Immunomodulatory Effect of the Ursolic Acid/Poly-β-cyclodextrin Complex in an Experimental Model of Multiple Sclerosis

Abstract

Multiple sclerosis (MS) is an autoimmune disease that affects the central nervous system, characterized by demyelination and inflammation. Although treatments are available, they have serious adverse effects and are expensive. Ursolic acid (UA) has anti-inflammatory and neuroprotective properties, but its low solubility hinders its application. This study developed and characterized a ursolic acid with poly-β-cyclodextrin (UA/pβCD) complex and evaluated its immunomodulatory potential in vitro and in vivo. Fourier transform infrared spectroscopy, thermal analyses, and phase solubility experiments confirmed the complexation. The complex improved UA solubility and thermal stability, reduced cytotoxicity in macrophage cell lines, and decreased nitric oxide production. In the experimental autoimmune encephalomyelitis (EAE) model, UA and UA/pβCD reduced neurological disability scores, indicating an attenuation of the inflammatory response. The complex group showed lower levels of IL-12p70 in the brain and spinal cord. These results indicate that the UA/pβCD complex is a promising strategy for the treatment of MS.

Keywords:
multiple sclerosis; experimental autoimmune encephalomyelitis; UA/pβCD complex; poly-β-cyclodextrin; ursolic acid

Introduction

Multiple sclerosis (MS) is a chronic and progressive autoimmune disease that affects the central nervous system (CNS), specifically the brain and spinal cord.¹ It is characterized by the immune-mediated deterioration of myelinated axons, which impairs the transmission of nerve impulses.² The pathogenesis of MS involves the activation of autoreactive T-lymphocytes that trigger an inflammatory reaction against autoantigens of the myelin sheath. This demyelination process leads to a series of manifestations, notably motor, sensory, cognitive, and visual, with the potential to cause permanent physical disability.² Although the exact etiology is unknown, genetic and environmental factors are believed to play a role in the disease’s development.¹

It is estimated that nearly 3 million people worldwide live with MS, with an average incidence of 2.1 per 100,000 people per year.³ Furthermore, there is a higher prevalence in women, with a 2:1 ratio compared to men, and the global average age of diagnosis is 32, a typical age for the economically active population.⁴ Consequently, MS is believed to be the leading neurological cause of disability in young adults, with a significant impact on the quality of life of those affected.⁴

The drugs used in MS treatment reduce inflammation and disease progression by inhibiting immune system cells involved in the disease’s inflammatory cascade.⁵,⁶ Medications available on the market include interferons, glatiramer acetate, rituximab, ocrelizumab, and alemtuzumab, as well as siponimod for progressive forms of MS. These drugs reduce CNS lesions and relapse rates, but do not offer a definitive cure.⁷ In addition, these therapies have a high cost, representing a barrier to access, especially in developing countries.⁸ Another difficulty with treatment is the side effects, which can include flu-like symptoms (interferons), injection site reactions (glatiramer acetate), increased risk of infections and progressive multifocal leukoencephalopathy (rituximab, ocrelizumab), and even bradycardia and changes in liver function (siponimod).⁸

Therefore, there is a need to develop more effective and safer drugs for the treatment of MS. In this development process, ursolic acid (UA) has been highlighted for its neuroprotective and anti-inflammatory effects.⁹,¹⁰ Studies demonstrate that UA effectively enhances myelin repair in adult mouse brains and stimulates exhausted oligodendrocyte progenitors to remyelination.¹¹ However, the therapeutic potential of UA, a natural pentacyclic triterpene found in various plants, is compromised by its low solubility, limited bioavailability, and rapid metabolism. Its reduced water solubility results in inefficient absorption and a short half-life in the body.¹⁰,¹²-¹⁵ To overcome these limitations, researchers have explored synthetic and semi-synthetic modifications of the UA structure to improve its solubility and optimize its therapeutic effects.⁹,¹⁰,¹⁴ When UA is encapsulated with cyclodextrins (CDs), its water solubility increases, leading to increased bioavailability, facilitated by micronization that helps in the intake of UA.¹⁵

CDs are cyclic oligosaccharides synthesized from starch or its derivatives.¹⁶ CDs have a truncated cage configuration with multiple hydroxyl groups at each end, making them effective complexing agents for lipophilic molecules.¹⁶ Complexation increases the solubility and bioavailability of compounds, as well as better targeting and absorption in skin and brain tissues.¹⁶,¹⁷ Natural CDs, such as α, β, and γCDs, can be chemically modified to generate more resistant and efficient derivatives in masking drugs complexed within their structure. Some of these derivatives are: hydroxypropyl β-cyclodextrin (HPβCD); methyl β-cyclodextrin (CH₃βCD); carboxymethyl β-cyclodextrin (cβCD); and poly β-cyclodextrin (pβCD).¹⁶,¹⁷ In the literature, it is possible to find works that demonstrate an improvement in the water solubility of UA, and thus in its low bioavailability, by complexing UA with βCD,¹⁵,¹⁸-²⁰ γCD,¹⁵,²¹,²² HPβCD²⁰,²³-²⁶ and HPγCD.²⁴ These complexes have been tested for antitumor¹⁸,²⁰,²⁴,²⁶ and hepatoprotective actions,¹⁵ but no studies have been found that evaluate the use of the UA/pβCD complex in MS. Furthermore, no works were found regarding the complexation of UA with pβCD. Given this, the present study aims to synthesize and characterize a UA/pβCD complex and evaluate the immunomodulatory effect of this complex in an experimental autoimmune encephalomyelitis (EAE) model, an in vivo model for the study of MS.

Experimental

Chemicals and reagents

The following reagents were used and acquired from Sigma-Aldrich (St. Louis, MO, USA) or Vetec (Rio de Janeiro, RJ, Brazil): β-cyclodextrin (βCD), epichlorohydrin, sodium hydroxide (NaOH) and hydrochloric acid (HCl), ursolic acid (UA), potassium bromide (KBr), dimethyl sulfoxide (DMSO), 3-(4,5-dimethylthiazol-2 yl)-2,5 diphenyltetrazolium bromide (MTT), Roswell Park Memorial Institute (RPMI)-1640 medium, Griess reagent, lipopolysaccharide (LPS), interferon gamma (IFN-γ), bovine serum albumin (BSA), L-glutamine, streptomycin-penicillin antibiotic, fetal bovine serum, phosphoric acid (H₃PO₄), phenylmethylsulfonyl fluoride (PMSF), sodium nitrite (NaNO₂), and ethylenediaminetetraacetic acid (EDTA). The cell lines used in the study were murine macrophages RAW 264.7 (provided by Imunocet, UFJF) and J774A.1 (acquired from the Rio de Janeiro Cell Bank, BCRJ). The kit for quantifying IL-12p70 cytokine was acquired from BD Biosciences (San Diego, CA, USA). The cytokine extraction buffer was prepared with 0.4 mol L^-1 NaCl, 0.05% Tween 20, 0.5% BSA, 0.1 mol L^-1 PMSF, and 0.1 mol L^-1 benzethonium chloride (BC). For EAE induction, MOG 35-55 peptide (GenOne, Campinas, SP, Brazil), Freund’s complete adjuvant, pertussis toxin (Sigma-Aldrich, St. Louis, MO, USA), and Mycobacterium tuberculosis H37RA (Difco, Detroit, MI, USA) were used. For euthanasia, the anesthetics ketamine and xylazine were acquired from Syntec (Brazil) and Calmiun^® (Agener União, Brazil), respectively. pβCD was synthesized as described in “Synthesis of the pβCD” sub-sub-section.

Synthesis and characterization of the complex

Synthesis of the pβCD

pβCD was synthesized via a polymerization reaction of βCD using epichlorohydrin as a chemical crosslinker in a basic medium (5 mol L^-1 NaOH). The reaction was carried out in a DHR1 hybrid dynamic rheometer (TA Instruments), which served as the reactor, under controlled stirring (1000 s^-1) and temperature (40 °C). After synthesis, the pβCD was purified by precipitation in acetone. The resulting solid was dried, redissolved in water, neutralized with HCl to pH 7, and reprecipitated in acetone.²⁷

Synthesis of UA/pβCD complex

The UA/pβCD complex was formed by combining UA and pβCD in a 1:1 molar ratio. For this purpose, 20 mL of an aqueous and/or methanolic solution was used, all at a concentration of 5 mmol L^-1. The solution was subjected to continuous stirring for 24 h at room temperature, after which it was rotary evaporated and lyophilized. The physical mixtures (PM) were prepared in the same 1:1 molar ratio by manual homogenization in a porcelain mortar with a pestle. The samples were then stored in microtubes at -4 °C.

Fourier transform infrared spectroscopy (FTIR)

Solid-state FTIR analyses were obtained on a Spectrum Two™ spectrometer (PerkinElmer) with potassium bromide (KBr) pellets. The spectra were collected with an average of 8 scans, a resolution of 2 cm^-1, in the 4000 400 cm^-1 range. Data was collected using the Spectrum ES software.²⁸

Thermogravimetric analysis (TGA) and differential thermal analysis (DTA)

Thermal analyses were conducted on an STA7200RV thermo-analytical module (Hitachi), as described by Carvalho et al.²⁹ Approximately 5 mg of the samples (UA/pβCD, UA, pβCD, and PM) were analyzed in a platinum crucible, using an empty crucible as a control, with a heating rate of 10 °C min^-1, in the range of 30 to 900 °C.

Phase solubility experiment (PSE)

Solutions of pβCD were prepared at concentrations from 0 to 12.3 mmol L^-1, which were preheated to 40 °C for 20 min. To each solution (3 mL), 1 mg of UA was added, in a way that this amount was never completely dissolved. The samples were kept under stirring at 30 °C for 48 h in a thermal shaker and subsequently filtered using a 0.45 μm membrane. The concentration analyses were performed by UV-Vis spectrophotometry (λ = 210 nm) using a Lambda 25 PerkinElmer instrument.²¹

From the solubility data, phase solubility diagrams were constructed by plotting the total concentration of UA as a function of the pβCD concentration. The apparent stability constant (K_1:1) was determined according to the Higuchi and Connors³⁰ model, which is applicable to a 1:1 stoichiometric complex, as shown in equation 1:

(1) $${\mathrm{K}}_{1:1}=\frac{\text{slope}}{{\mathrm{S}}_0(1-\text{slope})}$$

The complexation efficiency (CE), a dimensionless parameter, was calculated using equation 2 according to Inoue et al.:²¹

(2) $$\mathrm{CE}=\frac{\text{slope}}{1-\text{slope}}-{\mathrm{K}}_{1:1}\times {\mathrm{S}}_0$$

where S₀ is the solubility of UA in the absence of CD.

The percentage increase in the apparent solubility (AS) of UA per millimole of pβCD was obtained according to equation 3:

(3) $$\mathrm{AS}=\frac{[\mathrm{UAmax}]-\left[{\mathrm{S}}_0\right]}{\left[{\mathrm{S}}_0\right]\times [\mathrm{CDmax}]}\times 100$$

where UAmax is the highest observed solubility of UA in the presence of pβCD, S₀ is the solubility of free UA, and [CDmax] corresponds to the maximum concentration of pβCD in the equilibrium liquid phase.

Colloidal characterization by dynamic light scattering (DLS), zeta potential (ZP) and electrical conductivity

DLS and ZP analyses were performed at 25 °C, with Peltier temperature control, using a Malvern Zetasizer Nano ZS90. UA and UA/pβCD solutions were prepared in DMSO at a concentration of 30 mM. For the measurements, 25 injections of 20 μL of each sample were performed into 1.5 mL of ultrapure water. For DLS analyses, the samples were placed in polyethylene cuvettes. A He-Ne laser (4 mW, 633 nm) was used to measure the scattered light at a 90° angle. The hydrodynamic diameter (Dh) was determined from five independent measurements, each being the average of five counts, ensuring data robustness. ZP measurements were conducted by micro-laser-doppler electrophoresis (MELD) in a Dip Cell measurement cell. The light scattering angle was 173°, and an alternating potential difference (APD) of 10 V was applied. The ZP was calculated from the electrophoretic mobility using the Smoluchowski model, and the final value was obtained by averaging five independent measurements.

In vitro assays - J774.A1 and RAW 264.7 macrophages

Cell culture

The murine macrophage cell lines J774A.1 and RAW 264.7 were cultured in RPMI-1640 medium supplemented with L-glutamine (2 mmol L^-1), streptomycin and penicillin (100 µg mL^-1 each), and 5% fetal bovine serum. The cells were maintained in a humidified atmosphere at 37 °C with 5% CO₂. After reaching confluence, the cells were quantified and seeded into 96 well plates at a concentration of 2 × 10⁵ cells well^-1. The cultures were treated for 48 h with pβCD, UA, and the UA/pβCD complex at 15, 30, and 60 μmol L^-1 concentrations. For nitric oxide (NO) production analysis, cells were stimulated with 10 μg mL^-1 of LPS and 9 ng mL^-1 of IFN-γ. Stimulated, untreated cells were used as a positive control, while unstimulated and untreated cells served as a negative control.

Cell viability assay by MTT

Cell viability was determined by the colorimetric MTT assay, as described by Carvalho et al.²⁹ After 48 h of incubation, the supernatants were discarded, and 100 μL of RPMI-1640 medium and 10 μL of the MTT solution (5 mg mL^-1) were added to each well. After 4 h of incubation, the dye was removed and replaced with 100 μL of DMSO. Absorbance was measured at 570 nm using a microplate spectrophotometer (Multiskan™ FC Microplate Photometer).³¹

NO quantification by the Griess method

The NO concentration was indirectly estimated by quantifying nitrite using the Griess method, as described by Carvalho et al.²⁹ After 48 h of incubation, 100 μL of the supernatants were transferred to a new plate, and 100 μL of the Griess reagent was added. The nitrite concentration was determined using a standard nitrite curve, with absorbance measured at 540 nm using a microplate spectrophotometer (Multiskan™ FC Microplate Photometer).

In vivo assays - EAE

Animals and ethical aspects

Female C57BL/6 mice, 6-8 weeks old, were obtained from the Universidade Federal de Minas Gerais. The animals were housed in the Universidade Federal dos Vales do Jequitinhonha e Mucuri (UFVJM) bioterium (Teófilo Otoni-MG) at room temperature with ad libitum access to food and water, in compliance with the Brazilian Code for the Use of Laboratory Animals. All procedures were approved by the CEUA/Mucuri of UFVJM (No. 03-2023).

Experimental autoimmune encephalomyelitis induction

The EAE model induction followed the protocol described by Ramos et al.³² Each induced animal was immunized via subcutaneous (s.c.) injection of 100 μL on each side of the dorsal region near the tail base. The injection contained 100 μg of MOG_35-55 emulsified in Freund’s complete adjuvant at a 1:1 ratio, supplemented with 4 mg mL^-1 of Mycobacterium tuberculosis, H37RA strain. On the day of immunization and 48 h later, each animal received 0.3 μg of pertussis toxin via intraperitoneal (i.p.) injection. On the 21^st day post-immunization (dpi), the animals were euthanized by anesthetic deepening and exsanguination, using xylazine (10 mg kg^-1) and ketamine (100 mg kg^-1) i.p.. Brain and spinal cord samples were collected and stored at -80 °C for subsequent analyses.

Neurological disability score and body mass

Body mass was recorded daily from day zero, while the neurological disability score (NDS) began on the 8^th dpi. The NDS was determined using a neurological disability scale developed by De Paula et al.³³ Treatment was initiated on the 14^th dpi, with the onset of clinical signs. The animals were distributed into five groups (n = 10 mice per group). Group 1 (negative control, NC) consisted of non-induced and untreated animals. Groups 2, 3, and 4 were induced animals, treated with UA, pβCD, or UA/pβCD, respectively. Group 5 (positive control, PC) was induced animals not receiving treatment. All treatments were administered at 50 mg kg^-1 of body weight, via i.p. injection. The 50 mg kg^-1 dose was chosen based on previous studies⁹ demonstrating its efficacy and safety in experimental EAE models.

Cytokine quantification

The cytokine IL-12p70 was quantified by enzyme-linked immunosorbent assay (ELISA) in the mice’s brain and spinal cord samples. The tissues were weighed and macerated for sample preparation using 100 mg mL^-1 of cytokine extraction buffer. The homogenate was then centrifuged at 10,000 rpm for 15 min at 4 °C, and the supernatant was collected and stored at -80 °C until analysis. Cytokine quantification was performed using a commercial ELISA kit, following the manufacturer’s instructions (BD Biosciences^®).

Histology - assessment of inflammatory infiltrate and demyelination

To assess inflammatory infiltrate and demyelination, spinal cord and brain tissues from euthanized animals were perfused and fixed in 10% buffered formalin. Transverse sections (5 μm) were stained with hematoxylin-eosin (H&E) and Luxol fast blue (LFB) following strict protocols. A pathologist blindly performed a semi-quantitative evaluation of inflammatory infiltrate and demyelination from images captured at 100× and 400× magnification. The degree of inflammation and demyelination was scored on a scale from 0 (absence) to 5 (marked tissue involvement > 50%).³⁴

Statistical analysis

Data from the characterization of the complex were analyzed using Microcal Origin 9.0 software,³⁵ and chemical interactions were evaluated by comparing the profiles of the compounds. Data from the in vitro and in vivo experiments were analyzed using Prism 6.0 software.³⁶ The results are the mean of three independent experiments. Quantitative data are expressed as mean ± standard error of the mean (SEM). The normality of each data set was tested using the Kolmogorov-Smirnov and Shapiro-Wilk tests. For variables that presented parametric distribution (IL-12 quantification, cell viability, and NO quantification), one-way analysis of variance (ANOVA) was used, followed by Dunnett’s post-hoc test for comparisons between groups. For variables with nonparametric distribution (body mass, disability score, inflammation score, and demyelination score), the analysis was performed using the Kruskal-Wallis test, and multiple comparisons were made using Dunn’s post-hoc test. The significance level adopted for all analyses was p < 0.05.

Results and Discussion

Characterization of the complex

Fourier transform infrared spectroscopy (FTIR)

UA showed characteristic bands at 3432, 2921, 2854, and 1716 cm^-1 (Figure 1), which correspond respectively to the O-H stretching, superimpositions of C-H, C-H₂, and C-H₃ stretching, C-H methylene stretching, and C=O stretching.²²,³⁷,³⁸ pβCD exhibited bands at 3427, 2933, 1636, and 1036 cm^-1, associated with O-H stretching, C-H stretching, C-O vibration, and the α-1,4 linkage vibration, respectively, similar to observed to native cyclodextrins.²²,³⁷ The PM maintained the bands of UA with minimal alterations. The spectrum of the UA/pβCD complex showed modifications in the intensity and position of the bands when compared to the spectra of isolated UA and pβCD.

Figure 1
FTIR spectra of UA (ursolic acid), pβCD (poly-beta-cyclodextrin), UA/pβCD (ursolic acid poly-beta-cyclodextrin) complex, and PM (physical mixture) in the 4000 to 400 cm^-1 region in KBr.

These results indicate the establishment of intermolecular interactions, possibly of the hydrogen bond or hydrophobic type, between UA and the interior of the pβCD cavity. The difference between the PM and complex spectra confirms that the formation of the inclusion complex does not occur merely by mixing the components in the solid state but depends on the solution medium, where molecular mobility favors the partial or total insertion of UA inside the pβCD. Thus, the FTIR results corroborate the solubility data, pointing to an effective and specific complexation between UA and pβCD, with the direct involvement of functional groups essential for the system’s stability.

Thermogravimetric analysis (TGA) and differential thermal analysis (DTA)

The pβCD TGA curve (Figure 2a) showed mass losses of 3% below 230 °C, 60% in the 230-523 °C range, and 15% in the 525-1000 °C range. UA exhibited mass losses of 2% below 250 °C, 66% around 330 °C, 10% in the 370-515 °C range, and 17% in the 515-1000 °C range. These losses are associated with the loss of water molecules, degradation, and decomposition of the analyzed structures, respectively.

Figure 2
(a) TGA and (b) DTA curves of UA (ursolic acid), pβCD (poly-beta-cyclodextrin), UA/pβCD (ursolic acid poly-beta-cyclodextrin) complex, and PM (physical mixture).

The UA/pβCD complex showed distinct mass losses with lower intensity and greater thermal stability: 6% below 260 °C, 49% in the 260-575 °C range, 6% in the 575-863 °C range, and 29% in the 863-1000 °C range. The PM exhibited a profile similar to that of pβCD, with minor changes. The DTA curve (Figure 2b) shows that UA presented two exothermic events (T_max at 353 and 504 °C), attributed to its decomposition. pβCD exhibited peaks at 328 and 458 °C, related to degradation and carbonization, respectively. The UA/pβCD complex showed a single event with a T_max at 538 °C, attributed to melting followed by decomposition. The PM presented two peaks (362 and 523 °C), attributed to the decomposition process.

These results confirm the formation of a new supramolecular system with thermal properties distinct from the individual substances. The shift in decomposition temperatures and the fusion of exothermic events into a single peak in the complex indicate an effective interaction between UA and pβCD, forming a more thermally stable structure. The absence of direct superposition with the thermal events of the isolated components reinforces the idea that an inclusion complex was effectively formed, in contrast to the physical mixture, which maintained thermal characteristics similar to those of the free cyclodextrin.

Thus, the TGA and DTA data offer complementary evidence to the solubility and FTIR analyses, consolidating the conclusion that UA was efficiently complexed by pβCD, with structural modifications and a gain in thermal stability.

Solubility experiment

The phase solubility diagram of UA with pβCD showed a linear profile in the concentration range of 0 to 12.3 mmol L^-1, classifying it as an AL-type diagram. This suggests the formation of a 1:1 (drug:CD) complex according to the Higuchi and Connors³⁰ model (Figure 3). The AL-type profile indicates that the complexation between UA and pβCD occurs efficiently and proportionally to the increase in pβCD concentration, with no evidence of saturation in this range. This behavior is consistent with data previously reported for βCD and HPβCD¹⁹,²⁰,²⁵ and contrasts with the Bs-type profile observed with γCD, which indicates lower solubility of the formed complex.²¹

Figure 3
Phase solubility diagram for UA/pβCD (ursolic acid poly-beta-cyclodextrin). Results are expressed as the mean ± SEM.

From the equation of the straight line obtained in the solubility diagram, the K_1:1 of the UA/pβCD complex was calculated to be 33.12 L mol^-1. This value is considerably lower than the constant observed for γCD (50,000 L mol^-1), CD-metal organic frameworks (MOF)-1 (1,500 L mol^-1), and modified βCDs (up to 1,800 L mol^-1).²⁰,²¹ The lower affinity observed may be related to the polymeric nature of pβCD, which possibly imposes steric hindrance to the ideal fitting of UA inside the cavity. Even so, the K_1:1 is within the range considered suitable for low to moderate stability complexes, which can be advantageous in formulations aimed at controlled or reversible drug release.³⁹

The corresponding value for Gibbs free energy was ∆G^o = -8.41 kJ mol^-1, indicating the thermodynamic feasibility of the process. The CE was determined to be 2.075 × 10^-3, while the percentage increase in the AS of UA per millimole of pβCD was 3.31%. This was calculated based on the difference between UAmax (0.08809 mol L^-1) and its S₀ (0.06263 mol L^-1) in the presence of 12.3 mol L^-1 of pβCD. For comparison, reported CE values for UA with other CDs include UA/βCD in buffer solutions (pH = 9) with CE = 0.209,¹⁹ UA/γCD with CE = 0.1, and UA/CD-MOF-1 with CE = 3.5.²¹ None of these studies reports AS values for UA/CD systems, and both CE and AS can vary substantially depending on the drug/CD pair, as shown by Couto et al.⁴⁰ for carbamazepine, where AS ranged from 0.97 to 110.1 and CE from 0.006 to 2.33 depending on the cyclodextrin derivative. Although the CE and AS values may seem modest at first glance, they demonstrate a measurable gain in the solubility of a highly hydrophobic molecule like UA. As a polymer, pβCD offers additional advantages, such as greater physicochemical stability and the possibility of functionalization for specific applications, reinforcing its potential as a carrier in drug delivery systems.

Colloidal characterization by DLS, ZP and electrical conductivity

Phase solubility experiments have showed that complexation of UA with pβCD leaves to an increase of UA solubility in aqueous environment. However, although the UA/pβCD complex has greater solubility than pure UA, its absolute solubility remains low, suggesting that for future applications, it may require the use of a co-solvent to disperse the material in aqueous environment.

To address this, we have used the concept of hydrophobic nanoprecipitates (HNPs),⁴¹-⁴³ which are solid, water-insoluble particles typically formed in solvent mixtures where water is the primary component. One of the main advantages of this approach is the ability to develop simple formulations without relying on expensive additives or surfactants, functioning as a release system based on the dissolution of the solid phase.

To investigate the influence of pβCD on the formation of HNPs in a DMSO/water mixture, we performed measurements of hydrodynamic diameter (Dh, via DLS), zeta potential (ZP), and electrical conductivity (κ). Figures 4a-4c present the Dh, ZP, and κ values, respectively, plotted against the nominal concentration of UA. As shown, both UA and UA/pβCD form HNPs upon mixing their DMSO solutions with water, with particle size and zeta potential strongly influenced by concentration.

Figure 4
(a) Hydrodynamic diameter; (b) zeta potential; (c) electrical conductivity values; all of them plotted against the nominal concentration of UA. Experiments performed at 25 °C.

In both systems, particle size increases with concentration, likely due to suspension saturation and a higher probability of particle collisions. As can be observed, the Dh values are larger for the UA/pβCD system compared to the UA-only system, as result of different internal interactions and with the environment. Moreover, the presence of pβCD macromolecules can contribute to the higher Dh values.

Negative ZP values were observed in both systems, attributed to the presence of anionic species. In the UA system, this results from partial ionization (UA^-), while in the UA/pβCD system, it stems from the ionization of both UA and pβCD (UA^- and pβCD^n-). Interestingly, the UA/pβCD system exhibited lower ZP magnitudes up to a concentration of 2.5 mM, along with significantly higher conductivity values.

The elevated conductivity is likely due to the presence of Na⁺ ions, resultant from pβCD synthesis under basic conditions.²⁷ Regarding the reduced ZP values in the UA/pβCD system, we hypothesize that complexation promotes hydrogen bonding or ion-dipole interactions, which suppress UA ionization. Additionally, the abundance of Na⁺ ions may partially neutralize the electrical double layer surrounding the UA/pβCD complex, further decreasing its ZP values.

In vitro assays - J774.A1 and RAW 264.7 macrophages

Effects of the UA/pβCD complex on macrophage cell viability

The cytotoxicity of the UA/pβCD complex was evaluated by a cell viability assay in RAW 264.7 and J774A.1 macrophages (Figure 5). At concentrations up to 30 μmol L^-1, both UA and the UA/pβCD complex showed cell viability greater than 70%, which is considered non-cytotoxic.⁴⁴ At 60 μmol L^-1, a significant reduction in cell viability was observed for both substances. The J774A.1 macrophages showed greater sensitivity (ca. 7% viability) than the RAW 264.7 cells (ca. 11%).

Figure 5
Cell viability of RAW 264.7 and J774A.1 macrophages treated with UA (ursolic acid), pβCD (poly-beta-cyclodextrin), UA/pβCD (ursolic acid poly-beta-cyclodextrin) complex, and DMSO, at concentrations of 15, 30 and 60 μmol L^-1, after 48 h of culture. *Statistical difference (p < 0.05) versus RAW 264.7 or J774A.1 untreated control group. Results are expressed as mean ± SEM. One-way ANOVA was used, followed by Dunnett’s post-hoc test.

These findings indicate that the UA/pβCD complex has low cytotoxicity up to 30 μmol L^-1, making it safe for preliminary biological assays. The difference in response between the two cell lines is consistent with literature data,⁴⁵,⁴⁶ which reports intrinsic variability in cellular susceptibility to bioactive compounds.

In vitro nitric oxide production reduction by UA/pβCD

The NO production was estimated by nitrite quantification in the cell lines after 48 h of treatment (Figure 6). Notably, the UA/pβCD complex promoted a significant reduction in NO production at all concentrations tested in the RAW 264.7 cell line, with a more evident effect at higher concentrations. Similar results were observed in J774A.1, with significant inhibition at 60 and 30 μmol L^-1 concentrations.

Figure 6
Evaluation of NO production inhibition in RAW 264.7 and J774A.1 macrophages treated with UA (ursolic acid), pβCD (poly-beta-cyclodextrin), UA/pβCD (ursolic acid poly-beta-cyclodextrin) complex, and DMSO, at concentrations of 15, 30 and 60 μmol L^-1, and stimulated with LPS and IFN-γ, after 48 h of culture. *Statistical difference (p < 0.05) versus RAW 264.7 or J774A.1 untreated control group. ^#Statistical difference (p < 0.05) versus UA group at the same concentration. Results are expressed as mean ± SEM. One-way ANOVA was used, followed by Dunnett’s post-hoc test.

It is important to highlight that this anti-inflammatory activity was observed even at non-cytotoxic concentrations, reinforcing the complex’s therapeutic potential. Furthermore, the inhibitory effect was more pronounced in the UA/pβCD complex than in isolated UA, especially in the RAW 264.7 cell line, suggesting the possibility of synergism or increased bioavailability of UA in its complexed form with pβCD.⁹

The in vitro results indicate that the complexation of UA with pβCD does not compromise its cellular safety and may potentiate its anti-inflammatory effects, particularly through the inhibition of NO production, one of the main inflammatory mediators produced by activated macrophages.

In vivo assays - EAE

NDS significant reduction by UA and UA/pβCD

Figure 7 presents the NDS data and body mass variation of PC mice. The negative control group remained asymptomatic. The groups treated with UA and UA/pβCD showed a significant reduction in NDS between 16 and 19 dpi compared to the untreated PC group. No statistical difference was observed between the groups treated with UA and UA/pβCD.

Figure 7
(a) Neurological disability score (NDS) and (b) body mass variation of C57BL/6 mice induced with EAE. Induced animals were treated with UA (ursolic acid), pβCD (poly-beta-cyclodextrin), UA/pβCD (ursolic acid poly-beta-cyclodextrin). NC: consisted of non-induced and untreated animals; PC: positive control, consisted of induced and untreated animals. *Statistical difference (p < 0.05) versus PC. Dashed line: beginning of the treatment. Results are expressed as mean ± SEM. Kruskal-Wallis test was used, followed by Dunn’s post-hoc test.

The anti-inflammatory profile observed in vitro was reflected in the in vivo model, where both isolated UA and the UA/pβCD complex reduced the EAE NDS, with greater efficacy during the inflammatory peak. These results corroborate previous findings on the immunomodulatory potential of UA in EAE models.⁴⁷ The lack of a significant difference between the two treatments indicates that complexation with pβCD does not compromise the therapeutic activity of UA, thus representing a promising strategy to improve its pharmacokinetic properties without sacrificing efficacy. Furthermore, the maintenance of body mass in the treated groups suggests that the treatments were well-tolerated, with no apparent signs of systemic toxicity, which supports the toxicological safety of the formulation.

Production of the inflammatory cytokine IL-12p70 reduction by UA/pβCD

The analysis of pro-inflammatory cytokine levels in the brain and spinal cord of EAE-induced mice revealed that the group treated with UA/pβCD showed a significant reduction in the cytokine IL-12p70 when compared to the PC group. The PC group showed the highest levels in brain (9663.0 pg mL^-1) and in spinal cord (3883.5 pg mL^-1) (Figure 8).

Figure 8
Cytokine IL-12p70 levels in the (left) brain and (right) spinal cord of EAE-induced mice. Induced animals were treated with UA (ursolic acid), pβCD (poly-beta-cyclodextrin), UA/pβCD (ursolic acid poly-beta-cyclodextrin). PC: positive control, consisted of induced and untreated animals. *Statistical difference (p < 0.05) versus PC. Results are expressed as mean ± SEM. One-way ANOVA was used, followed by Dunnett’s post-hoc test.

These findings indicate that the UA/pβCD complex promoted a significant and selective reduction of IL-12p70, which points to a more specific immunomodulatory effect. This is particularly relevant as IL-12p70 is a key cytokine in the differentiation of Th1 lymphocytes, playing a central role in promoting inflammation in autoimmune diseases like MS.⁴⁸ Thus, the suppression of this cytokine suggests that treatment with UA/pβCD may directly interfere with this pathogenic inflammatory pathway.

In addition, the significant reduction in NO production observed in the in vitro assays, especially with the UA/pβCD complex treatment, combined with the suppression of IL-12p70 in the CNS, may indicate a functional redirection of macrophages toward a profile compatible with M2 polarization. This polarization is characterized by anti-inflammatory properties and suppression of the Th1 inflammatory pathway.⁴⁹ Although specific phenotypic markers of M2 macrophages were not directly evaluated, the congruence between the in vitro data (NO reduction) and ex vivo data (IL-12p70 suppression) points to a systemically coordinated effect that manifests in peripheral innate immunity and central neuroinflammation.

This modulation can contribute to the attenuation of EAE clinical signs in the treated animals, given that the IL 12/Th1 axis is strongly implicated in the CNS pathogenesis of autoimmune diseases.²,⁴⁹-⁵¹ Therefore, the data presented point to the potential of the UA/pβCD complex as a therapeutic strategy for MS, through inhibiting the Th1 pathway and promoting a more regulatory immunological microenvironment.

Histology - assessment of inflammatory infiltrate and demyelination

Histopathological analysis was conducted to provide direct morphological evidence of inflammation and demyelination in the CNS, correlating these findings with clinical and immunological data. The evaluation of inflammatory infiltrate in the brain (Figure 9e) did not reveal significant distinctions between the groups treated with UA and UA/pβCD, which exhibited similar and significantly lower levels of inflammation compared to the PC group. These results corroborate the suppression of IL 12p70 observed in the same region, suggesting that cytokine modulation is reflected in the reduction of inflammatory cell accumulation. In the spinal cord (Figure 9f), there were no statistically significant differences in inflammation among the groups, although the UA/pβCD group showed a trend toward lower inflammatory scores. The absence of a statistically significant reduction in the spinal cord is consistent with the lack of change in IL-12p70 levels in this region, which was expected.

Figure 9
Histopathological analysis of inflammation in the CNS of mice induced for EAE. (a,c) Brain and spinal cord slides, respectively, at 10× magnification; (b,d) 40× magnifications of the respective slides. Arrows indicate inflammatory foci. The scale bar in all images is 200 µm; (e,f) graphs represent the inflammation score in the brain and spinal cord, respectively. UA: treated with ursolic acid; pβCD: treated with poly-β-cyclodextrin; UA/pβCD: treated with ursolic acid poly-β-cyclodextrin complex; NC: non-induced control; PC: induced control, untreated. The number of animals per group was n = 4. *p < 0.05 compared to the PC group; Kruskal-Wallis test was used, followed by Dunn’s post-hoc test.

The analysis of demyelination in the brain (Figure 10c) demonstrated that the UA and pβCD groups produced a significant reduction in demyelination levels relative to the PC group. It is interesting to note that despite pβCD not reducing the clinical EAE score, it led to a reduction in both the IL-12p70 cytokine and demyelination in the brain, indicating that this cyclodextrin may have an important action in EAE. In the spinal cord (Figure 10d), the demyelination analysis was even more striking, showing that the UA and UA/pβCD groups had a significant reduction in demyelination levels. The reduction in demyelination observed here, without a significant change in the inflammatory score, suggests that the compounds have a direct protective effect on the myelin sheath, independent of the intensity of the cellular infiltrate.

Figure 10
Histopathological analysis of demyelination in the CNS of mice induced for EAE. (a,b) Brain and spinal cord slides, respectively, at 10× magnification. Arrows indicate foci of demyelination. The scale bar in all images is 200 µm; (c,d) graphs represent the demyelination score in the brain and spinal cord, respectively. UA: treated with ursolic acid; pβCD: treated with poly-β-cyclodextrin; UA/pβCD: treated with ursolic acid poly-β cyclodextrin complex; NC: non-induced control; PC: induced control, untreated. The number of animals per group was n = 4. *p < 0.05 compared to the PC group. Kruskal-Wallis test was used, followed by Dunn’s post-hoc test.

Conclusions

The physicochemical characterization confirmed the formation of the UA/pβCD complex, which has a lower equilibrium constant compared to native UA complexes with non-polymeric cyclodextrins. Colloidal characterization demonstrated that the complex forms stable hydrophobic nanoprecipitates (HNPs) with a larger hydrodynamic diameter and distinct surface charge properties compared to uncomplexed UA, indicating altered interactions in the aqueous environment. In in vitro tests, this complex stood out for its potent anti-inflammatory activity, capable of reducing NO production by RAW 264.7 and J774A.1 macrophages without compromising cell viability, demonstrating superior efficacy over the isolated compounds.

In the EAE model, UA/pβCD significantly reduced the clinical scores of the disease, without impacting body mass, and modulated the immune response by decreasing IL-12p70 production in the central nervous system. These results were corroborated by histopathological findings, which revealed a significant reduction in both inflammatory infiltrate and demyelination in the brain. Notably, in the spinal cord, the complex also caused a marked reduction in demyelination, suggesting a direct protective effect on the myelin sheath that is potentially independent of the intensity of the cellular infiltrate.

Collectively, these results show that UA/pβCD has high efficacy and therapeutic potential, both in vitro and in vivo, and stands out as a promising strategy for treating neuroinflammatory and demyelinating diseases, such as MS.

Acknowledgments

The authors gratefully acknowledge the financial support provided by the National Council for Scientific and Technological Development (CNPq) and by the Minas Gerais State Research Foundation (FAPEMIG) under grant number APQ 02052/21, APQ 02056/22, and APQ 02762/25. We also thank Universidade Federal dos Vales do Jequitinhonha e Mucuri (UFVJM), and the Universidade Federal de Juiz de Fora, Governador Valadares Campus (UFJF/GV) for providing research infrastructure. J. C. S. Viana: CNPq scholarship holder, Brazil; V. A. Lopes: CNPq scholarship holder, Brazil.

Data Availability Statement

All data generated or analyzed during this study are included in this published article.

References

• ¹ de Oliveira, J. C.; Ferrari, C. A.; Pentagna, D. C.; Lino, L. L.; Cavalcante, R. S. G.; Braz. J. Health Rev. 2024, 7, e72161. [Crossref]
  » Crossref
• ² Haki, M.; AL-Biati, H. A.; Al-Tameemi, Z. S.; Ali, I. S.; Al-hussaniy, H. A.; Medicine 2024, 103, e37297. [Crossref]
  » Crossref
• ³ Multiple Sclerosis International Federation (MSIF); Atlas of MS 2020: Mapping Multiple Sclerosis around the World - Key Epidemiology Findings; MSIF: London, 2020. [Link] accessed in October 2025
  » Link
• ⁴ Walton, C.; King, R.; Rechtman, L.; Kaye, W.; Leray, E.; Marrie, R. A.; Robertson, N.; La Rocca, N.; Uitdehaag, B.; van der Mei, I.; Wallin, M.; Helme, A.; Napier, C. A.; Rijke, N.; Baneke, P.; Mult. Scler. J. 2020, 26, 1816. [Crossref]
  » Crossref
• ⁵ Samjoo, I. A.; Worthington, E.; Drudge, C.; Zhao, M.; Cameron, C.; Häring, D. A.; Stoneman, D.; Klotz, L.; Adlard, N.; J. Comp. Eff. Res. 2021, 10, 495. [Crossref]
  » Crossref
• ⁶ Wei, W.; Ma, D.; Li, L.; Zhang, L.; Front. Pharmacol. 2021, 12, 724718. [Crossref]
  » Crossref
• ⁷ Callegari, I.; Derfuss, T.; Galli, E.; Presse Med. 2021, 50, 104068. [Crossref]
  » Crossref
• ⁸ Burman, J.; Mult. Scler. Relat. Disord. 2021, 54, 103134. [Crossref]
  » Crossref
• ⁹ Scherrer, E. C.; Valadares, Y. M.; Alves, C. C. S.; Carli, A. P.; Fernandes, B. G. R.; Carvalho, P. E.; Ramos, K. A.; Salvador, M. R.; da Silva, J. G.; Silva, F. S.; Denadai, Â. M. L.; Castro, S. B. R.; J. Braz. Chem. Soc. 2023, 34, 1250. [Crossref]
  » Crossref
• ¹⁰ Zafar, S.; Khan, K.; Hafeez, A.; Irfan, M.; Armaghan, M.; ur Rahman, A.; Gürer, E. S.; Sharifi-Rad, J.; Butnariu, M.; Bagiu, I.-C.; Bagiu, R. V.; Cancer Cell Int. 2022, 22, 399. [Crossref]
  » Crossref
• ¹¹ Honarvar, F.; Hojati, V.; Zare, L.; Bakhtiari, N.; Javan, M.; J. Mol. Neurosci. 2022, 72, 2081. [Crossref]
  » Crossref
• ¹² Lei, P.; Li, Z.; Hua, Q.; Song, P.; Gao, L.; Zhou, L.; Cai, Q.; Int. J. Mol. Sci. 2023, 24, 14771. [Crossref]
  » Crossref
• ¹³ Liu, G.; Qin, P.; Cheng, X.; Wu, L.; Wang, R.; Gao, W.; Front. Vet. Sci. 2023, 10, 1251248. [Crossref]
  » Crossref
• ¹⁴ Sun, Q.; He, M.; Zhang, M.; Zeng, S.; Chen, L.; Zhou, L.; Xu, H.; Fitoterapia 2020, 147, 104735. [Crossref]
  » Crossref
• ¹⁵ Žaloudková, L.; Tichá, A.; Nekvindová, J.; Pavlíková, L.; Zadák, Z.; Živný, P.; J. Evidence-Based Complementary Altern. Med. 2020, 2020, 4074068. [Crossref]
  » Crossref
• ¹⁶ Wüpper, S.; Lüersen, K.; Rimbach, G.; Biomolecules 2021, 11, 401. [Crossref]
  » Crossref
• ¹⁷ Alshati, F.; Alahmed, T. A. A.; Sami, F.; Ali, M. S.; Majeed, S.; Murtuja, S.; Hasnain, M. S.; Ansari, M. T.; Curr. Pharm. Des. 2023, 29, 2853. [Crossref]
  » Crossref
• ¹⁸ Fajardo, J. B.; Vianna, M. H.; Ferreira, T. G.; Lemos, A. S. O.; Souza, T. F.; Campos, L. M.; Paula, P. L.; Andrade, N. B.; Gamarano, L. R.; Queiroz, L. S.; Oliveira, B. A.; da Silva, A. D.; Chedier, L. M.; Denadai, Â. M. L.; Tavares, G. D.; Barradas, T. N.; Fabri, R. L.; ACS Omega 2025, 10, 12906. [Crossref]
  » Crossref
• ¹⁹ Huang, Y.; Quan, P.; Wang, Y.; Zhang, D.; Zhang, M.; Li, R.; Jiang, N.; J. Biomed. Res. 2017, 31, 395. [Crossref]
  » Crossref
• ²⁰ Song, S.; Gao, K.; Niu, R.; Yi, W.; Zhang, J.; Gao, C.; Yang, B.; Liao, X.; J. Mol. Liq. 2019, 296, 111993. [Crossref]
  » Crossref
• ²¹ Inoue, Y.; Motoda, A.; Tanikawa, T.; Takao, K.; Arce Jr., F.; See, G. L.; Ishida, Y.; Nakata, D.; Terao, K.; J. Drug Delivery Sci. Technol. 2023, 89, 104986. [Crossref]
  » Crossref
• ²² Zong, W.; Bi, S. M.; Adv. Mater. Res. 2012, 403-408, 712. [Crossref]
  » Crossref
• ²³ Li, R.; Quan, P.; Liu, D.-F.; Wei, F.-D.; Zhang, Q.; Xu, Q.-W.; AAPS PharmSciTech 2009, 10, 1137. [Crossref]
  » Crossref
• ²⁴ Oprean, C.; Mioc, M.; Csányi, E.; Ambrus, R.; Bojin, F.; Tatu, C.; Cristea, M.; Ivan, A.; Danciu, C.; Dehelean, C.; Paunescu, V.; Soica, C.; Biomed. Pharmacother. 2016, 83, 1095. [Crossref]
  » Crossref
• ²⁵ Quan, P.; Liu, D.; Li, R.; Zhang, Q.; Qian, Y.; Xu, Q.; J. Inclusion Phenom. Macrocyclic Chem. 2009, 63, 181. [Crossref]
  » Crossref
• ²⁶ Soica, C.; Oprean, C.; Borcan, F.; Danciu, C.; Trandafirescu, C.; Coricovac, D.; Crăiniceanu, Z.; Dehelean, C. A.; Munteanu, M.; Molecules 2014, 19, 4924. [Crossref]
  » Crossref
• ²⁷ Renard, E.; Deratani, A.; Volet, G.; Sebille, B.; Eur. Polym. J. 1997, 33, 49. [Crossref]
  » Crossref
• ²⁸ PerkinElmer; Spectrum 10 Enhanced Security (Spectrum 10 ES), version 10.03.08.0133; PerkinElmer, Waltham, MA, USA, 2024.
• ²⁹ Carvalho, P. E.; Salvador, M. R.; Fernandes, B. G. R.; Souza, C. A.; Carneiro, M. A. G.; Cachuba, R. M.; Amaro, B. R.; Oliveira, A. R.; Alves, C. C. S.; Carli, A. P.; da Silva, J. G.; Denadai, Â. M. L.; Castro, S. B. R.; J. Braz. Chem. Soc. 2025, 36, e-20250062. [Crossref]
  » Crossref
• ³⁰ Higuchi, T.; Connors, K. A. In Advances in Analytical Chemistry and Instrumentation, vol. 4; Reilly, C. N., ed.; Wiley-Interscience: New York, 1965, p. 117.
• ³¹ Castro, S. B. R.; Junior, C. O. R.; Alves, C. C. S.; Dias, A. T.; Alves, L. L.; Mazzoccoli, L.; Zoet, M. T.; Fernandes, S. A.; Teixeira, H. C.; Almeida, M. V.; Ferreira, A. P.; Chem. Biol. Drug Des. 2012, 79, 347. [Crossref]
  » Crossref
• ³² Ramos, K. A.; Soares, I. G. M.; Oliveira, L. M. A.; Braga, M. A.; Soares, P. P. C.; Guarneire, G. J.; Scherrer, E. C.; Silva, F. S.; Lima, N. M.; La Porta, F. A.; Andrade, T. J. A. S.; Preet, G.; Castro, S. B. R.; Alves, C. C. S.; Carli, A. P.; Curr. Issues Mol. Biol. 2024, 46, 8726. [Crossref]
  » Crossref
• ³³ De Paula, M. L.; Rodrigues, D. H.; Teixeira, H. C.; Barsante, M. M.; Souza, M. A.; Ferreira, A. P.; Int. Immunopharmacol. 2008, 8, 1291. [Crossref]
  » Crossref
• ³⁴ Pant, A. B.; Wang, Y.; Mielcarz, D. W.; Kasper, E. J.; Telesford, K. M.; Mishra, M.; Haque, A.; Channon, J. Y.; Kasper, L. H.; Begum-Haque, S.; J. Neuroimmunol. 2017, 303, 22. [Crossref]
  » Crossref
• ³⁵ OriginLab Corporation; Origin, version 9.0; OriginLab Corporation, Northampton, MA, USA, 2012.
• ³⁶ GraphPad Software, Inc.; GraphPad Prism, version 6.0; GraphPad Software, San Diego, CA, USA, 2012.
• ³⁷ Pratt, D. Y.; Wilson, L. D.; Kozinski, J. A.; Mohart, A. M.; J. Appl. Polym. Sci. 2010, 116, 2982. [Crossref]
  » Crossref
• ³⁸ Yang, L.; Sun, Z.; Zu, Y.; Zhao, C.; Sun, X.; Zhang, Z.; Zhang, L.; Food Chem. 2012, 132, 319. [Crossref]
  » Crossref
• ³⁹ Feng, Y.; Chen, S.; Li, Z.; Gu, Z.; Xu, S.; Ban, X.; Hong, Y.; Cheng, L.; Li, C.; Crit. Rev. Food Sci. Nutr. 2023, 63, 4744. [Crossref]
  » Crossref
• ⁴⁰ Couto, A. R. S.; Ryzhakov, A.; Larsen, K. L.; Loftsson, T.; ACS Omega 2019, 4, 1460. [Crossref]
  » Crossref
• ⁴¹ Bittencourt, V. C. E.; Moreira, A. M. D. S.; da Silva, J. G.; Gomides, A. F. F.; Velloso-Rodrigues, C.; Kelmann, R. G.; Mendonça, L. M.; Lula, I. S.; Denadai, Â. M. L.; Heliyon 2019, 5, e02013. [Crossref]
  » Crossref
• ⁴² Bittencourt, V. C. E.; Moreira, A. M. S.; Ferreira, A. A.; Ferreira, E. A. E.; Gomides, A. F. F.; Rodrigues, C. V.; Martins, G. F.; da Silva, J. G.; Borges, K. B.; Denadai, Â. M. L.; J. Braz. Chem. Soc. 2021, 32, 2023. [Crossref]
  » Crossref
• ⁴³ Moreira, A. M. S.; Bittencourt, V. C. E.; Costa, F. L. S.; de Lima, M. E.; Lopes, M. T. P.; Borges, W. S.; Martins, G. F.; Nascimento Jr., C. S.; da Silva, J. G.; Denadai, Â. M. L.; Borges, K. B.; J. Agric. Food Chem. 2018, 66, 7275. [Crossref]
  » Crossref
• ⁴⁴ International Organization for Standardization (ISO); ISO 10993-5:2009 - Biological Evaluation of Medical Devices - Part 5: Tests for in vitro Cytotoxicity; ISO: Geneva, 2009. [Link] accessed in October 2025
  » Link
• ⁴⁵ Feito, M. J.; Casarrubios, L.; Oñaderra, M.; Gómez-Duro, M.; Arribas, P.; Polo-Montalvo, A.; Vallet-Regí, M.; Arcos, D.; Portolés, M. T.; Colloids Surf., B 2021, 208, 112110. [Crossref]
  » Crossref
• ⁴⁶ Gunawardena, D.; Karunaweera, N.; Lee, S.; van Der Kooy, F.; Harman, D. G.; Raju, R.; Bennett, L.; Gyengesi, E.; Sucher, N. J.; Münch, G.; Food Funct. 2015, 6, 910. [Crossref]
  » Crossref
• ⁴⁷ Zhang, Y.; Li, X.; Ciric, B.; Curtis, M. T.; Chen, W.-J.; Rostami, A.; Zhang, G.-X.; Proc. Natl. Acad. Sci. 2020, 117, 9082. [Crossref]
  » Crossref
• ⁴⁸ Upadhye, M.; Wilhelm, C. R.; Rogers, K. J.; Vemulawada, C.; Borcherding, N.; Boyden, A. W.; Legge, K. L.; Karandikar, N. J.; J. Neuroinflammation 2025, 22, 54. [Crossref]
  » Crossref
• ⁴⁹ Chu, F.; Shi, M.; Lang, Y.; Chao, Z.; Jin, T.; Cui, L.; Zhu, J.; Clin. Exp. Immunol. 2021, 204, 199. [Crossref]
  » Crossref
• ⁵⁰ Jana, M.; Mondal, S.; Jana, A.; Pahan, K.; Cytokine 2024, 174, 156457. [Crossref]
  » Crossref
• ⁵¹ Viswanathan, L. G.; Babu, G. S.; Dey, S.; Kavya, P. V.; Mullapudi, T.; Thennarasu, K.; Ravindranadh, M.; Kenchaiah, R.; Nagappa, M.; Seshagiri, D.; Asranna, A.; Sinha, S.; Debnath, M.; J. Neuroimmunol. 2025, 400, 578532. [Crossref]
  » Crossref

Edited by

• Editor handled this article:
  Brenno A. D. Neto (Editor-in-Chief)

Publication Dates

• Publication in this collection
  28 Nov 2025
• Date of issue
  2025

History

• Received
  21 Aug 2025
• Published
  08 Oct 2025