Degradation kinetics of Andrographolide in aqueous solution, product identification and biological activity evaluation

The thermal stability of 1 was assessed kinetically in aqueous solutions at different pH levels relevant to the formulation of commercial products. Andrographolide (1) is the major bioactive metabolite of A. paniculata and is thus a driver for the development of value-added products, either as a single material or as an ingredient in a standardized mixture. The use of standardized preparations of A. paniculata leaves or leaf extracts as functional ingredients for medicinal plant products is also important for commercial development. Knowledge regarding the thermal stability of 1 is therefore of significance for the development of high-quality functional products for the cosmetic and nutraceutical industries. Such standardized extracts and cosmetic products may be exposed to elevated temperatures at different pH levels, depending on the production processes, and may experience a variety of formulation, storage, and manufacturing operations where metabolite integrity could be compromised^13,27,28.

Identification of degradation products

The HPLC method for the analysis of 1 was adapted from the American Herbal Pharmacopoeia²⁹. The separation of 1 and its major degradation products from the pH-adjusted solutions in methanol (MeOH) was accomplished within a 25 min time frame. The isolates were characterized using nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry and comparison with published spectroscopic data (Supplementary Tables S3 and S4, Figures S1−S10). In the acidic environment (pH 2.0), the degradation products 2^{30,31,32,33,34,35} and 3^24,36 eluted at retention times (t_R) of 4.1 min and 7.6 min, respectively (Fig. 1). Isoandrographolide (2), a white amorphous powder, possessed a molecular formula of C₂₀H₃₀O₅ as determined by (+)-LC-MS-QTOF [M + Na]⁺ at m/z 373.1976 (Supplementary Figure S11). The ¹H- and ¹³C-NMR spectra of 2 were similar to those of 1, except for a methyl group at C-17 (δ_H 1.19) and an oxymethine proton at C-12 (δ_H 4.58), instead of an exo-methylene group H-7α (δ_H 4.67) and H-7β (δ_H 4.89) and an olefinic proton H-12 (δ_H 6.85) in 1, respectively (Supplementary Table S3). The ¹³C NMR data of 2 showed two oxymethine carbons at C-8 (δ_C 83.2) and C-12 (δ_C 72.3) instead of the two resonances of olefinic carbons in 1 (Supplementary Table S4). The NMR data of 2 agree with those reported³⁵. 8,9-Didehydroandrographolide (3), white amorphous powder, C₂₀H₃₀O₅, showing (–)-LC-MS-QTOF at m/z 395.2069 ([M + HCOO]^– (Supplementary Figure S12). The ¹H- and ¹³C-NMR data were closely related to those of 1, except for an endo-olefinic group at C-8 (δ_C 129.2), C-9 (δ_C 134.9), and C-17 (δ_C 20.9) in 3, instead of an exo-methylene group in 1 (Supplementary Table S4). The NMR spectral data of 3 agreed with those previously reported²⁴.

For the pH 8.0 solution, the degradation products 4^3,24,37,38, 5³⁹, and 6^19,20,24 eluted at t_R of 5.5, 10.2, and 19.8 min, respectively. Compound 6 was not detected in strongly basic conditions (pH 10.0–12.0) (Fig. 1). 15-Seco-andrographolide (4) was isolated as a white amorphous powder, C₂₀H₃₂O₆, m/z 367.2123 ([M – H]^–) (Supplementary Figure S13). The structure of 4 was confirmed by comparison of the NMR data with 1, through the presence of an acyclic oxy-methylene group at C-15 (δ_H 3.75 and δ_H 3.61; δ_C 66.3) instead of the cyclic oxy-methylene protons (δ_H 4.16 and δ_H 4.46; δ_C 76.1) in 1 (Supplementary Table S3). The NMR data of 4 agreed with those reported²⁴. 14-Deoxy-15-methoxyandrographolide (5) was obtained as a white solid, C₂₁H₃₂O₅, m/z 365.2328 ([M + H]⁺) (Supplementary Figure S14). The NMR data were similar to those of 1, except for resonances for a methylene at C-14 (δ_H 3.09, 2.67; δ_C 33.5) and di-oxymethine at C-15 (δ_H 5.55; δ_C 104.4), instead of resonances for an oxymethine and oxymethylene in 1. The structure of 5 was confirmed by comparison of the NMR data with those reported³⁹. The structure of 11,14-dehydro-14-deoxy-andrographolide (6) was confirmed by analysis of the ¹H-NMR data with those reported²⁴. A trans-olefinic group was observed at C-11 (δ_H 6.86; δ_C 136.54) and C-12 (δ_H 6.16; δ_C 122.5). and by the resonances for an olefinic group at C-13 (δ_C 129.6) and C-14 (δ_H 7.44; δ_C 146.7) instead of an oxy methine in 1. The (+)-LC-MS-QTOF at m/z 333.2060 ([M + H]⁺) confirmed the molecular formula to be C₂₀H₂₈O₅ (Supplementary Figure S15).

The HPLC chromatograms illustrating the degradation products formed at pH 2.0, pH 6.0, and pH 8.0 are presented in Fig. 1 and Supplementary Figures S16–S18. In experiments using DMSO, the degradation product observed under acidic conditions (pH 2.0) produced similar signals as in MeOH, whereas under strongly basic conditions (pH 10.0 and pH 12.0), only 4 was formed from 1 (Supplementary Figure S19). Polarity or nucleophilicity therefore plays an important role in the degradation pathways of 1 since MeOH is a stronger nucleophile than DMSO and thus 5 could be observed in the MeOH conditions.

The persistence of the degradation products from 1 in pH 2.0 solution was evident with compounds detected for at least 7 days (Supplementary Figure S16), whereas at pH 6.0 and pH 8.0 4 was the major degradation product, and 5 and 6 were observed in small amounts under these conditions (Fig. 1 and Supplementary Figures S17 and S18). The formation of products 4, 5, and 6 was initially detected after 2 days at pH 6.0, and after 1 h at pH 8.0 (Supplementary Figures S17 and S18). These studies support the previous reports on the kinetic degradation of 1 in solution²⁴. Andrographolide (1) therefore undergoes distinctive acid and base-catalyzed degradation pathways which provides a critical insight into the degradation mechanisms of 1 and highlights the importance of pH on the stability and transformation of this bioactive compound when different pharmaceutical or cosmeceutical formulations are being considered.

Kinetics of the degradation of 1

HPLC chromatographic analysis revealed a direct relationship between the pH of the solution and the rate of the degradation reaction. Specifically, at pH 2.0 and pH 4.0, the degradation of 1 occurred at a slower rate compared to solutions at higher pH levels (pH 6.0 to pH 12.0) (Figs. 2 and 3). The HPLC chromatogram of 1 at pH 2.0 over a period of 7 days at 70 °C indicated two degradation products, 2 and 3 (Supplementary Figure S16), while no degradation was detected at pH 4.0 in the same time frame and temperature (Supplementary Figure S20). Thus, 1 exhibited greater stability in pH 4.0 buffer solution compared to pH 2.0, consistent with previous findings that noted the optimal stability of 1 within the pH range of 3–5⁴⁰. Although the degradation rates increase below pH 3.0, 1 remains largely present at pH 2.0, a typical human gastric value, and making this low pH relevant in the context of food, cosmetic, and drug processing.

The metabolite remained stable at the boiling point of MeOH (64.7 °C for 28 days), whereas in DMSO, the degradation rate of 1 depended on the temperature and the composition of the solvent²³. The rate constant for the degradation of 1 was investigated at pH 2.0, pH 6.0, and pH 8.0, and at three individualized, elevated temperatures of 70, 77, and 85 °C (Table 1). Chemical kinetic parameters and profiles for the degradation of 1 are shown in Table 2. As expected, in each case the apparent kinetic rate constant (k) increased with increasing temperature. Strong correlation coefficients (0.9800 < r² < 0.9978) from the plot of ln (C) against reaction time (day) were found (Table 1). The rate constant obtained from Eq. (1) was fitted to an Arrhenius-type equation in each kinetic model studied to determine the effect of temperature on the chemical reaction (Fig. 2). The k values indicated the decreased thermal stability of 1 as the temperature was increased. Under different conditions, the k value of solid-state andrographolide under heat-accelerating conditions was reported to be 3.8 × 10^− 6 per day¹⁷ and 6.58 × 10^− 6 per day¹⁸ while in pH-dependent solutions the k value was revealed as 6.5 × 10^− 5 per day (at pH 2.0), 2.5 × 10^− 3 per day (at pH 6.0), and 9.9 × 10^− 2 per day (at pH 8.0). Thus 1 decomposed faster in acid and basic solutions than in the solid state. This is the first report documenting the thermal degradation kinetics of 1 at specific pH conditions. These results align with earlier studies that characterized the degradation of 1 as following a first-order reaction model in solution through intermolecular interactions²³. First-order kinetics indicated that the degradation of 1 is concentration dependent, therefore the amount of 1 degrading per unit of time is not constant for the ambient pH conditions. Most drugs tend to degrade with either zero- or first-order kinetics⁴¹. On the other hand, solid state 1 decomposed through second-order degradation kinetics under accelerated conditions through intramolecular interactions^19,20. Second-order degradation occurs when the rate is influenced by the concentration of two separate or identical reactants. Therefore, the varied matrix or formulation of 1 under study may be the cause of the identified differences in degradation kinetics. The intermolecular interactions of 1 influence the rate of degradation^17,18,19,20. However, the intramolecular interactions of 1, particularly through hydrogen bonds and hydrophobic interactions are crucial for stability and function. The calculated activation energies (E_a) derived from the curves showed that the values for 1 at pH 2.0, pH 6.0, and pH 8.0 were 118.9, 82.8, and 79.4 kJ/mol¹, respectively (Table 2). A high E_a generally indicates that the reaction is less sensitive to temperature fluctuations¹⁵. Hence, the higher E_a value of 1 at pH 2.0 portends a lower sensitivity to temperature-induced degradation than in pH 6.0 and pH 8.0 solutions. The stability of 1 in acidic conditions may be due to the presence of free –OH groups causing clustering, followed by strong intermolecular hydrogen bonding. The shelf-life values for 1 at pH 2.0, pH 6.0, and pH 8.0 at 25 °C were 4.3 years, 41 days, and 1.1 days, respectively (Table 2) which is highly significant for product formulation studies.

First-order plot and Arrhenius plots of the degradation of andrographolide based on pH, (A) pH 2.0, (B) pH 6.0, and (C) pH 8.0.

First-order plot of the degradation of andrographolide (1) in MeOH between pH 2.0 and pH 4.0 at 70 °C for 35 days.

Formation of degradation products

The transformation of 1 into 2 can occur through (a) allylic rearrangement of the hydroxyl group on the lactone ring, (b) protonation of the exo-methylene group, and (c) cyclization of the tetrahydrofuran ring³⁰. The formation of 3 was visualized to occur through isomerization of the 8,17-double bond by (a) protonation of the exo-methylene moiety, and (b) abstraction of the proton at C-9[22] (Scheme 1). However, the¹³C-NMR spectrum indicated a mixture of formation olefinic isomers of 3 for which the precise configuration could not be established. Compound 4 could be produced through fragmentation of the lactone ring of 1, and 6 could arise through E2 elimination by base following abstraction of the δ-proton in 1 leading to 1,4-elimination from an allylic alcohol²⁴ (Scheme 1). The formation of 5 may occur through methoxylation at C-15 involving an enol lactone intermediate (m/z 333.2060)⁴² which may be derived from 6 (Scheme 1 and Supplementary Figure S21). However, the¹H-NMR spectrum indicated either a mixture of C-15-epimers of 5, or only one epimer for which the precise configuration could not be established. This mechanistic possibility was revealed from the HPLC analysis of 5 when 6 was treated at pH 8.0 in MeOH solution (Supplementary Figure S22). This chemical reaction supports the biogenetic pathway for 5 that is proposed to occur in plants³⁹. However, 5 could also be an artifact due to the presence of MeOH in the solution which acts as a nucleophile to react with an enol lactone intermediate (Figure 4 and Scheme 1). This analysis enumerates some of the mechanisms underlying the degradation of 1 and highlights the potential for exploring reaction opportunities towards further analogues for biological assessment while retaining their stability.

Chemical structures of 1 and its degradation products.

Reaction mechanisms for the degradation of 1 under acidic and basic conditions.

Biological activity assessments

To investigate the changes in the biological profile of the degradation products in comparison with 1, two bioassays were performed. The first assay evaluated the inhibitory effect on lipopolysaccharide (LPS)-induced nitric oxide (NO) production in RAW264.7 macrophages. The results of this anti-inflammatory bioassay for 1 and its degradation products are summarized in Table 3. The presence of a newly formed tetrahydrofuran ring and an olefinic bond in 2 did not enhance the anti-inflammatory activity when compared to 1 and the other degradation products. This affirmed that the conjugated Δ¹²⁽¹³⁾-double bond and the hydroxy group at C-14 are critical structural elements for the inhibition of NO production³³. In addition, the anti-inflammatory activity of 5 and 6 was not enhanced relative to 1, 3, and 4 through the introduction of a methoxy group at C-15 in 5, and a conjugated double bond in compound 6. These results align with molecular docking studies on the nitric oxide production inhibition activity of 1 and its derivatives⁴³. The formation of a C-8 vinylic methyl group, as in 3, and 4 where the lactone ring is opened, were less active than 1. These data confirm that, since in the degradation products of 1 the strong anti-inflammatory activity is not retained, the stability of 1 in formulated medicinal products must be monitored over time to avoid diminished efficacy for the patient. In this study it was affirmed that 1 was extensively degraded under strongly basic conditions. Therefore, alkaline products of 1, such as soaps and shampoos, can be explored for other activities, recognizing that the anti-inflammatory activity of 1 will have been lost.

The cytotoxic activities of the degradation products and 1 were assessed against the SW480 human colon cancer cell line, with the resulting IC₅₀ values presented in Table 3. Compounds 2 and 6 exhibited no activity at the tested concentrations, in agreement with earlier studies across a selection of cancer cell lines^{42,43,44,45,46,47}. In this investigation, the parent compound 1 was identified as the most cytotoxic with a modest IC₅₀ value of 4.17 µM, suggesting the important role of the allylic hydroxyl lactone moiety of 1 in imparting cytotoxic effects. Compounds with a C-8 vinylic methyl group, e.g., 3, and with the lactone ring-opened, e.g., 4, demonstrated weaker cytotoxic activity compared to 1. In summary, the structural integrity of 1 is necessary for maintaining both the anti-inflammatory and cytotoxic activities in developed products.