Beneficial Modulation of Lipid Mediator Biosynthesis in Innate Immune Cells by Antirheumatic Tripterygium wilfordii Glycosides

Tripterygium wilfordii glycosides (TWG) is a traditional Chinese medicine with effectiveness against rheumatoid arthritis (RA), supported by numerous clinical trials. Lipid mediators (LM) are biomolecules produced from polyunsaturated fatty acids mainly by cyclooxygenases (COX) and lipoxygenases (LOX) in complex networks which regulate inflammation and immune responses and are strongly linked to RA. The mechanism by which TWG affects LM networks in RA treatment remains elusive. Employing LM metabololipidomics using ultra-performance liquid chromatography-tandem mass spectrometry revealed striking modulation of LM pathways by TWG in human monocyte-derived macrophage (MDM) phenotypes. In inflammatory M1-MDM, TWG (30 µg/mL) potently suppressed agonist-induced formation of 5-LOX products which was confirmed in human PMNL and traced back to direct inhibition of 5-LOX (IC50 = 2.9 µg/mL). TWG also efficiently blocked thromboxane formation in M1-MDM without inhibiting other prostanoids and COX enzymes. Importantly, in anti-inflammatory M2-MDM, TWG (30 µg/mL) induced pronounced formation of specialized pro-resolving mediators (SPM) and related 12/15-LOX-derived SPM precursors, without COX and 5-LOX activation. During MDM polarization, TWG (1 µg/mL) decreased the capacity to generate pro-inflammatory 5-LOX and COX products, cytokines and markers for M1 phenotypes. Together, suppression of pro-inflammatory LM but SPM induction may contribute to the antirheumatic properties of TWG.


Introduction
Lipid mediators (LM) encompass oxygenated polyunsaturated fatty acids (PUFA) which are involved in maintenance of normal hemostasis but also display significant roles in host defense as well as in pain, fever and in inflammation and its resolution [1][2][3]. They are mainly derived from arachidonic acid (AA), eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) that are liberated from membrane phospholipids by phospholipases (PL)A 2 upon cell stimulation [4,5]. Cyclooxygenases (COX)-1/2, six lipoxygenases (LOX) in humans, and CYP enzymes convert these free PUFA towards a broad spectrum of ter was produced by a Sartorius Arium 611 UV water purification system (Göttingen, Germany). Deuterated and non-deuterated LM standards for ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS-MS) were purchased from Cayman Chemicals (Ann Arbor, MI, USA). All other chemicals and reagents were obtained from Sigma-Aldrich (Munich, Germany), unless stated otherwise.

High-Performance Liquid Chromatography (HPLC) Analysis of TWG and Celastrol
For analysis of TWG and celastrol determination we applied a gradient-based C18reversed phase high-performance liquid chromatography (RP-HPLC) using an Agilent 1260 Infinity II system equipped with a diode-array-detection system (all Agilent Technologies Inc., Santa Clara, CA, USA) and a Xterra ® RP18 5 µm (4.6 mm × 250 mm) column (Waters Corporation, Milford, MA, USA). Detection was performed at 421 nm. The software OpenLAB CDS Rev. C.01.07 (Agilent Technologies) was used for method control and analysis. As mobile phase A acetonitrile (HPLC grade) and as mobile phase B 1% phosphoric acid in ultrapure water (both Carl Roth GmbH + Co. KG, Karlsruhe, Germany) were used. Each run was conducted with an injection volume of 50 µL. Elution was performed with 1 mL/min at 38 • C using the following gradient: 0-3 min 58A:42B, 3-8 min 58A:42B → 95A:5B, 8-13 min 95A:5B, 13-15 min 95A:5B → 58A:42B, 15-18 min 58A:42B. A calibration curve was established with celastrol (abcr GmbH, Karlsruhe, Germany) as reference substance in a concentration range from 0.5 to 250 µg/mL. Correlation coefficient displayed linearity with a value of 0.99997. The limit of detection (LOD) was 231.5 ng/mL, and the limit of quantification (LOQ) 241.7 ng/mL. One tablet containing 10 mg TWG was grinded, 1.5 mL acetonitrile were added and treated by ultrasonication. After centrifugation at 14,500 rpm for 5 min, the supernatant was analyzed by RP-HPLC. The amount of celastrol in the tablets containing 10 mg of TWG was determined at 15.8 µg or 0.158% based on the amount of extract, respectively. A representative chromatogram of TWG including detection of celastrol is shown in Supplementary Figure S1.

Isolation of Cells from Human Blood
For cell isolation, leukocyte concentrates from freshly withdrawn blood (16 IU heparin/mL blood) from healthy adult volunteers were obtained from the Department of Transfusion Medicine at the University Hospital of Jena, Germany. The experimental protocols were approved by the local ethical committee and were performed in accordance with guidelines and regulations; informed consent was obtained. Peripheral blood mononuclear cells (PBMC) were separated using dextran sedimentation, followed by centrifugation on lymphocyte separation medium (C-44010, Promocell, Heidelberg, Germany). PBMC were seeded in RPMI 1640 (Thermo Fisher Scientific, Schwerte, Germany) containing 10% (v/v) heat-inactivated fetal calf serum (FCS), 100 U/mL penicillin, and 100 mg/mL streptomycin in cell culture flasks (Greiner Bio-one, Frickenhausen, Germany) for 1.5 h at 37 • C and 5% CO 2 for adherence of monocytes. Differentiation of monocytes to macrophages and polarization towards M1-monocyte-derived macrophages (MDM) and M2-MDM was performed as recently described [27]. Briefly, M0 GM-CSF and M0 M-CSF were generated by incubating monocytes with 20 ng/mL GM-CSF or M-CSF (Cell Guidance Systems Ltd., Cambridge, UK), respectively, for 6 days in RPMI 1640 supplemented with 10% FCS, 2 mmol/L L-glutamine and 100 U/mL penicillin-streptomycin. Then, 100 ng/mL LPS and 20 ng/mL IFNγ (Peprotech, Hamburg, Germany) were added to M0 GM-CSF for 48 h to obtain M1-MDM, while 20 ng/mL IL-4 (Peprotech) were added to M0 M-CSF for 48 h to obtain M2-MDM. To obtain polymorphonuclear leukocytes (PMNL), contaminating erythrocytes of pelleted PMNL after the density centrifugation step were removed by hypotonic lysis using water. PMNL were washed twice in ice-cold PBS and finally resuspended in PBS pH 7.4 containing 1 mg/mL glucose and then incubated as described below for the analysis of 5-LOX product formation. Platelet-rich plasma was obtained from the supernatants after density gradient centrifugation, mixed with PBS pH 5.9 (3:2 v/v), centrifuged (2100× g, 15 min, room temperature), and the pelleted platelets were resuspended in PBS pH 5.9/0.9% NaCl (1:1, v/v). Washed platelets were finally resuspended in PBS pH 7.4 and 1 mM CaCl 2 .

Determination of LM Formation in Human MDM
M2-MDM (2 × 10 6 /mL) were incubated with vehicle (0.1% DMSO) and TWG (3, 30 µg/mL) in PBS containing 1 mM CaCl 2 for 180 min at 37 • C. The reaction was stopped by transferring supernatants (1 mL) into 2 mL ice-cold MeOH. In another set of experiments, with the scope to elucidate the effect of TWG on challenged M1-MDM, cells (2 × 10 6 /mL) were pre-treated with TWG (3, 30 µg/mL) in PBS containing 1 mM CaCl 2 for 15 min prior to stimulation with 1% Staphylococcus aureus 6850 wt-conditioned medium (SACM) for another 90 min. Cultivation of S. aureus and preparation of SACM was performed as previously described [28]. The vehicle group received 0.1% DMSO. The reaction was stopped by transferring supernatants (1 mL) into 2 mL ice-cold MeOH.

Lipid Mediator Metabololipidomics by UPLC-MS-MS
Samples obtained from incubated MDM were kept at −20 • C for at least 60 min to allow protein precipitation. After centrifugation (1200× g, 4 • C, 10 min), acidified H 2 O (8 mL, final pH = 3.5) was added and samples were subjected to solid phase cartridges (Sep-Pak ® Vac 6cc 500 mg/6 mL C18; Waters, Milford, MA, USA). The columns had been equilibrated with 6 mL methanol and 2 mL H 2 O before sample loading. After washing with 6 mL H 2 O and then with 6 mL n-hexane, LM were eluted with 6 mL methyl formate. The samples were brought to dryness using an evaporation system (TurboVap LV, Biotage, Uppsala, Sweden) and resuspended in 100 µL methanol/water (50/50, v/v) for UPLC-MS-MS analysis. LM were analyzed with an Acquity™ UPLC system (Waters, Milford, MA, USA) and a QTRAP 5500 Mass Spectrometer (ABSciex, Darmstadt, Germany) equipped with a Turbo V™ Source and electrospray ionization. LM were eluted using an ACQUITY UPLC ® BEH C18 column (1.7 µm, 2.1 mm × 100 mm; Waters, Eschborn, Germany) at 50 • C with a flow rate of 0.3 mL/min and a mobile phase consisting of methanol-water-acetic acid of 42:58:0.01 (v/v/v) that was ramped to 86:14:0.01 (v/v/v) over 12.5 min and then to 98:2:0.01 (v/v/v) for 3 min [29]. The QTRAP 5500 was operated in negative ionization mode using scheduled multiple reaction monitoring (MRM) coupled with informationdependent acquisition. The scheduled MRM window was 60 s, optimized LM parameters were adopted [29], and the curtain gas pressure was set to 35 psi. The retention time and at least six diagnostic ions for each LM were confirmed by means of an external standard (Cayman Chemical/Biomol GmbH, Hamburg, Germany). Quantification was achieved by calibration curves for each LM. Linear calibration curves were obtained for each LM and gave r 2 values of 0.998 or higher. Additionally, the limit of detection for each targeted LM was determined [29].
Purified 5-LOX (0.5 µg) in PBS pH 7.4 containing EDTA (1 mM) and ATP (1 mM) was pre-incubated with vehicle (DMSO 0.1%), TWG (0.1, 0.3, 1, 3, 10 µg/mL) or zileuton (3 µM) for 15 min at 4 • C. Then, samples were pre-warmed for 30 s at 37 • C, and 2 mM CaCl 2 plus 20 µM AA were added to start 5-LOX product formation. The reaction was stopped after 10 min by addition of one volume of ice-cold methanol, and the formed metabolites were analyzed by RP-HPLC. For the extraction of 5-LOX products, 200 ng of internal PGB 1 standard was added to each sample and samples were centrifuged at 2000 rpm for 10 min. 5-LOX products were purified by solid phase extraction. After elution with 300 µL methanol, samples were analyzed for all-trans isomers of LTB 4 and 5-H(p)ETE by RP-HPLC using a C-18 Radial-PAK column (Waters, Eschborn, Germany) as previously reported [31].

Determination of 5-LOX Activity in Human PMNL
Freshly isolated PMNL (5 × 10 6 ) were pre-incubated with vehicle (DMSO, 0.1%), TWG (0.1, 0.3, 1, 3, 10 µg/mL), or zileuton (3 µM) for 10 min at 4 • C. Then, 2.5 µM A23187 was added together with 20 µM AA for another 10 min at 37 • C in order to induce 5-LOX product formation. The reaction was stopped by adding 1 mL of ice-cold MeOH to the cell suspension on ice. For the extraction of 5-LOX products (all-trans isomers of LTB 4 , LTB 4 and 5-H(p)ETE), 200 ng of internal PGB 1 standard was added, samples were centrifuged at 2000× g for 10 min, and solid phase extraction and analysis by RP-HPLC was performed as described above.

Determination of mPGES-1 Activity in a Cell-Free Assay
Microsomes of A549 cells stimulated by IL-1β (2 ng/mL for 48 h) were used as source for microsomal prostaglandin E 2 synthase (mPGES)-1, as described elsewhere [32]. In brief, A549 cells were incubated with ice-cold homogenization buffer (0.1 M potassium phos-phate buffer pH 7.4, 1 mM phenylmethanesulphonyl fluoride, 60 µg/mL soybean trypsin inhibitor, 1 µg/mL leupeptin, 2.5 mM glutathione and 250 mM sucrose). After sonication, the lysate was first centrifuged at 10,000× g for 10 min, and then at 174,000× g for 1 h at 4 • C. The pelleted microsomal fraction was then resuspended into 1 mL of homogenization buffer and diluted in a potassium phosphate buffer (0.1 M, pH 7.4) containing 2.5 mM glutathione. Afterwards, TWG, vehicle (0.1% DMSO), or the positive control MK886 (10 µM) were added for 15 min on ice in 100 µL incubation volume. The reaction was started after addition of PGH 2 (20 µM) and stopped after 1 min at 4 • C using 100 µL of a stop solution (40 mM FeCl 3 , 80 mM citric acid, and 10 µM of 11β-PGE 2 as internal standard). PGE 2 and 11β-PGE 2 were extracted by solid phase extraction using acetonitrile as eluent and quantified by RP-HPLC.

Determination of Thromboxane A Synthase Activity
Thromboxane A synthase (TXAS) activity was determined in lysates of human platelets. Freshly isolated cells were resuspended in ice-cold PBS pH 7.4 containing 1 mM EDTA (1 × 10 8 cells/mL) and sonicated (4 × 10 s) on ice. The reaction was initiated by the addition of 20 µM PGH 2 for 1 min at 4 • C in cell homogenates (1 mL) after pre-incubation with TWG (1 and 10 µg/mL), vehicle (0.1% DMSO) or the positive control ozagrel (OZA, 50 µM) for 15 min at 4 • C and terminated by the addition of ice-cold MeOH (2 mL). Then, samples were processed as described for LM metabololipidomics analysis and formed TXB 2 was determined by UPLC-MS-MS, as reported above.

Statistical Analysis
Results are expressed as mean + S.E.M. of each independent experiment, where n represents the indicated numbers from separate donors performed on different days. Statistical analysis and graphs were made by using GraphPad Prism 8 software (San Diego, CA, USA). Paired t-test was used to analyze experiments for comparison of two groups; while for multiple comparisons, ANOVA with Bonferroni or Dunnett multiple comparison tests were applied as indicated. A p-value ≤ 0.05 is a criterion for statistical significance.

TWG Modulates LM Formation in Activated Pro-Inflammatory Macrophages
Human MDM were polarized with IFNγ and LPS for 48 h towards a pro-inflammatory M1-like phenotype and preincubated with TWG (3 or 30 µg/mL, as low or high dose) for 15 min prior to activation with SACM [28] to induce LM biosynthesis. LM profile signatures in the medium were analyzed by UPLC-MS-MS after 90 min of these M1-MDM incubations, and revealed substantial amounts of COX-and 5-LOX-derived LM, with only minor formation of 12/15-LOX products (Table 1), as reported before [27,28]. Analysis of the cell viability by MTT assay revealed no detrimental effects of TWG within 180 min up to 30 µg/mL ( Figure 1A). TWG at the high dose of 30 µg/mL potently suppressed formation of all 5-LOX products (LTB 4 , t-LTB 4 , 5-HETE, 5-HEPE and 7-HDHA) with minor efficiency at the low dose of 3 µg/mL (Table 1, Figure 1B,D). The sum of COX-derived products was not altered by TWG, however, PGE 2 and PGD 2 were elevated but formation of TXB 2 was strongly diminished (Table 1, Figure 1B,C), suggesting that conversion of the COX product PGH 2 by terminal prostanoid synthases is differentially affected by TWG. Of interest, TWG suppressed generation of 12-LOX products (i.e., 14-HDHA, 12-HETE and 12-HEPE) and RvD5, while 15-LOX products (i.e., 17-HDHA, 15-HETE and 15-HEPE) were rather elevated (Table 1), at least at 30 µg/mL TWG. The release of PUFA as LM substrates was not markedly affected by TWG.

TWG Selectively Inhibits 5-LOX among LM-Biosynthetic Enzymes
Intrigued by the significant and differential modulation of certain LM by TWG in activated MDM, we assessed the effects of TWG on enzymatic activities of isolated LMbiosynthetic enzymes in cell-free assays. TWG concentration-dependently inhibited the activity of human recombinant 5-LOX with IC 50 of 2.9 µg/mL (Figure 2A,B). COX-1 activity was not inhibited, and COX-2 activity was moderately suppressed at 10 µg/mL TWG, the highest concentrations tested (Figure 2A). Surprisingly, TXAS was not affected despite efficient suppression of TXB 2 formation in MDM (see Figure 1C), but mPGES-1 activity was reduced by 44% at 10 µg/mL TWG (Figure 2A). To explore 5-LOX inhibition we also assessed the effects of TWG in a robust and well-recognized cell-based assay, that is, A23187-activated human PMNL. Again, TWG concentration-dependently inhibited 5-LOX activity ( Figure 2C), like in MDM.

TWG Induces the Formation of SPM and 12/15-LOX Products in Anti-Inflammatory Macrophages
Novel recent pharmacological approaches for treatment of chronic and excessive inflammatory disorders favor the formation of SPM, besides blocking pro-inflammatory PGs and LTs [8,33]. We employed the use of anti-inflammatory M2-MDM polarized with IL-4 that possess high capacities for SPM production due to substantial expression of 15-LOX-1, with moderate formation of PGs and LTs [27]. Like for M1-MDM, cell viability analysis by MTT assay showed no detrimental effects by TWG up to 30 µg/mL within 180 min for M2-MDM ( Figure 3A). Intriguingly, exposure of M2-MDM to TWG at 30 µg/mL caused a massive formation of the SPMs PDX, RvD5 and MaR2 along with substantial formation of their monohydroxylated precursors 17-HDHA and 14-HDHA as well as other 12/15-LOX products with minor effects at 3 µg/mL TWG (Table 2, Figure 3B,C). The very low amounts of 5-LOX products in M2-MDM were not diminished by TWG (Table 2, Figure 3B). COX products were much less efficiently elevated by TWG as compared to SPM ( Figure 3C), again with differential effects depending on the individual prostanoids: PGD 2 was most potently elevated while TXB 2 was not altered (Table 2). Together, these data indicate that TWG suppresses pro-inflammatory 5-LOX products in M1-but elevates inflammation-resolving SPM and 12/15-LOX products in M2-MDM.

TWG Modulates the Expression of LM-Biosynthetic Enzymes in Macrophages during Polarization
Next, we investigated if TWG may also affect LM biosynthetic pathways during polarization of the MDM with consequence for LM-producing capacities of the cells. MDM were pre-treated with TWG for 15 min, polarized for 48 h towards M1-MDM using IFNγ plus LPS or towards M2-MDM using IL-4, harvested, and then activated with SACM to elicit LM production. Analysis by MTT assay showed that during 48 h exposure to TWG, both M1-and M2-MDM were susceptible to TWG ≥ 3 µg/mL with loss of viability ( Figure 4A), and thus we applied lower TWG concentrations of only 1 µg/mL. As shown in Table 3 and Figure 4B, 5-LOX and COX products formed by M1-MDM were significantly impaired when cells had been pre-treated with TWG, and also 15-LOX products were diminished, while 12-LOX products remained elevated; release of PUFA was not markedly altered. In M2-MDM, a significant but moderate reduction of 5-LOX products was observed without significant alterations of other LM (Table 3, Figure 4C). We then assessed if TWG affects the amounts of LM-biosynthetic enzymes during polarization; dexamethasone (DEX, 100 nM) was used as reference drug. Western blot analysis with M1-MDM showed no significant changes of cPLA 2 , 5-LOX, COX-1 and COX-2 protein amounts by TWG, but mPGES-1 was significantly reduced; DEX suppressed COX-2 protein levels as expected without affecting other enzymes addressed ( Figure 5A,B). In M2-MDM, TWG impaired the amounts of 5-LOX protein (yet no statistical significance was reached) without affecting 15-LOX-1 levels, while DEX caused the opposite: it slightly decreased the protein amounts of 15-LOX-1 without affecting those of 5-LOX ( Figure 5C,D).

TWG Suppresses Pro-Inflammatory Cytokines and Impacts Macrophage Polarization
Since TWG suppressed pro-inflammatory COX and 5-LOX pathways during polarization of M1-MDM, we studied if also pro-inflammatory cytokines are affected by TWG. In fact, TWG (1 µg/mL) blocked release of TNF-α and IL-1β in M0 M-CSF MDM, similar as the reference drug DEX ( Figure 6A). Finally, we investigated if TWG may affect macrophage polarization by assessing CD54 and CD80 as M1 markers and CD163 and CD206 as markers for M2 macrophages [27]. In agreement with the impaired capacities to generate pro-inflammatory LM (i.e., COX/5-LOX products) and cytokines (TNF-α and IL-1β) but increasing anti-inflammatory SPM, the markers for M1-MDM, i.e., CD54 and CD80 were significantly decreased but those for M2 (CD163, CD206) were slightly increased ( Figure 6B). Together, TWG apparently impairs polarization of pro-inflammatory macrophages, characterized by suppression of pro-inflammatory cytokines and LM in M1, while promoting inflammation-resolving SPM in M2 macrophages.

Discussion
Employing a comprehensive LM metabololipidomics approach we show here that TWG beneficially modulates the biosynthesis of LM networks in various innate immune cells, that is, suppressing the formation of pro-inflammatory 5-LOX products and thromboxane in M1-MDM and PMNL, but elevating the levels of inflammation-resolving SPM in anti-inflammatory M2-MDM. Our results suggest multiple points of attack of TWG in the LM networks, such as direct inhibition of 5-LOX activity and blocking 5-LOX expression, reducing the expression and activity of mPGES-1, and intriguingly, induction of the activation of 12/15-LOXs. Since these bioactions of TWG occurred at fairly low effective concentrations (i.e., 1-30 µg/mL) that might be of pharmacological relevance, our data suggest that such beneficial switch from pro-inflammatory to pro-resolving LM production in innate immune cells may contribute to the anti-rheumatic features of TWG documented in numerous clinical trials of RA [16,18,19,21]. Future studies on how TWG modulates LM production in co-culture system that more closely recapitulate RA pathology such as human RA synovial fibroblasts or synovial cell lines with macrophages may sustain the direct link with RA.
Modulation of LM networks in innate immune cells by TWG has not been reported yet, to the best of our knowledge. Only one study was published that demonstrated suppression of PGE 2 formation due to inhibition of COX-2 expression [24]. Therefore, modulation of cellular LOX activities and thus LT and SPM formation by TWG are novel findings. We confirmed PGE 2 suppression using pro-inflammatory human M1-MDM that acquire substantial amounts of COX-2 protein during polarization [27]. When TWG was present during M1-MDM polarization, the cells showed reduced capacities to generate PGE 2 , but also other COX-derived prostanoids such as PGD 2 , PGF 2 α and TXB 2 upon subsequent stimulation. Interestingly, we discovered that besides interference with COX-2, TWG also inhibited the induction of mPGES-1 protein and slightly impaired its enzymatic activity, which may sustain the reduced PGE 2 formation by TWG. Among the three PGE 2 synthases, mPGES-1 is an inducible isoform and strongly linked to inflammatory diseases [34], including RA [35].
While long-term treatment of M1-MDM with TWG clearly impaired formation of all COX-derived products possibly due to COX-2 suppression, short term exposure of these cells to TWG selectively blocked formation of only TXB 2 , suggesting an inhibitory effect on the biosynthetic branch from PGH 2 towards TXA 2 [36] potentially by acting on TXAS. Surprisingly, TWG however failed to inhibit TXAS in a cell-free assay, excluding direct interaction of TWG with the enzyme. It is reasonable that TXAS inhibition requires the intracellular environment, for example to convert the responsible bioactive ingredient(s) into the active form, a phenomenon well-known for naturally occurring quinones (like celastrol), which act as potent inhibitors of 5-LOX as reduced hydroquinones [37].
Our study reveals 5-LOX as a direct target of TWG by using cell-free assays and inhibition of 5-LOX products in M1-MDM and in PMNL; 5-LOX and its products, especially LTB 4 , have been implicated in RA as well [6,11], playing essential roles in the induction of pain and bone damage [38]. Also, 5-LOX was strongly expressed in lining and sublining macrophages, neutrophils and mast cells of RA synovial biopsies which was suppressed by glucocorticoid treatment [39]. Results of several studies underline the crucial role for LTB 4 and its receptor BLT1 in the pathogenesis of inflammatory arthritis [40]. Therefore, the potent impairment of LT formation in pro-inflammatory M1-MDM and PMNL due to 5-LOX inhibition may reasonably contribute to amelioration of RA by TWG treatment. This is also supported by our recent finding that the pentacyclic triterpenoid quinone methide celastrol, as major bioactive TWG constituent [23], potently inhibits 5-LOX in cell-free and cell-based assays at 0.1 to 1 µM [26]. Our RP-HPLC analysis and calculation revealed a celastrol content of 0.158% in TWG, implying that at the IC 50 of 2.9 µg/mL TWG for 5-LOX, 0.0044 µg/mL or~0.1 µM celastrol is present, which fits to the reported IC 50 of 0.19 µM for celastrol under the same 5-LOX assay conditions [26].
Although our results are in favor of beneficial properties of TWG for treatment of RA, the well-known toxicity of TWG observed in clinical trials is still a concern [18,19]. We have carefully considered the issue of potential cytotoxicity and thus avoided the use of high concentrations of TWG, i.e., ≤30 µg/mL in short-term and 1 µg/mL in long-term incubations, along with exclusion of cytotoxic effects under the experimental conditions that we employed.
We suggest that promoting a switch from pro-inflammatory LTs and PGs towards inflammation-resolving SPM by smart manipulation of LM networks using TWG might a beneficial strategy for RA treatment. In contrast to LTs and PGs that initialize and maintain persistent and excessive inflammation-promoting various inflammatory pathologies [1,2], the SPM are anti-inflammatory immunoresolvents and promote the resolution of inflammation leading to tissue repair and return to homeostasis [3,9,41]. Recent studies indicated that RA may arise from a decreased ability of the immune response to engage resolution programs that terminate inflammation and prevent chronicity [15]. Experimental models of delayed or non-resolving joint inflammation showed that SPM, i.e., RvD3, were downregulated [42]. In arthritic patients, synovial levels of RvE2 correlated with decreased joint pain [43]. Recently, SPM levels in peripheral blood of RA patients were linked with both DMARD responsiveness and disease pathotype [44] and strategies to increase SPM production have been shown to be connected with decreased joint inflammation and promotion of joint protection [45]. Intriguingly, exposure of M2-MDM to 30 µg/mL TWG caused massive induction of 12-/15-LOX product formation including the biosynthesis of SPM. Note that in contrast to reduced capacities of MDM to generate COX and 5-LOX products upon long-term treatment with TWG, the remedy did not impair the ability to form 12/15-LOX-derived LM. How such potent 12-/15-LOX activation is induced by TWG remains to be investigated but might be again caused by celastrol that induced SPM formation in M2-MDM at 1 µM [26], the calculated celastrol concentration present in 30 µg/mL TWG.
Besides celastrol, other bioactive ingredient(s) contained in TWG might be responsible for the observed actions on LM pathways, especially related to the suppression of the expression of the LM biosynthetic enzymes COX-2, mPGES-1 and 5-LOX. The diterpenoid triepoxide triptolide is another major constituent of TwHF with glucocorticoid-like properties that mediates many of the pharmacological actions and the anti-rheumatic activity of TWG [23]. Studies on celastrol and triptolide with focus on modulation of expression of LM pathways under long-term conditions are currently ongoing in our laboratory.
Taken together, TWG causes beneficial modulation of LM biosynthesis in prime innate immune cells by suppressing pro-inflammatory PG and LT pathways via multiple points of attack and by promoting SPM formation. In view of the well-established detrimental impact of PG and LT in RA pathology on one hand and the beneficial features of SPM on the other, our findings may help to explain the ameliorating effects of TWG in RA treatment and encourage future evaluation of the efficacy and safety of such remedies.

Data Availability Statement:
The data presented in this study are available on reasonable request from the corresponding author. The data are not publicly available due to privacy.