Comparison of the composition and antiplasmodial activity of Artemisia annua teas using an untargeted metabolomic approach

Abstract

The use of teas made from locally cultivated Artemisia annua to fight malaria in remote areas where access to care is difficult is a matter of debate. This study aimed at document differences in the composition of A. annua teas cultivated in Benin to be sold as antimalarial teas, and in France, and their impact on antiplasmodial activity. A. annua teas were prepared with plants from one location in south France and from ten different plantations in Benin. Artemisinin was quantified in herbal teas with a liquid chromatography system coupled to mass spectrometry and multiple reaction monitoring detection methods. The herbal teas were tested against chloroquine-sensitive 3D7 strain of Plasmodium falciparum using isotopic microtest to determine IC50 values and calculate the concentration of artemisinin corresponding to the IC50 of the teas [ART(tea)_IC50]. Chemical profiles were determined by liquid chromatography coupled to high resolution mass spectrometry and a metabolomic analysis was performed to annotate compounds statistically linked to the antiplasmodial properties of the teas. Artemisinin content varied between 0.3 mg/L for tea with plants from France to 15.7 mg/L for teas made with plants from Benin with differences between locations. Artemisinin content was decreasing after a one-year storage of the plant for 3 localities in Benin with loss of 33%, 48% and 24% (P < 0.05). Artemisinin concentrations and antiplasmodial activity of teas were positively correlated although the comparison of ART(tea)_IC50s to IC50 of pure artemisinin suggested that other compounds present in the tea were involved in the activity, either enhancing or limiting it. Unknown alkaloids in A. annua teas correlated to antiplasmodial activity were also detected. These findings suggest that A. annua teas deserve further studies to identify other metabolites of interest and determine their role in antiplasmodial activity in relation to other molecules, particularly artemisinin.

Fig 1. Location of the ten A. annua plantations in Benin that took part in the study.

The maps of West Africa and Benin used for this figure come from the geoBoundaries project: Runfola, D. et al. (2020) geoBoundaries: A global database of political administrative boundaries. PLoS ONE 15(4): e0231866. https://doi.org/10.1371/journal.pone.0231866.

Fig 2. ART concentrations in herbal teas from Benin and France.

(A) Concentrations were compared for each series of preparation between locations. To avoid overloading the graph, comparisons with Soual_01 and Soual_02 are symbolized by * for series 1 (the highest P value was considered between comparison to Soual_01 or Soual_02); comparisons with Banigbé (AB) are symbolized by § for both series; (B) the table shows the P values of the other comparisons between locations. (C) Concentrations were compared for each location between the two series of preparations carried out one year apart (series 1 and 2 prepared in Feb. 2021 and Feb. 2022); for Soual in France, the symbol $ indicates the comparison between two harvest periods (January and May 2021). Comparisons were carried out using a t-test. One symbol means P < 0.05, two symbols P < 0.005, three symbols P < 0.0005.

Fig 3. Principal component analysis (PCA) of A. annua herbal teas prepared with plants from different origins in Benin and France.

Samples from Benin are called AS, AD, AODH, ADJR, AC, ACO, ACDT, AMLH, AB and AK. For these samples, numbers 1 and 2 refer to two series of tea preparations (Feb. 2021 and Feb. 2022). Samples from France are called Soual_01 and Soual_02 and were harvested at two different periods (Jan. 2021 and May 2021). a. PCA with French and Beninese samples. b. PCA with Beninese samples only.

Fig 4. IC₅₀ values of herbal teas from Benin and France.

The y-axis on the left is for Soual_01 and Soual_02, the one on the right is for all the other localities in the two series. (A) IC₅₀ values were compared for each series of preparation between locations. To avoid overloading the graph, comparisons with Soual_01 and Soual_02 are symbolized by * for series 1 (the highest P value was considered between comparison to Soual_01 or Soual_02); comparisons with Banigbé (AB) are symbolized by § for both series; (B) the table shows the P values of the other comparisons between locations. (C) IC_50s were compared for each location between the two series of preparations carried out one year apart (series 1 and 2 prepared in Feb. 2021 and Feb. 2022); for Soual in France, the symbol $ indicates the comparison between two harvest periods (January and May 2021). Comparisons were carried out using a t-test. One symbol means P < 0.05, two symbols P < 0.005, three symbols P < 0.0005.

Fig 5. ART content as a function of IC₅₀ of the teas.

The curve is modelled as a power curve (Excel software) for the first (A) and the second series (B). Spearman correlation test was used to test the association between values.

Fig 6. IC₅₀ values based only on the ART contained in A. annua teas (ART(tea)_IC₅₀) from Benin and France.

(A) ART(tea)_IC_50s were compared for each location between the two series of preparations carried out one year apart (series 1 and 2 prepared in Feb. 2021 and Feb. 2022); for Soual in France, the symbol $ indicates the comparison between two harvest periods (January and May 2021). (B) ART(tea)_IC_50s were compared for each series of preparation between locations. Since the values were similar between the two series of preparation, the values from both series were considered for this comparison. * shows the comparisons to the IC₅₀ values of pure ART; (C) the table shows the P values of the other comparisons between locations. Comparisons were carried out using a t-test. One symbol means P < 0.05, two symbols P < 0.005, three symbols P < 0.0005.

Fig 7. LC-HRMS chromatograms of samples AB_1 (series 1) and Soual_02.

Main features described in Tables 2–4 are highlighted. a: proline, b: leucine/isoleucine, c: phenylalanine, d: tryptophan, e: chlorogenic acid, f: unknown alkaloid and 5-feruoylquinic acid, g: unknown alkaloid, h: dicaffeoylquinic acid (mixture of isomers), i: chrysosplenol D, j: arteannuin B, k: ART, l: deoxyartemisinin.