀ and amino-f-Gd₃N@C₈₀, respectively, at 1 μM in serum-free media for 20 h, treated with LPS for 5 h, followed by intracellular ROS staining with 5 μM 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) and counter-stained by DAPI. Fluorescence imaging was performed at DAPI and FITC channels. Control cells were treated with serum-free media. (A) Intracellular ROS production (green fluorescence) was significantly increased in LPS group, which was inhibited by pretreatment of both amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀. Images were taken at 200× magnification. Scale bar = 100 μm. (B) Quantification of ROS fluorescence signal showed LPS increased ROS production by ~9-fold compared to control with more than ~60% inhibitory effects in amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ pretreated groups compared to LPS group. (C) Both amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ at 1 μM reduced the LPS increased mRNA expression of inos (*p < 0.05 vs control ^#p < 0.05 vs LPS). (D) Neither amine functionalized C₈₀ metallofullerene induced inos mRNA expression of macrophages at 1 μM for 24 h. Cells were seeded and treated in Nunc Lab-Tek Chamber Slide 8-well system in triplicates. The experiments were performed independently three times.

Amino-f-Sc₃N@C₈₀ and Amino-f-Gd₃N@C₈₀ Decreased LPS Induced Pro-inflammatory Genes in Macrophages Possibly through p38 MAPK/NF-κB Pathway.

It was reported that high concentration of ROS production may favor induction of M1-like pro-inflammatory macrophages during onset and disease progression,⁴⁹ in which pro-inflammatory mediators such as IL-1, IL-6, TNF-α, and prostaglandin E2 (PGE2) are elevated, thus triggering progressive activation of immune cells and neuropathic pain in degenerative disc diseases.¹⁵ Our RT-PCR data suggested that at an in vitro dose of both 1 μM amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ significantly reduced LPS induced elevation of pro-inflammatory genes including tnf-α (Figure 6A), il-1 (Figure 6B), and cox-2 (Figure 6D). Amino-f-Sc₃N@C₈₀, but not amino-f-Gd₃N@C₈₀, reduced il-6 mRNA expression (Figure 6C), which is possibly because these two amine functionalized nanoparticles attenuate inflammation via an alternative mechanism.

Figure 6.

Amine functionalized metallofullerenes decreased LPS induced pro-inflammatory genes in macrophages. Raw 264.7 cells were treated with nanoparticles at 1 μM in serum-free media for 20 h before LPS (100 ng/mL) stimulation for 5 h. Control cell was treated with serum-free media. Both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ significantly attenuated LPS induced mRNA expression of pro-inflammatory genes including (A) tnf-α, (B) il-1, and (D) cox-2. (C) Amino-f-Sc₃N@C₈₀ was effective in inhibiting LPS elicited mRNA expression of il-6, while amino-f-Gd₃N@C₈₀ did not show beneficial effect. *p < 0.05 vs control, ^#p < 0.05 vs LPS-treated groups. Cells were seeded and treated in 24-well plates. The experiments were performed independently three times.

MAPK and NF-κB are common inflammation regulators.^50,51 Oxidative stress can activate the MAPK-signaling pathway that regulates transcriptional activity of NF-κB subunit p65.^52,53 Several studies have shown that blocking p38 activity attenuates the transcriptional activity of the proinflammatory transcription factor NF-κB. After 30 min stimulation, LPS induced phosphorylation of p38 mitogen-activated protein kinase (MAPK) and NF-κB p65 subunit (*p < 0.05 vs control, Figure 7A-C), both of which were markedly inhibited by amino-f-Sc₃N@C₈₀ by ~86% and ~67% (^#p < 0.05 vs LPS), respectively. Amino-f-Gd₃N@C₈₀ did not change LPS elevated phosphor-p65 (*p < 0.05 vs control and n.s. vs LPS) but modestly reduced phosphor-p38 by ~30% (*p < 0.05 vs control and ^#p < 0.05 vs LPS). Phosphor-p38 peaked after 30 min of LPS treatment on our preliminary time-dependent test (Figure S6). In addition, immunostaining data showed that LPS induced p65 nuclear translocation (>95%) consistent with literature,⁵⁴ while both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ effectively prevented p65 nuclear translocations by ~60–70% (Figure 7D,E). These data suggested that amino-f-Sc₃N@C₈₀ may affect the upstream p38 MAPK pathway, while amino-f-Gd₃N@C₈₀ may mainly intervene on the transcription factor NF-κB. The detailed molecular mechanisms warrant further investigation.

Figure 7.

Amine functionalized metallofullerenes inhibited LPS activated p38 MAPK and NF-κB pathway in macrophages. Raw 264.7 cells were treated with nanoparticles at 1 μM in serum-free media for 20 h before LPS (100 ng/mL) stimulation for 30 min. Control cell was treated with serum-free media. (A) Western blots show that amino-f-Sc₃N@C₈₀ markedly inhibited LPS induced phosphorylation of p38 MAPK and p65 subunit, while amino-f-Gd₃N@C₈₀ only moderately reduced phosphorylation of p38. Quantification of (B) p-p38 and (C) p-p65 and normalized to pan protein level. (D) Immunostaining of p65 shows that both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ prevented LPS induced p65 nuclear translocation. (E) Quantification of p65 nuclear translocation in various treatment groups. In D, scale bar represents 50 μm. White arrow indicates p65 in cell cytosol, and yellow arrow shows p65 translocation into cell nuclei. *p < 0.05 vs control, ^#p < 0.05 vs LPS-treated groups. Cells were seeded and treated in 24-well plates. The experiments were performed independently three times.

Amino-f-Sc₃N@C₈₀ and Amino-f-Gd₃N@C₈₀ Effectively Alleviated Mechanical Hyperalgesia in Mouse Model of Radiculopathy and Neuroinflammation in DRG Explant Culture.

In additional to macrophage induced inflammatory responses, neuroinflammation in DRGs and nerve roots are also highly indicative in the pathology of disc related radicular pain. As we previously established, at early time points, for example, POD 3 and POD 6, excessive inflammatory cytokine IL-1 and IL-6 and macrophage were detected at the disc implantation site.³⁵ To test the analgesic effect of amine functionalized C₈₀ metallofullerenes, extracted mouse tail NP tissue was bathed in amino-f-Sc₃N@C₈₀ (1 μM), amino-f-Gd₃N@C₈₀ (1 μM), or saline (0.9% NaCl) before implantation. For the ipsilateral hind paw, the saline group demonstrated significantly increased mechanical sensitivity for up to POD 12 compared to the presurgery baseline, while both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ groups exhibited less mechanical sensitivity compared to saline group (*p < 0.05, saline vs amino-f-Gd₃N@C₈₀; #p < 0.05, saline vs amino-f-Sc₃N@C₈₀) with no significant difference compared to the baselines (n.s. vs baseline) throughout the observation period (Figure 8A). For contralateral mechanical threshold, no significant difference was observed among the three groups (Figure 8B). Histology confirmed abundant cell infiltration around implanted discs at the posterior spines in all groups (Figure S7). To explore the possible effects of amine functionalized metallofullerenes in neuroinflammation, we also performed in vitro mouse DRG explant culture. TNF-α treatment significantly elevated il-6 (Figure 8C) and cox-2 (Figure 8D) mRNA and IL-6 protein levels (Figure 8E) (*p < 0.05 vs control DRGs). Both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ treatment significantly decreased these inflammatory markers induced by TNF-α (~40–50% decline, #p < 0.05 vs TNF-α group). Discogenic pain (radicular pain induced by disc herniation or disc degeneration) involves a complex cascade of inflammatory and neuro-inflammatory events in the pathological microenvironment of disc, involving immune cells and dorsal roots. IL-6 plays an important role in both the inflammatory process and pain. Although amino-f-Gd₃N@C₈₀ did not attenuate the LPS induced IL-6 mRNA level in the macrophage cell model (Figure 6C), it inhibited LPS induced phosphorylation of p38 in macrophages (Figure 7B), which are upstream regulators of many inflammatory cytokines. The discrepancy may be caused by temporal activation of IL-6. Moreover, amino-f-Gd₃N@C₈₀ inhibited the expression of IL-6 at both mRNA and protein levels in the explant culture of dorsal root ganglia (Figure 8C,E) that are mainly responsible for pain sensation.

Figure 8.

Amine functionalized metallofullerenes effectively alleviated mechanical hyperalgesia in a mouse model of radiculopathy and neuroinflammation in DRG explant culture. Metallofullerenes were diluted in saline (0.9% NaCl) solution. Coccygeal disc tissues were bathed in amino-f-Sc₃N@C₈₀ (1 μM), amino-f-Gd₃N@C₈₀ (1 μM), or saline for 30 s before implantation. (A) For the ipsilateral hind paw, saline group demonstrated significantly increased mechanical sensitivity for up to POD 12 compared to the presurgery baseline, while both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ groups exhibited less mechanical sensitivity compared to the saline group (*p < 0.05 for saline vs amino-f-Gd₃N@C₈₀; ^#p < 0.05 for saline vs amino-f-Sc₃N@C₈₀). (B) For contralateral hind paw, no significant difference in mechanical sensitivity was observed among amino-f-Sc₃N@C₈₀, amino-f-Gd₃N@C₈₀, and saline groups (n.s.). n = 6 mice/group in animal behavior tests. In DRG culture, metallofullerenes were diluted in serum-free F12 media. TNF-α treatment significantly induced neuroinflammation as shown in elevated mRNA expression of (C) il-6, (D) cox-2, and (E) secreted IL-6 protein in media (*p < 0.05 vs control DRGs), while (C–E) both amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ effectively attenuated these inflammatory cytokines (~40–50% decline, ^#p < 0.05 vs TNF-α group). In each experiment, bilateral DRGs from 3 to 4 mice were extracted and randomly assigned with 8–10 DRGs in each condition per well. DRGs were cultured in a 48-well plate. Triplicated wells were used per condition per group with experiments independently repeated three times.

There are a few limitations of this study, which can be addressed in future work. Depending on the disease stages, macrophages may display a broad range of phenotypes and functions. Future investigation on how these nanoparticles alter macrophage phenotype and functions along the disease progression might be necessary to fully define the therapeutic potential of these nanomedicine candidates. In this study, a previously established mouse model of NP implantation induced radiculopathy was used; however, other animal models might be utilized to evaluate in vivo effects such as disc herniation induced lumbar radiculopathy.⁵⁵ Future work may also involve how amino-f-Gd₃N@C₈₀ and amino-f-Sc₃N@C₈₀ affect disc tissue homeostasis and whether they could reverse degenerative changes of discs. Meanwhile, MRI imaging might be performed in amino-f-Gd₃N@C₈₀ treated groups to further demonstrate the diagnostic potential. Alternatively, the carbon cage can be functionalized with ¹²⁴I for PET scanning.²⁹

CONCLUSIONS

As shown in Figure 9, we have synthesized and characterized four functionalized metallofullerenes with different surface modifications and encapsulated metals by XPS, MALDI-TOF MS, and UV–vis. We have shown that amine functionalized C₈₀ metallofullerenes amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ possessed superior radical scavenging capabilities, antioxidative stress, and anti-inflammatory responses in macrophages, possibly via inhibiting activity of p38 MAPK/NF-κB p65 signaling pathways. Furthermore, using a mouse model of lumbar radiculopathy, we exhibited that amino-f-Sc₃N@C₈₀ and amino-f-Gd₃N@C₈₀ effectively alleviated mechanical pain sensitivity. In particular, the amino-f-Gd₃N@C₈₀ metallofullerene provided a theranostic MRI contrast agent with suppressed release of Gd³⁺ ions from the C₈₀ fullerene cage.⁵⁶ Since disc degeneration induced low back/leg pain involves complicated interactions among nerve root, disc, and infiltrating leukocytes, our study provides a great opportunities for developing pleiotropic treatment strategies to tackle the inflammation, a critical pathological marker in degenerative disc diseases and associated pain.

Figure 9.

Illustration on mechanism of action of functionalized metallofullerenes including strong free radical scavenging, attenuation of LPS-induced oxidative stress and inflammatory responses in macrophages possibly via regulating p38 MAPK or NF-κB pathways, and amelioration of TNFα elicited neuroinflammation in DRGs, which collectively contributed to the alleviation of mechanical hyperalgesia in an established mouse model of lumbar radiculopathy.