Analysis of Boswellic Acid Contents and Related Pharmacological Activities of Frankincense-Based Remedies That Modulate Inflammation

Extracts of frankincense, the gum resin of Boswellia species, have been extensively used in traditional folk medicine since ancient times and are still of great interest as promising anti-inflammatory remedies in Western countries. Despite their common therapeutic use and the intensive pharmacological research including studies on active ingredients, modes of action, bioavailability, pharmacokinetics, and clinical efficacy, frankincense preparations are available as nutraceuticals but have not yet approved as a drug on the market. A major issue of commercially available frankincense nutraceuticals is the striking differences in their composition and quality, especially related to the content of boswellic acids (BAs) as active ingredients, mainly due to the use of material from divergent Boswellia species but also because of different work-up and extraction procedures. Here, we assessed three frequently used frankincense-based preparations for their BA content and the interference with prominent pro-inflammatory actions and targets that have been proposed, that is, 5-lipoxygenase and leukotriene formation in human neutrophils, microsomal prostaglandin E2 synthase-1, and inflammatory cytokine secretion in human blood monocytes. Our data reveal striking differences in the pharmacological efficiencies of these preparations in inflammation-related bioassays which obviously correlate with the amounts of BAs they contain. In summary, high-quality frankincense extracts display powerful anti-inflammatory effectiveness against multiple targets which can be traced back to BAs as bioactive ingredients.


Introduction
Frankincense, the gum resin from various Boswellia species, is an ancient remedy used for centuries in Asian and African folk medicine, especially in traditional Ayurvedic medicine in India [1][2][3]. Preparations from frankincense have been externally applied for the treatment of wounds and skin disorders but also internal use has long been commonly practiced in order to cure a variety of inflammatory and infectious diseases [2]. In Western societies and modern folk medicine, frankincense preparations are among the most popular natural remedies for treatment of inflammatory disorders such as rheumatoid arthritis and osteoarthritis, inflammatory bowel diseases, multiple sclerosis, respiratory diseases (αBA, βBA) were found to be higher than the respective acetylated non-ketylated BAs (AαBA, AβBA) in "Sallaki ® Tablets" and "BOSWELLIASAN ® " compared to "H15 Ayurmedica ® " as can be seen in Table 1.

Modulation of Lipid Mediator Production in Exotoxin-Stimulated Neutrophils by Frankincense-Based Remedies
Next, we assessed how the frankincense-based preparations affect 5-LOX activity in intact human neutrophils that are major producers of leukotrienes [15], challenged with Staphylococcus aureus-conditioned medium (SACM) containing exotoxins for eliciting 5-LOX product formation [16]. Freshly isolated neutrophils from human blood were preincubated with the frankincense-based remedies for 15 min prior to addition of 1% SACM The reaction was started with 20 µM PGH 2 . After 1 min at 4 • C, the reaction was terminated using a stop solution containing FeCl 2 and 11β-PGE 2 . Formed PGE 2 was purified by SPE and analyzed using RP-HPLC. Data, expressed as percentage of control, are given as mean ± S.E.M., n=5; repeated measures ANOVA + Dunnett post hoc test with raw data; ** p < 0.01; *** p < 0.001 vs. vehicle. (b) MK-886 (10 µM) was used as positive control under the same experimental conditions as above. Paired t-test; ** p < 0.01 vs. vehicle.

Effects of Frankincense-Based Remedies on Pro-Inflammatory Cytokine Secretion
For analysis of the effects of the remedies on pro-inflammatory cytokine release [19], primary monocytes isolated from human peripheral blood were pre-incubated with the test items for 30 min and then stimulated with LPS (10 ng/mL) for 4 h to measure TNF-α or for 18 h to measure IL-1β and IL-6 release in the cell supernatant by ELISA; dexamethasone (100 nM) was used as reference drug. None of the frankincense-based remedies significantly blocked the release of IL-6 or TNF-α, while dexamethasone suppressed both responses, especially IL-6 release ( Figure 5). Of interest, "BOSWELLIASAN ® " concentrationdependently inhibited the secretion of IL-1β with 55% suppression at 50 µg/mL; "Sallaki ® Tablets" and "H15 Ayurmedica ® " were not significantly active at the highest test concentration of 50 µg/mL, and dexamethasone (100 nM) reduced IL-1β release by 90% ( Figure 5). . Effects of frankincense-based remedies on cytokine production. Freshly isolated human monocytes (1.5 × 10 6 /mL) were pre-incubated with 2, 10, or 50 µg/mL of frankincense-based remedies, 100 nM dexamethasone, or vehicle (0.1% EtOH) for 30 min. Cells were then stimulated with LPS (10 ng/mL) for 4 h (TNF-α) or 18 h (IL-6, IL-1β). Determination of cytokines in the supernatants was performed using ELISA. Data, expressed as percentage of control, are given as mean ± S.E.M., n = 3 for TNF-α and IL-6, and n = 5 for IL-1β. Statistics: Repeated measures ANOVA + Dunnett post hoc test with raw data. # p < 0.05; **,## p < 0.01 vs. vehicle.

Discussion
Here, we studied three frequently used frankincense-based remedies for their BA content and in parallel for inhibition of the key targets of BAs, namely 5-LOX and mPGES-1, as well as for suppression of pro-inflammatory LT and cytokine production in human primary leukocytes. We found considerable differences in the quality between the analyzed products, in terms of BA content as well as regarding their pharmacological efficiency. To our surprise "H15 Ayurmedica ® " revealed only trace amounts of the characteristic BAs, that is, KBA, AKBA, αBA, AαBA, βBA, and AβBA, and accordingly, the pharmacological activities in vitro were much less efficient as compared to "BOSWELLI-ASAN ® " and "Sallaki ® Tablets". In fact, these two products contain substantial amounts of BAs and display potent pharmacological effects in vitro that might be beneficial for treatment of inflammatory diseases. Thus, the observed dual inhibition of 5-LOX and mPGES-1 along with a shift of the LM class switch towards 12/15-LOX products by the extracts is of pivotal interest with high therapeutic potential, representing a pharmacological concept which is currently pursued as a novel and innovative approach in the therapy of inflammatory diseases [20][21][22].
In general, ingredients of dietary supplements are presented in the same order as the corresponding amount in the product. Since the amount of Boswellia serrata extract per capsule is not declared on the label and based on the fact that Boswellia serrata is listed as penultimate ingredient prior to titanium dioxide used as colorant for the capsule shell, Figure 5. Effects of frankincense-based remedies on cytokine production. Freshly isolated human monocytes (1.5 × 10 6 /mL) were pre-incubated with 2, 10, or 50 µg/mL of frankincense-based remedies, 100 nM dexamethasone, or vehicle (0.1% EtOH) for 30 min. Cells were then stimulated with LPS (10 ng/mL) for 4 h (TNF-α) or 18 h (IL-6, IL-1β). Determination of cytokines in the supernatants was performed using ELISA. Data, expressed as percentage of control, are given as mean ± S.E.M., n = 3 for TNF-α and IL-6, and n = 5 for IL-1β. Statistics: Repeated measures ANOVA + Dunnett post hoc test with raw data. # p < 0.05; **,## p < 0.01 vs. vehicle.

Discussion
Here, we studied three frequently used frankincense-based remedies for their BA content and in parallel for inhibition of the key targets of BAs, namely 5-LOX and mPGES-1, as well as for suppression of pro-inflammatory LT and cytokine production in human primary leukocytes. We found considerable differences in the quality between the analyzed products, in terms of BA content as well as regarding their pharmacological efficiency. To our surprise "H15 Ayurmedica ® " revealed only trace amounts of the characteristic BAs, that is, KBA, AKBA, αBA, AαBA, βBA, and AβBA, and accordingly, the pharmacological activities in vitro were much less efficient as compared to "BOSWELLIASAN ® " and "Sallaki ® Tablets". In fact, these two products contain substantial amounts of BAs and display potent pharmacological effects in vitro that might be beneficial for treatment of inflammatory diseases. Thus, the observed dual inhibition of 5-LOX and mPGES-1 along with a shift of the LM class switch towards 12/15-LOX products by the extracts is of pivotal interest with high therapeutic potential, representing a pharmacological concept which is currently pursued as a novel and innovative approach in the therapy of inflammatory diseases [20][21][22].
In general, ingredients of dietary supplements are presented in the same order as the corresponding amount in the product. Since the amount of Boswellia serrata extract per capsule is not declared on the label and based on the fact that Boswellia serrata is listed as penultimate ingredient prior to titanium dioxide used as colorant for the capsule shell, substantial amounts of Boswellia serrata extract were not expected in this product. Nevertheless, the fact that only traces of BAs could be detected fell far short of expectations. This might be attributed to inadequate extract preparation or to the application of another Boswellia species such as Boswellia frereana that contains only traces of BAs and other triterpene acids like lupeolic acid [13].
There is also a striking difference between the ratios of the individual BAs in the tested products. Hence, Boswellia serrata extracts contained in "Sallaki ® Tablets" and "BOSWELLIASAN ® " in contrast to "H15 Ayurmedica ® " are characterized by a higher amount of non-acetylated/non-ketylated BAs compared to the acetylated ones. This is clearly reflected in the ratios of the respective BAs. Thus, in both "Sallaki ® Tablets" and "BOSWELLIASAN ® " the ratios of αBA/AαBA and βBA/AβBA were comparable, being 2.66 and 2.11 for αBA/AαBA in "Sallaki ® Tablets" and "BOSWELLIASAN ® ", respectively, as well as 2.05 and 1.7 for βBA/AβBA "Sallaki ® Tablets" and "BOSWELLIASAN ® ", respectively. In addition, the total sum of the determined BAs make up around 31% of the applied Boswellia serrata extract in both products. This corresponds to the known distribution and amounts of BAs in Boswellia serrata extracts [12,23]. On the other hand, this distribution in the individual BAs could not be detected in "H15 Ayurmedica ® " raising further questions with regard to the origin of the extract applied in "H15 Ayurmedica ® ".
Last but not least, it should be noticed that the determined total sum of the analyzed BAs may correspond to the declared amount of 400 mg Boswellia serrata for "Sallaki ® Tablets" and "H15 Ayurmedica ® " or 300 mg Boswellia serrata extract for "BOSWELLIASAN ® " on the label. Note that for technical reasons, we analyzed 300 mg of the Boswellia serrata extract contained in the product "BOSWELLIASAN ® " that is actually composed of 400 mg of this extract. Finally, different analytical methods may result in different BA contents depending on whether HPLC, a titration method, or a very specific LC-MS/MS method was applied as in the present case [24][25][26][27].
BAs display a multitude of pharmacological activities [4]. 5-LOX and mPGES-1 are well-defined molecular targets that directly bind BAs and are inhibited by these triterpene acids [4,18,28,29]. mPGES-1 is an inducible enzyme in the biosynthesis of pro-inflammatory PGE 2 from COX-2-derived PGH 2 , and is proposed as an alternative drug target to COX enzymes that are blocked by NSAIDs, to effectively and safely intervene with inflammatory disorders [30,31]. In fact, lipophilic extracts of various Boswellia species blocked mPGES-1 activity [32] and the abundant β-BA efficiently inhibited PGE 2 under inflammatory conditions in vivo after oral application to rats [29]. Recently, we showed that AKBA binds an allosteric site in 5-LOX thereby shifting the regiospecificity towards a 12-lipoxygenating enzyme, which in neutrophils led to decreased LT and 5-HETE levels but increased the amounts of 12/15-LOX products and elevated anti-inflammatory specialized pro-resolving mediators (SPM) in 5-LOX-transfected HEK cells [18]. Such a pattern was also revealed in the present study for "Sallaki ® Tablets" and "BOSWELLIASAN ® " that inhibited the formation of LTs and 5-HETE in human primary neutrophils that had been challenged by physiological relevant bacterial exotoxins [16] and at the same time, shifted AA conversion towards 12-HETE, as observed for AKBA [18]. BAs were shown to inhibit constitutively activated NF-kappaB signaling by blocking IkappaB kinase activity [33], which might underly the suppressive effects on pro-inflammatory cytokine release [19]. Note that the suppression of cytokine release from LPS-stimulated monocytes was much less sensitive to the frankincense products as compared to LT formation, implying a subordinated effect as compared to interference with LM formation. Future analysis on these products may also reveal additional anti-inflammatory effects such as the observed interference with the formation of reactive oxygen species [34,35].
It is tempting to speculate that "Sallaki ® Tablets" and "BOSWELLIASAN ® " may act as LM class switch inducer also in inflammation models in vivo, that is, to lower LT and PGE 2 levels but increase the amounts of SPM. There is a current trend towards concepts of immunoresolvent therapies in inflammation in order to support resolution [36], especially by exploiting SPM rather than blocking the inflammatory process using glucocorticoids or NSAIDs that act as immunosuppressants with severe side effects [21,22]. Our findings support the application of high-quality frankincense-based remedies in inflammatory diseases to effectively push the LM class switch towards SPM in order to promote endogenous programs of inflammation resolution without immunosuppression. Future studies with appropriate experimental models and test systems will reveal if frankincense-based remedies may act in this respect and foster SPM formation.
Finally, we point out that the strong promises of frankincense-based products in the therapy of many diseases are mainly built upon experience-based medicine, such as case reports, rather than on evidence-based medicine. Thus, there are limitations with respect to efficacy in disease treatment, for example in oncology, where the therapeutic potential for cancer treatment is still speculative and well-designed clinical studies are required to validate the clinical usefulness [6,7].

Standard Preparation
Stock standard solutions of each BA were prepared by weighing into a 20 mL volumetric flask 20 mg of each BA standard and diluting it with 20 mL methanol to yield a concentration of 1 mg/mL of each BA. Mixed spike solutions were prepared by mixing the appropriate amount of each BA stock standard solution with methanol to yield spike solutions K1 (4 µg/mL), K2 (12 µg/mL), and K3 (24 µg/mL).

Sample Preparation
For "Sallaki ® Tablets" and "BOSWELLIASAN ® ", the contents of the tablets/capsules were pulverized and well mixed. An equivalent to 100 mg extract was weighed in a 50 mL centrifuge tube (Eco, PP, Roth, Art. AN78.1). In the case of the "H15 Ayurmedica ® " which is an oily substance, 1 capsule was carefully cut in half in a 50 mL centrifuge glass with a scalpel. Then, 20 mL of methanol (volumetric pipette) and two glass pearls were added to each product and shaken for 60 min at 200 rpm on a vertical shaker. Afterwards, the samples were treated in an ultrasonic bath for 30 min and centrifuged for 10 min at 2000 rpm. Four aliquots of 100 µL (or 25 µL depending on the content in the oil capsules) of the clear supernatant were then transferred into four 10 mL volumetric flasks. Three of these aliquots for each product were diluted with 1 mL of the three spike solutions K1, K2, K3, respectively, to yield three differently spiked samples. To one aliquot, no spike solution was added. Finally, 20 µL of each sample solution was injected into the chromatographic system.

LC-MS/MS Method
The BA content was determined according to a previously developed sensitive LC-MS/MS method. In brief, LC was performed on an Agilent 1200 series consisting of a gradient pump with vacuum degasser, an autosampler, and a column oven. A Hypersil BDS RP C18 column (100 × 4 mm; 3 µm; Thermo scientific) and an upstream Gemini security guard cartridge (4 × 3 mm; Phenomenex, Germany) were used for chromatography. Separation was achieved using a gradient program starting with 90% mobile phase A (methanol: water 90:10, 400 mg/L ammonium formate) and 10% mobile phase B (methanol: water 80:20, 400 mg/L ammonium formate) rising to 100% mobile phase A within 20 min. This was kept constant for 14 min before returning to the initial conditions during 1 min. The total run time was 35 min at a flow rate of 0.4 mL/min. The column oven was set to 40 • C and the autosampler was kept at room temperature. A standard addition method at three concentration levels (4 µg/mL, 12 µg/mL, and 24 µg/mL) was chosen for the quantification of BAs. The resulting plots of concentration versus peak area were linear with R 2 > 0.9984 for all BAs. The percent recovery, determined by comparing the calculated amounts of the BAs with the real amount spiked to the samples, ranged from 96.87% to 100.03% for KBA, 99.94% to 100.11% for AKBA, 99.89% to 100.56% for βBA, 98.43% to 100.87% for αBA, 99.98% to 100.01% for AβBA, and 97.70% to 101.28% for AαBA. The reproducibility in terms of relative standard deviation at each level ranged from 0.24% to 2.12% for KBA, 0.2% to 1.83% for AKBA, 0.48% to 4.39% for βBA, 0.3% to 2.72% for αBA, 0.29% to 2.64% for AβBA, and 0.27% to 2.46% for AαBA.

Induction of mPGES-1 in A549 Cells, Isolation of Microsomes, and Determination of mPGES-1 Activity
Microsomes of IL-1β-activated A549 cells were prepared and mPGES-1 activity determined as described [29]. In brief, cells were treated with IL-1β (1 ng/mL) at 37 • C and 5% CO 2 , harvested after 72 h, and frozen in liquid nitrogen. After reuptake of the cells in ice-cold homogenization buffer (0.1 M potassium phosphate 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) and incubation for 15 min, cells were sonicated on ice (3 × 20 s) and subjected to differential centrifugation at 10,000× g for 10 min and 174,000× g for 1 h at 4 • C. The microsomal fraction (pellet) was resuspended in homogenization buffer, analyzed for its protein content using a protein assay kit (Bio-Rad laboratories GmbH, Munich, Germany), and diluted in potassium phosphate buffer (0.1 M, pH 7.4) containing 2.5 mM glutathione. After pre-incubation with the test items for 15 min at 4 • C, the reaction (100 µL total volume) was initiated by the addition of 20 µM PGH 2 and terminated after 1 min by the addition of 100 µL stop solution (40 mM FeCl 2 , 80 mM citric acid, and 10 µM of 11β-PGE 2 as internal standard). PGE 2 was separated by solid phase extraction on RP-C18 material using acetonitrile (200 µL) as eluent, and analyzed by RP-HPLC (30% acetonitrile aqueous, 0.007% TFA (v/v), Nova-Pak ® C18 column, 5 × 100 mm, 4 µm particle size, flowrate 1 mL/min) with UV detection at 195 nm.

Isolation of Human Leucocytes
Leucocyte concentrates from freshly withdrawn peripheral blood of adult healthy donors were provided by the Institute of Transfusion Medicine of the University Hospital Jena, Germany. The experimental protocol was approved by the ethical committee of the University Hospital Jena. All methods were performed in accordance with the relevant guidelines and regulations. Neutrophils and monocytes were immediately isolated as described before [39]. In brief, cells were isolated by dextran sedimentation and Ficoll-Histopaque 1077-1 (Sigma-Aldrich) density centrifugation. To purify neutrophils, the remaining erythrocytes were removed by hypotonic lysis. Neutrophils were finally resuspended in PBS pH 7.4 plus 1 mg/mL glucose at the cell density of 5 × 10 6 cells /mL. Monocytes were separated from peripheral blood mononuclear cells by adherence to cell culture flasks (Greiner Bio-one, Frickenhausen, Germany) for 1 h (37 • C, 5% CO 2 ), followed by cell scraping and resuspension in RPMI 1640 supplemented with 5% fetal calf serum, 2 mmol/L L-glutamine (Biochrom/Merck, Berlin, Germany), and 100 U/mL penicillin; 100 µg/mL streptomycin (Biochrom/Merck).

Preparation of Exotoxin-Containing Staphylococcus Aureus-Conditioned Medium (SACM)
A single colony of Staphylococcus aureus (strain 6850) grown for 24 h at 37 • C on Columbia-agar plates containing 5% sheep blood (Altmann Analytic, Munich, Germany) was picked and inoculated in 50 mL of brain heart infusion broth (BHI, Sigma-Aldrich) for a further 24 h at 37 • C under shaking (250 rpm). The OD 600 was adjusted to 0.05 by dilution with BHI and 50 mL of the corresponding solution was cultivated for another 24 h. Finally, 15 mL were centrifuged for 10 min at 3350× g and sterile filtered using a 0.22 µm PVDF syringe filter (Carl Roth GmbH).

Analysis of LOX Product Formation in Human Neutrophils
Human neutrophils (5 × 10 6 /mL) were preincubated with a test item or vehicle (0.1% EtOH) for 15 min at 37 • C. The production of LOX products was induced by addition of SACM (1%) and CaCl 2 (1 mM). After 90 min at 37 • C, the reaction was stopped with 2 mL ice-cold MeOH containing deuterium-labeled internal standards d 8 -5S-HETE and d 4 -LTB 4 (200 nM each), and 10 µM d 8 -AA (Cayman Chemical/Biomol GmbH, Hamburg, Germany). Sample preparation was conducted as described previously [40]. In brief, samples were kept at −20 • C overnight to allow protein precipitation. After centrifugation (1200× g, 4 • C, 10 min), supernatants were transferred to 7 mL acidified H 2 O (pH 3.5, HCl). The LOX products in these samples were purified by solid phase extraction (SPE). Solid-phase C18 cartridges (Sep-Pak ® Vac 6cc 500 mg/6 mL C18; Waters, Milford, MA) were equilibrated with 6 mL MeOH and 6 mL H 2 O. Next, samples were loaded onto the columns. After washing with 6 mL H 2 O and additional 6 mL n-hexane, LMs were eluted with 6 mL methyl formate. Finally, 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 ultraperformance liquid chromatography-tandem mass spectrometry measurements (UPLC-MS-MS). LOX product profiling was conducted by means of 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