Background
Rheumatoid arthritis (RA) is an autoimmune disease that is characterized by chronic articular inflammation with progressive joint destruction [
1]. The current first-line therapy for RA patients includes the use of conventional disease-modifying antirheumatic drugs (DMARDs), such as methotrexate or leflunomide (LEF), in combination with short-term glucocorticoids. Moreover, the use of biological agents is employed as an alternative therapy in patients whose disease failed to respond to conventional DMARDs [
2].
LEF is an isoxazole derivative with a potent immunosuppressive activity that was approved for the treatment of RA in 1998 [
3]. LEF is a prodrug that is converted in vivo to its primary active metabolite A771726 (also known as teriflunomide). LEF blocks lymphocyte proliferation and hence the clonal expansion of autoreactive T cells in RA patients by inhibiting dihydroorotate dehydrogenase (DHODH), the mitochondrial rate-limiting enzyme in the de novo synthesis of pyrimidine ribonucleotides [
4,
5]. The use of LEF is normally reserved for RA patients whose disease failed to respond to first-line DMARDs, before the introduction of biological DMARDs [
6]. Nevertheless, despite it having demonstrated safety and efficacy, a substantial proportion of patients (around 30–40%) do not have an appropriate response to LEF [
7]. Therefore, it is highly desirable to discover novel DHODH inhibitors as lead compounds for the development of new DMARD candidates.
Lapachol (LAP; 2-hydroxy-3-(3methyl-2-butenyl)-1,4-naphthoquinone) is a nonpolar naturally occurring naphthoquinone found in some Brazilian medicinal plants [
8]. LAP and others naphthoquinones have been described as having a range of biological actions, including microbicidal, anti-inflammatory and antiproliferative activities [
9‐
15]. In fact, it was demonstrated that LAP has potent antitumoral activity which was characterized by its ability to inhibit DNA and RNA synthesis in neoplastic cells [
16]. Moreover, it was reported that LAP can reduce proliferation of the human keratinocytes in vitro, suggesting that it has potential antipsoriatic effects [
17]. Despite the molecular mechanisms associated with these effects remaining poorly elucidated, it has been described that LAP and other naphthoquinones derivatives, such as lawsone and atovaquone, can inhibit DHODH activity [
9]. However, the biological relevance of this effect was poorly characterized. In the present study, we investigated the potential immunosuppressive properties of LAP.
Methods
Preparation of LAP sodium salt
To a solution of LAP (500 mg, 2.06 mmol) in ethanol (20 ml) was added NaOH (112 mg, 0.28 mmol) and the reaction mixture was stirred for 24 h. After consumption of LAP, the reaction mixture was concentrated under reduced pressure and the solid residue was washed with dichloromethane (4×) and petroleum ether (4×) to afford a purple solid of 518 mg, 95% yield (1H NMR (300 MHz, D2Od6) δ7.66 (br s, 1H), 7.64 (br s, 1H), 7.53 (br t, J = 9 Hz, 1H), 7.41 (br t, J = 9Hz, 1H), 5.15 (br t, J = 9Hz, 1H), (3.06, d, J = 6Hz, 2H), 1.72 (s, 3H), 1.63 (s, 3H); 13C NMR (101 MHz, DMSOd6) δ 187.2, 178.4, 169.6, 136.0, 133.2, 131.3, 129.7, 127.5, 125.76 124.81, 124.4, 117.9, 25.59, 20.8, 14.1; HRMS-ESI m/z cald for: [M + Na]+ = 265.0835; found = 265.0834).
Molecular modeling and docking procedures
We used nine human DHODH high-resolution crystal structures in complex with the following inhibitors: DHO1B0033 (PDB id: 4LS0); DSM338 (PDB id: 4OQV); O57 (PDB id: 4JS3); a brequinar analogue (PDB id: 4JTU); 221290 (PDB ID: 2WV8); amino-benzoic acid inhibitor 715 (PDB id: 3KVL); LEF derivative inhibitor 1 (PDB id: 3F1Q); another brequinar analogue (PDB id: 2B0M); and antiproliferative agent A771726 (PDB id: 1D3H). The crystal structure of hDHODH in complex with antiproliferative agent A771726 (PDB id: 1D3H) has been considered for flexible docking with A771726 and LAP, and the calculations were carried out using the GOLD (Genetic Optimisation for Ligand Docking) 5.2 software [
18]. GOLD was comprehensively validated, reliably identifying the correct binding mode for a large range of test set cases, in a vast set of independent studies, with a rate of success in 70–80% of the PDB protein-ligand structures thus analyzed, such as reported in the literature [
18,
19].
Here, a parameter set including a population of 100 conformers, 100,000 operations, 95 mutations, and 95 crossovers has been used. The simulations were then performed inside a selected region of the active site (sphere of 8.5 Å radius centered at
x = 49.65,
y = 42.13,
z = –1.54), keeping the Leu46 side chain flexible (using a rotamers library). The number of docking simulations to be performed with each inhibitor was specified under 10 GA (genetic algorithm) runs, once each docking run can evolve to different ligand poses (pose = conformation + orientation). Thus, ten poses of highest score (top-ranked GOLD solutions) obtained for each compound were selected by using the CHEMPLP score function. In this case, a Piecewise Linear Potential (fPLP) is used to model the steric complementarity between protein and ligand, and for CHEMPLP the distance- and angle-dependent hydrogen and metal bonding terms from other fitness function also implemented in GOLD, so called ChemScore, are considered. CHEMPLP has been found to give the highest success rates for docking pose prediction as well as virtual screening experiments against diverse validation test sets and it was here chosen as the fitness function. Based on this CHEMPLP function, GOLD classifies the orientations of the molecules by a decreasing order of affinity (scores) with the binding site of the receptor [
19].
Previous to the docking calculations and after the removal of the ligand as well as crystallographic waters of the hDHODH/A771726 complex structure, hydrogen atoms of the residues side chains were added and oriented in the active site region. Also, suitable 3D structures of the inhibitors A771726 and LAP were previously built and optimized with molecular mechanics (MMFF force field), followed by Hartree-Fock/Density Functional methods (full optimization at B3LYP/6-31G* level of calculation), using the Spartan’06 software.
Pharmacokinetics study design
For administration to Wistar rats, LAP was dissolved in DMSO:Tween 80:glucose 5% in a proportion of 15:5:80 (v/v/v), resulting in a solution of 1 mg/ml (for intravenous (i.v.) administration) and another of 5 mg/ml (for oral administration). LAP was administered to rats as an i.v. bolus dose at 2 mg/kg (
n = 7) and at two oral doses of 10 (
n = 8) and 25 mg/kg (
n = 6). LAP salt was administered as i.v. (2 mg/kg,
n = 6) and oral doses (30 mg/kg equivalent to 27.5 mg/kg of LAP,
n = 8). The i.v. doses were injected into the lateral tail vein, and oral doses were given by gavage. The doses were chosen based on previous toxicological and pharmacodynamic studies [
20]. At predetermined time points (30 min before dosing and at 0.08, 0.25, 0.5, 1, 2, 4, 6, 8, 12, and 24 h) after LAP i.v. administration, blood samples (200–250 μl) were withdrawn into heparinized tubes via puncture of the lateral tail vein, opposite to the vein used for drug dosing. The same procedure was carried out after oral administration of LAP, with blood sampling at 0.25, 0.5, 1, 1.5, 2, 3, 6, 12, 24, and 30 h. After LAP sodium salt i.v. and oral dosing, blood samples were harvested up to 10 and 12 h, respectively. Plasma was obtained by centrifugation of blood samples (6800 × g, at 4 °C for 10 min) and stored at –80 °C until analysis by UPLC-MS/MS.
Pharmacokinetic analysis
LAP and LAP sodium salt pharmacokinetic parameters after i.v. and oral administration were determined from individual plasma profiles by a noncompartmental approach (NCA). The peak plasma concentration (Cmax) and the time of maximum concentration (Tmax) were obtained by visual inspection of the data from the plasma concentration–time curve after oral dosing. Pharmacokinetic parameters such an elimination rate constant (λ), area under the curve (AUC0–∞), clearance (CLtot), half-life (t1/2), volume of distribution (Vdss), mean residence time (MRT), and bioavailability (Fabs) were determined using classical equations. The compartmental analyses were performed using SCIENTIST v.2.0.1 software (MicroMath®, USA). One- and two-compartment models with or without weighting schemes were evaluated. The best model to fit the data was chosen based on the random distribution of residuals, the correlation coefficient, and the model selection criterion (MSC) given by the software.
The individual plasma profiles of LAP and LAP sodium salt after i.v. administration were best described by a two-compartmental open model. Plasma profiles after oral administration with two different doses of LAP and one dose of LAP sodium salt were best described by the one-compartmental model.
Plasma analysis by UPLC-MS/MS
LAP concentration in plasma samples was determined by a validated (FDA, US Food and Drug Administration, 2001) UPLC-MS/MS method [
21]. Analyses were run on an Acquity UPLC BEH (Waters Acquity™) C18 column (2.1 × 50 mm, 1.7 μm particle size), with a flow of 300 μl/min at 35 °C. A gradient constituted of water (A) and acetonitrile (B) acidified with 0.1% acetic acid was used as follows: 0 min (90% A), 1 min (75% A), 7 min (50% A), 8.5 min (0% A) and 9.5 min (100% A). For the triple quadrupole, MS parameters were set as follows: capillary voltage (2.20 kV); extractor (3.0 V) source temperature (150 °C), desolvation temperature (300 °C), cone gas flow (50 l/h), and desolvation gas flow (700 l/h). For quantification, a multiple reaction monitoring method (MRM) was applied. For LAP, the transition of m/z 243 > 187 using cone energy of 24 V and collision energy of 19 V was determined as most appropriate for quantification (Calibration curves between 1 and 20,000 ng/ml of LAP,
R > 0.99, low quantification limit of 1 ng/ml, and detection limit of 0.1 ng/ml).
Sample preparation for pharmacokinetics studies
A total of 200 μl cold acetonitrile containing internal standard (2-methyl-amino-lapachol) at 5 μg/ml and 0.05% trifluoroacetic acid was added to 100 μl of plasma and vortexed for 20 s. Precipitated protein was removed by centrifugation (6800 × g at 4 °C for 10 min). A total of 200 μl of the supernatant was diluted with purified water 1:1 and filtered by a 0.22-μm membrane before analysis. To prepare the calibration curves, blank plasma samples were spiked with LAP and further processed as indicated. Animal samples with concentrations at the higher upper limit of the calibration curve were diluted with blank plasma before processing.
Enzymatic assay
hDHODH activity was assessed using a colorimetric continuous assay that monitors 2,6-dichloroindophenol (DCIP) reduction. Change in absorbance at 610 nm was monitored over a period of 60 s at 25 °C using a microplate reader (Molecular Devices, SpectraMax 384 Plus, California, USA). The enzymatic reaction was analyzed in a total volume of 195 μl containing 50 mmol/l Tris, pH 8.15, 150 mmol/l KCl, 0.1% Triton X-100, 1 mmol/l l-dihydroorotate, 100 μmol/l CoQ0, and 60 μmol/l DCIP. The assay was started with 5 μl of 0.8 μmol/l stock of enzyme prepared in 50 mmol/l HEPES, pH 7.7, 400 mmol/l NaCl, 10% glycerol, 0.05% Thesit, and 1 mmol/l EDTA in a final concentration of enzyme at 20 nmol/l. A reference measurement was obtained by preparing the same solution without enzyme.
LAP was analyzed in quadruplicate for each concentration used. LAP sodium salt was prepared as a 10 mmol/l stock in DMSO. From this solution, dilutions were prepared in the assay mixture to achieve the range of 100 μmol/l to 0.35 nmol/l. Control enzyme activity in the absence of inhibitor was taken as 100%. The percentage of activity versus log of LAP concentration graph was drawn. The half maximal inhibitory concentration (IC
50) values were calculated using a nonlinear fitting of the concentration–response data to the equation:
$$ activity\ \left(\%\right)= Bottom+\left[\frac{Top- Bottom}{10^{\log \left[ I\right]- \log \left[ I{C}_{50}\right]}+1}\right] $$
Animals
Collagen-induced arthritis (CIA) and antigen-induced arthritis (AIA) models were carried out in male DBA1/J mice (10–12 weeks old) and male C57BL/6 mice (6 weeks old), respectively. The mice were bred and housed in the animal facility of the Ribeirão Preto Medical School (FMRP) at University of São Paulo. For the pharmacokinetic studies, male Wistar rats (200–300 g) were purchased from the State Foundation for the Research and Production in Health (FEPPS, Porto Alegre, Brazil). Animals received water and food ad libitum. All protocols were conducted in accordance with ethical guidelines and approved by the Animal Welfare Committee of FMRP and the Federal University of Rio Grande do Sul (Protocols: 53/2013 and 20244, respectively).
Isolation of CD4 T cells
Human CD4 T cells were purified from the whole blood of healthy volunteers. Briefly, peripheral blood mononuclear cells (PBMCs) were isolated by Percoll gradient (Sigma-Aldrich, St. Louis, MO, USA). CD4 T cells were isolated from PBMCs using Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany). For murine CD4 T cells, lymph nodes from naive C57BL/6 male mice were harvested, and CD4 T cells were purified using Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s recommendations.
Proliferation assay
A total of 1 × 105 CD4 T cells were labeled with 1 μM Dye Efluor® 670 (eBioscience, San Diego, CA, USA) for 15 min at 37 °C and cultured in RPMI-1640 10% FBS for 4 days in a 96-well U-bottom plate (Falcon, Franklin Lakes, New Jersey, USA) with the LAP salt diluted in RPMI-1640 (10, 30, and 100 μM), LEF diluted in RPMI-1640 (10, 30, and 100 μM; Arava®, Sanofi, France) and/or uridine (30, 100, and 300 μM; Sigma-Aldrich, St. Louis, MO, USA) in the presence of anti-CD3 (3 μg/ml) and anti-CD28 (1.5 μg/ml). The proliferation of CD4 T cells was determined by dye dilution in flow cytometry analysis. The results were expressed as the percentage of suppression using the following formula: [proliferation of CD4 T cells only – (proliferation of T CD4 cells with LEF or LAP)/proliferation of CD4 T cells only] × 100.
Collagen-induced arthritis (CIA)
Male DBA/1 J mice were injected i.d. at the base of the tail with 200 μg bovine type II collagen (CII; a gift from Dr. David D. Brand, University of Tennessee Health Science Center) emulsified in Freund’s complete adjuvant (CFA) on day 0. Mice were boosted i.d. with CII (200 μg emulsified in Freund’s incomplete adjuvant (IFA)) on day 21. Mice were monitored daily for signs of arthritis. Scores were assigned based on erythema, swelling, or ankylosis present in each paw on a scale of 0 to 3, giving a maximum score of 12 per mouse. After arthritis induction, mice were treated orally with LAP (3 mg/kg and 10 mg/kg) or LEF (3 mg/kg) or saline daily. The clinical score was addressed every day after collagen boost. All mice were euthanized for histologic assessment of the hind limbs 4 weeks after the boost.
Histological analysis
Femur-tibial joints were collected 4 weeks after CII boost, fixed in 4% (vol/vol) buffered formalin and decalcified in 10% EDTA for 2–3 weeks. The tissues were then trimmed, dehydrated in ethanol, and embedded in paraffin for the preparation of the slides. Histological assessment was carried out following routine staining. Joint sections were stained with hematoxylin and eosin (H&E) to analyze synovitis (inflammatory cell influx and synovial hyperplasia) or Safranin-O to visualize proteoglycan depletion and cartilage destruction. The severity of the joint damage was scored according to the criteria described by Wang et al. [
22]: 0 = no destruction; 1 = minimal erosion; 2 = slight to moderate erosion in a limited area; 3 = more extensive erosion; 4 = general destruction. The degree of synovial pathology (i.e., synovitis) was scored using a scoring system that measured the thickness of the synovial cell layer on a scale of 0–3 (0 = 1–2 cells, 1 = 2–4 cells, 2 = 4–9 cells, and 3 = 10 or more cells) and cellular density in the synovial stroma on a scale of 0–3 (0 = normal cellularity, 1 = slightly increased cellularity, 2 = moderately increased cellularity, and 3 = greatly increased cellularity) [
23].
Cytokine quantification
Interleukin (IL)-17 and interferon (IFN)-γ cytokines were measured by enzyme-linked immunosorbent assay (ELISA) from hind paw homogenate of an individual mouse using antibodies according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN, USA). The results were expressed as pg of cytokine/mg of tissue.
Myeloperoxidase assay
Myeloperoxidase (MPO) activity in tissue homogenates was used as an index of neutrophil infiltration into paws from CIA mice as previously described [
24].
Measurement of liver enzymes
Serum concentrations of aspartate aminotransferase (AST) or alanine aminotransferase (ALT) were measured using commercial kits according to the manufacturer’s instructions (ID Labs Biotechnology Inc., London, Canada).
Antigen-induced arthritis (AIA)
Mice were immunized with methylated bovine serum albumin (mBSA; Sigma-Aldrich, St. Louis, MO, USA) as described previously [
25]. Briefly, mice were immunized with subcutaneous injection of an emulsion with mBSA (500 μg; Sigma-Aldrich, St. Louis, MO, USA) and CFA (2 mg/ml of inactivated
Mycobacterium tuberculosis; Sigma-Aldrich, St. Louis, MO, USA). Booster injections of mBSA in IFA were given at 7 and 14 days after the first immunization. On day 21 after the first immunization, arthritis was induced by an intra-articular injection of mBSA (30 μg). During the AIA protocol, LAP (10 mg/kg) or saline (vehicle) was given orally every day from 12 to 21 days after first immunization.
Determination of joint leucocyte infiltration
Leucocyte infiltration into the joints was assessed 6 h after intra-articular challenge with mBSA as previously described [
26]. Briefly, articular infiltration of leukocytes was determined by washing the femur-tibial joint three times with 3.3 μl phosphate-buffered aline (PBS) + EDTA (0.2 M) and subsequent cell counting was performed in a Neubauer chamber. The results were expressed as the numbers of leucocyte × 10
4 (mean ± SEM)/joint.
Anti-mBSA antibody titer measurement
The titers of serum anti-mBSA antibody were measured by ELISA as previously described [
26].
Recall experiments
Cell suspension (1 × 105 cells) of draining lymph nodes (inguinal) and spleen from naive and mBSA-immunized (treated or not with LAP) mice were stimulated in a 96-well round-bottom plate with mBSA (100 μg/ml) for 96 h. Next, the supernatant was collected to measure the levels of IL-17A, IFN-γ, and IL-4 by ELISA (R&D Systems, Minneapolis, MN, USA).
Statistical analysis
Statistical analyses were performed using one-way nonparametric analysis of variance (ANOVA) followed by Bonferroni’s t test (for three or more groups) comparing all pairs of columns, or two-tailed Student’s t test (for two groups). P < 0.05 was considered statistically significant. Statistical analysis was performed with GraphPad Prism (GraphPad Software, San Diego, CA, USA).
Discussion
In the present study, we conducted a series of in silico, in vitro and in vivo studies describing the biological activity and pharmacokinetic properties of LAP, which is a novel immunosuppressive drug that attenuates experimental autoimmune arthritis through inhibition of DHODH activity. Firstly, we synthetized LAP and performed chemical modifications to improve its solubility in water. In accordance with a previous report [
9], we found that LAP can inhibit the enzymatic activity of hDHODH in vitro. Moreover, we also provided a convincing model for the interaction of LAP with hDHODH by computational docking studies, indicating similar interactions observed with A771726, the active metabolite of LEF. Specifically, the narrow and relatively good hydrophobic pocket of hDHODH allows a suitable accommodation of hydrophobic prenyl and aromatic moieties from LAP. In this case, the analyses predicted a consensual binding mode amongst all the poses calculated for LAP, which additionally interacts by hydrogen bonds with Arg136 and Tyr356 of hDHODH, residues well conserved amongst the mammalian enzymes [
5].
LAP is a naturally occurring naphthoquinone that has been reported to exhibit antitumor, anti-inflammatory, and antimicrobial activities, but the molecular mechanism underlining these effects is poorly understood [
9‐
15]. It was previously reported that some naphthoquinones derivatives, including LAP, can inhibit DHODH activity [
9], but the biological relevance of this observation was not investigated. DHODH is a mitochondrial enzyme that catalyzes the rate-limiting step of the de novo pyrimidine synthesis [
5]. Using lymphocyte proliferation assays, we demonstrated that LAP has a potent immunosuppressive activity on human and murine lymphocytes. Supplementation with uridine, which overcomes the inhibition of pyrimidine synthesis, reversed the antiproliferative activity of LAP on lymphocytes in vitro, demonstrating that the molecular mechanism underlying the antiproliferative effect is mainly due to DHODH inhibition. Importantly, we found that LAP exhibits a greater ability to suppress the proliferation of T cells than observed with LEF in vitro. These results suggest that LAP has immunosuppressive activity on lymphocytes through its direct ability to block DHODH activity and, consequently, inhibit pyrimidine synthesis.
In the pathogenesis of RA, it is well accepted that the influx and proliferation of T cells in the synovial space play a critical role in the articular inflammation and joint destruction [
1,
27,
30]. In fact, autoreactive activated T cells in the joint stimulate plasma cells, mast cells, macrophages, and synovial fibroblasts to produce inflammatory mediators, which in turn stimulate matrix degradation [
4]. Therefore, compounds that inhibit T-cell proliferation have been introduced into the therapeutic schedule of RA [
2]. LEF is a widely used antiproliferative and immunosuppressive drug for treatment of RA that targets DHODH [
4]. However, around 30–40% of RA patients do not have an appropriate response to LEF [
7]. Thus, identification of new small molecule inhibitors targeting DHODH constitutes an attractive therapeutic approach for RA. Taking into account that LAP shows a great ability to inhibit DHODH in vitro, we hypothesized that LAP could have a therapeutic potential in the context of arthritis by interfering with T-cell proliferation. In accordance with its immunosuppressive activity in vitro, we found that LAP effectively attenuated arthritis development and progression in two well-established T cell-dependent models of autoimmune arthritis. Moreover, mice treated with LAP showed a reduction in joint inflammation and articular damage at similar effectiveness as LEF.
Synovial tissue infiltrating inflammatory cells from RA patients are more resistant to apoptotic events, contributing to their accumulation and, consequently, the persistence of inflammation [
31]. The exact mechanism that drives the leucocyte resistance to apoptosis in RA remains unclear, but it is believed that proinflammatory cytokines released in the synovial fluid microenvironment are responsible for this phenomenon [
32]. Since LAP is reducing the production of T cell-dependent proinflammatory cytokines in vivo, it could be indirectly interfering with the apoptosis of inflammatory cells. Thus, LAP and its derivate comprise a potential option for the development of novel lead candidates for treating RA based on DHODH inhibition. Indeed, β-lapachone, a closely related secondary metabolite of LAP, is a promising drug candidate currently in Phase II clinical trials for the treatment of cancer based on its ability to inhibit DHODH (ClinicalTrials.gov identifier: nCT01502800; ClinicalTrials.gov identifier: nCT02514031). However, further studies are needed to determine whether LAP will be effective in inhibiting proliferation of T cells from RA patients who show an inadequate response to LEF.
Acknowledgements
We thank Daniel Callejon for the help in the pharmacokinetic studies, and Sergio Rosa, Ieda Schivo, Ana Katia Santos, and Giuliana Bertozi for the technical support.