Materials
Baltic amber teething necklaces (33 cm Prestige Necklace, Little Smiles Genuine Baltic Amber Beads, Narangba, AUS) were purchased from a baby and toddler supplies store in Brisbane, Queensland. Individual beads were stratified by colour into light, medium and dark categories, based on some claims that lighter coloured beads contained more of the active ingredient, succinic acid [16]. Average bead weight was obtained by individually weighing 85 beads (24 light, 35 medium, 26 dark). Succinic acid, paracetamol, ibuprofen, hydrocortisone and phorbol myristate acetate (PMA) were purchased from Sigma Aldrich (St Louis, USA). Lipopolysaccharide (LPS) from E. coli was purchased from Enzo Life Sciences (Farmingdale, USA).
Infrared spectroscopy
Beads contained in a sealed plastic bag were crushed into a fine powder using a hammer. The powder was analysed using an infrared spectrometer (Shimadzu IRPrestige-21 FTIR) using 20 scans for transmittance at the highest resolution setting, and transmittance was analysed using IRsolution 1.3 software (Shimadzu). The spectra were compared with published data for Baltic amber [17] after adaptation using the WebPlotDigitizer v3.21 tool.
Succinic acid content determination
The melting point of powdered beads was 360 °C, as determined using Crown Scientific Melting Point Apparatus. As succinic acid melts at 185 °C and boils at 235 °C, an alternative method was needed in order to separate succinic acid from the amber without having to melt the beads using extremely high temperatures. Therefore, a method was developed to dissolve the beads. The beads were immersed in a range of solvents, including octanol, methanol, hydrochloric acid, or hexane. Sulfuric acid was found to affect the beads, so six beads (comprised of two beads from each colour group) were dissolved in 20 mL of concentrated H2SO4 for 16 h. 5 mL aliquots (3 replicates) of the bead-acid solutions were then neutralized using 11 mL of 18 M NaOH for analysis by HPLC for succinic acid content. The peak area of the dissolved beads was compared with a standard curve of succinic acid prepared in sulfuric acid and neutralised in the same way.
Succinic acid release assay
To investigate if the beads would release succinic acid under various conditions, 22 beads of each colour were submerged in 10 mL of either phosphate buffered saline (PBS) at pH 5.5 (approximate pH of human skin) or octanol (organic phase to mimic hydrophobic conditions of human skin layers). Beads were left in these solutions for up to 24 weeks at 37 °C, and samples of supernatant were collected after 4, 8, 12, 16, 20 and 24 weeks and analysed by HPLC for succinic acid content. Succinic acid was quantified using a standard curve of succinic acid dissolved in PBS (0.005–10 mg/mL) or octanol (0.005–0.1 mg/mL).
HPLC
The samples were filtered using 0.45 μm PVDF membrane filters. Following filtration, samples (50 μL) were injected on a Shimadzu HPLC using a 250 × 4.6 mm Vydac Denali C18 column (Grace Davison Discovery Sciences, Illinois, US) fitted with a guard column. The mobile phase was isocratic sulphate buffer comprised of 1 mM sulfuric acid and 8 mM sodium sulfate with pH 2.7, run at 1 mL/min for 20 min. A diode array detector (Shimadzu Prominence SPD-M20A) recorded absorbance across the range 200–800 nm with concentrations calculated using the area under the curve at 210 nm. This method for quantification was validated for linearity, specificity, sensitivity (limit of detection (LOD) and quantitation (LOQ)), and inter-day repeatability in accordance to the International Conference on Harmonization guidelines [18]. A 5-point calibration curve was constructed with standard solutions in the range of 0.005–10 mg/mL in PBS (pH 5.5). This was prepared in triplicate with the linearity assessed using linear regression analysis by least squares. Specificity was examined by contrasting the wavelength of the samples and standards across a range of 200–400 nm to determine whether bead excipients or other contributing factors affected the maximum absorbance wavelength. Sensitivity was determined by calculating the LOD = 3.3σ/S, and the LOQ = 10σ/S, with σ = standard deviation of the response and S = slope of the calibration curve. Intermediate precision (inter-day repeatability) was performed by calculating the % of relative standard deviation (%RSD) of 7 measurements for standard solutions at 0.1 mg/mL in PBS or octanol conducted over a period of 6 months.
Assessment of anti-inflammatory activity
The potential anti-inflammatory properties of succinic acid were tested using the THP-1 human monocyte cell line (ATCC No TIB-202) [19]. THP-1 cells (2 × 105) were differentiated into macrophages by stimulation with 50 nM PMA [19], before being treated with various control drugs (paracetamol, ibuprofen or hydrocortisone) or with succinic acid. Paracetamol and ibuprofen were selected due to their common usage in treatment of teething symptoms in children [20], and hydrocortisone was selected as a positive control due to its well-known anti-inflammatory effects and mechanisms [21]. Concentrations of the drugs were based on published observations of IC50 values in the inhibition of COX-1 and COX-2: hydrocortisone ~ 5 μM, ibuprofen ~ 8 μM, paracetamol ~ 25 μM [22, 23]. Several concentrations above and below these IC50 values were selected to capture likely effects of each drug on inflammation. After 2 h of incubation with respective drugs or media only, cells were then stimulated by adding LPS (1 μg/mL) from E. coli to cause release of prostaglandins and inflammatory cytokines [19]. For quantification of inflammatory cytokines, supernatants were collected after 4 h of LPS stimulation; for quantification of prostaglandins, supernatants were collected after 20 h of LPS stimulation. Cytokines and prostaglandins were then quantified by enzyme-linked immunosorbent assay (ELISA), with three independent experiments performed for each assay.
Cytokine ELISA
In order to quantify cytokine release by activated THP-1 cells, ELISA was performed for the pro-inflammatory cytokines IL-1α, IL-1β, IL-8 and TNFα using kits from BioLegend (Thermo Fisher Scientific; San Diego, USA). All assays were performed following the manufacturer’s instructions. Briefly, ELISA plates were coated in primary anti-cytokine antibody overnight and washed twice. Culture supernatant samples or cytokine standards were added to the plates and incubated for 3 h on a plate shaker and washed twice. An enzyme-conjugated secondary antibody was then added to the plates and incubated for 2 h on a plate shaker and washed twice. Tetramethylbenzidine (TMB) substrate was then added to plates and left to develop for 15 min before the reaction was stopped by addition of 1 M H2SO4. Optical density was immediately read using an ELISA plate reader (iMark Microplate Reader, Bio-Rad) at 450 nm.
Prostaglandin ELISA
In order to quantify prostaglandin release by activated THP-1 cells, ELISA was performed for PGE2 using a kit from Enzo Life Sciences according to the manufacturer’s instructions. Briefly, PGE2 standards and culture supernatants were incubated in wells of a pre-coated plate in the presence of PGE2-enzyme conjugate and anti-PGE2 antibody for 2 h on a plate shaker and washed three times. Paranitrophenylphosphate (pNpp) substrate was then incubated for 45 min before the reaction was stopped by addition of trisodium phosphate stop solution. Optical density was immediately read using an ELISA plate reader at 405 nm.
Statistical analysis
Standard curves from HLPC analysis to quantify succinic acid were generated by linear regression, and standard curves from ELISA analysis to quantify cytokines were generated by 4-parameter logistic regressions. Each ELISA test was performed three times in independent experiments; for each experiment data were normalised against the LPS-only control, such that these control cells were 100%, thus a value of 200% indicates twice as much as the control and a value of 50% indicates half as much as the control. Normalised ELISA data were analysed by ANOVA using a non-parametric model with Holm-Sidak post-hoc comparisons. A cut-off alpha level of p = 0.05 was selected for determining statistical significance, and Prism v7 (Graphpad software, San Diego, CA) was used for all analyses.