Malaria is a dreadful and chronic infectious disease. It is caused by Plasmodium parasites and transmitted through the bites of female Anopheles mosquito, presenting a severe threat to human health and lives1. Globally, the number of malaria cases has been on a downward trend since 2015. However, the rate of decline has decelerated. In recent years, the emergence and spread of the COVID-19 pandemic, which led to a substantial increase in the number of cases and deaths, has also hindered the efforts to combat malaria2.

Artemisinin (ART) is a sesquiterpene lactone antimalarial drug. It contains an active peroxide bridge, which is a crucial active site. This site can disrupt haemoglobin metabolism, target specific proteins, and act on mitochondrial membranes. ART can directly kill intraerythrocytic Plasmodium, and its is characterized by rapid, high efficiency, and low toxicity. Nevertheless, it has drawbacks such as poor water solubility, a short half-life, and a high relapse rate3,4. In some regions, the sole use of ART has led to a decrease in the sensitivity of Plasmodium, which serves as a warning sign of Plasmodium resistance5. Since the first detection of Plasmodium falciparum resistance in Cambodia, similar ART resistant strains have continuously emerged in some areas of the Greater Mekong Subregion6,7 As a result, the World Health Organization (WHO) has recomended replacing the single use of ART with ART-based combination therapies (ACTs).

Traditional Chinese Medicine, with its multi-component, multi-target, and multi-systemic actions, is less likely to induce drug resistance. Some inactive components may even prevent or reverse the emergence of drug resistance through multiple pathways8,9,10. These include interfering with metabolic and excretory processes and directly influencing drug concentrations in the blood, target organs, or cells. These molecules can be considered as candidate resistance inhibitors or reversal agents for drugs derived from homologous plants. We purified a polymethoxylated flavonol, chrysosplenetin (CHR), from the acetone layer of ART industrial extraction waste. Our previous research indicated that it might function as a potential ART resistance inhibitor through various mechanisms, and the optimized combination ratio between ART and CHR was 1:211,12,13. However, the underlying mechanism remains unclear to this day.

Metabolomics has advanced with the development of high-resolution platforms11,12. It uses1H NMR and MS techniques to analyze metabolic changes at the level of Plasmodium or in the infected host13,14,15. Our laboratory previously applied the non-targeted metabonomic method of1H NMR and UHPLC-QTOF-MS to characterize the metabolic changes in mice infected with ART-sensitive and ART-resistant Plasmodium berghei K173 before and after treatment with ART alone and the ART-CHR combination16,17. We filtered out metabolites that changed mutually and specifically including endogenous amino acids, glucose, pyruvate, and lactic acid. However, non-targeted metabolomic studies are less accurate and less quantitative. In contrast, the “close-up mode” of precision targeted metabolomics, combined with the “microscopic” level, enables in-depth qualitative and quantitative interpretation of known differential metabolites18.

The LC-MS/MS platform is the most commonly used tool in targeted metabolomics for studying thermally unstable, volatile, and polar metabolites like amino acids. Although metabolomics techniques have been successfully applied in several parasite infection studies, there has been relatively little research on the amino acid metabolism disrupted by Plasmodium berghei and how small-molecule inhibitors regulate the infection state. Therefore, in this study, we quantified thirteen pre-screened distinct metabolites in biological samples infected with sensitive and resistant Plasmodium before and after treatment with ART alone and ART-CHR (1:2) combination using HPLC-MS/MS method19. Through multivariate statistical analysis and unidimensional statistical analysis, we identified differential amino acids, aiming to gain insights into the metabolic pathways closely related to the ART resistance and the regulatory function of CHR.

Materials and methods

Chemicals and reagents

Twelve tested amino acid standards, namely glycine (Gly), L-alanine (Ala), L-serine (Ser), L-proline (Pro), L-valine (Val), L-threonine (Thr), taurine (Tau), L-isoleucine (Ile), L-lysine (Lys), L-glutamine (Gln), L-phenylalanine (Phe), L-tryptophan (Trp) along with trimethy-amine N-oxide (TMAO) or internal standard α-aminobutyric acid (AABA) were all procured from Beijing Solarbio Science & Technology Co., Ltd. (China). ART crystal (with a purity of over 98%) was purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. (China). CHR, serving as the partner drug of ART, was self-prepared and identified in our laboratory, boasting a purity of over 98%.

Methanol, acetonitrile, formic acid, and ammonium acetate of chromatographic purity grade was obtained from Beijing Dikema Technology Co., Ltd. (China). Sodium carboxymethyl cellulose (CMC-Na) used for suspending the drugs was purchased from Sinopharm Chemical Reagents Co., Ltd. (China). Bovine serum albumin was sourced from Beyotime Biotech. Inc. (China). Purified water was of Milli-Q grade from a Millipore purification system. Sinopharm Chemical Reagents Co., Ltd. (China) also supplied the following reagents for passaging and freezing mouse Plasmodium strains: sodium citrate, potassium dihydrogen phosphate, trisodium hydrogen phosphate, sodium chloride, and glycerol.

Preparation of the chemical stock solutions

4.0 g of bovine serum albumin (BSA) was dissolved in 100 mL of pure water to prepare a 4% BSA solution, which served as a mimic substrate. CMC-Na (0.50 g) was accurately weighed and dissolved in 100 mL of pure water under continuous stirring. After complete swelling, the solution was stored at 4℃ until use. ART (0.04 g) and its combination with CHR at a ratio of 1:2 was precisely weighed and suspended in 10 mL of 0.5% CMC-Na solution under ultrasonic-assisted condition for 2 h. All these stock solutions were stored at 4℃ before use.

The standard substances of amino acids and TMAO were separately (1 mg) dissolved in a 5% acetonitrile-water solution (V/V) at a 1:1 (W/V) ratio to obtain the respective stock solutions at a concentration of 1.0 mg/mL and were stored at -80 °C in an ultra-low temperature refrigerator.

The inner standard AABA stock solution was freshly prepared by dissolving 1.0 mg of AABA in 5% acetonitrile-water (V/V) at a 1:1 (W/V) ratio, resulting in a concentration of 1.0 mg/mL and stored at -80 °C. Subsequently, 10 µL of the AABA stock solution was diluted with acetonitrile (containing 0.2% formic acid) to prepare an acetonitrile protein working solution containing 10.0 ng/mL of AABA and stored at 4 °C in the refrigerator before use.

Parasite infection with or without drug administrations followed by sample harvests and processes

Parasite passage and invasion

ART-sensitive and -resistant strains of Plasmodium berghei K173 were donated by the laboratory of Professor Yong Dai (School of Basic Medicine, Chengdu University of Chinese Medicine). After 42 generations of passaging using the standard mutiplier increment method, the resistant strain maintained a resistance index of 14.153 as previously reported20.

Male ICR mice (18–22 g) were obtained from the Center for Comparative Medicine, College of Veterinary Medicine, Yangzhou University (SYXK Su, 2017-0044). We confirm that all methods described in this study were indeed carried out in strict accordance with relevant guidelines and regulations. Before the experiments, the protocol was submitted to and approved by the University Ethics Committee, Yangzhou University (YXYLL-2021-10). In addition, all experimental animals adhere to the ARRIVE guidelines throughout the study. The total number of mice is approximately 100, with 10 mice allocated in each group to ensure that there are at least 6 experimental animals in each group after parasite-infection. Only male mice were ultimately ultilized in our study because the immunity of female mice is weaker than that of male mice which results in a considerably high mortality rate when female mice are infected with Plasmodium. Additionally, male mice have relatively more stable behavioral characteristics and are free from the interference of estrogen. All experimental animals were housed in polycarbonate cages and acclimated in an environmentally controlled room (23 ± 2ºC, with adequate ventilation and a 12 h light/dark cycle) for at least 7 days prior to use. They were provided with standard pellet diet and water ad libitum before and throughout the entire experiments. Cervical dislocation was used to euthanize the mice.

After a rapid resuscitation of the parasites, healthy male mice were intraperitoneally injected with 0.2 mL of blood, each containing an average of 106 red blood cells infected with Plasmodium, as the first-generation of malaria transmission. Three days after inoculation, blood was collected from the tail veins of the mice that had an infection rate of 20% and were designated as breeding animals. This blood, which contained parasites, was then used to infect healthy mice. The final infection rate was determined via microscopic analysis with Giemsa stain and the process continued until the infection rate of the newly-infected mice reached 5%. Experimental animals are initially categorized into a non-administered group and an administered group. The non-administered group encompasses a healthy control group, a sensitive Plasmodium-infected group, and a resistant Plasmodium-infected group. Each of these groups is solely administered 0.5% CMC-Na, with a sample size of n = 6 per group. The administered group is also randomly partitioned into the following subgroups: a sensitive-CMC-Na control group (receiving 0.5% CMC-Na, n = 6), a resistant-CMC-Na control group (receiving 0.5% CMC-Na, n = 6), a sensitive ART single-use group (administered at a dose of 40 mg/kg, n = 6), a resistant ART single-use group (administered at a dose of 40 mg/kg, n = 6), a sensitive combination group (with a ratio of 1:2 and doses of 40:80 mg/kg, n = 6), and a resistant combination group (with a ratio of 1:2 and doses of 40:80 mg/kg, n = 6). The dosages were calculated by the equivalent conversion method between mice and human based on body surface area, and the optimized combination ratio 1:2 between ART and CHR was screened out in our previous work21,22,23. After 7 days of continuous gavage administration, the mice were sacrificed after collecting serum samples. Liver, spleen, kidney, and intestine tissues were subsequently harvested and processed according to the steps below, respectively.

Blood and tissue sample harvest and treatment

Approximately 700 µL of blood was collected from the orbit of each mouse into a 1.5-mL centrifuge tube using a capillary tube and left to stand at 4℃ for 1 h to allow solidification and stratification. After centrifuged at 4 °C (800 × g) for 10 min, the supernatant was transferred to a clean centrifuge tube and centrifuged again (4 °C, 12,814 × g, 10 min). The final supernatant was sub-packed into a 0.2-ml centrifuge tube and stored at -80 °C. Later, the frozen serum sample was gradually thawed under a programmed temperature increase from − 80℃ to − 20℃ (2 h), then to 4℃ (1 h), and finally to room temperature (30 min). 200 µL of acetonitrile protein precipitation solution was added into 50 µL of serum sample. The mixture was vortexed for 30 s, and centrifuged at 4℃, and 12,814 × g for 10 min. The supernatant was dried using a Nitrogen Blowing Concentrator. Before use, 200 µL of acetonitrile was added to redissolve the supernatant, with thorough vortexing for 1 min, sonication for 15 min, and then filtration through a 0.22 μm organic-phase membrane. The filtrate was stored at 4℃, and detected by LC-MS/MS.

The liver, spleen, kidney, and intestine tissues were carefully removed from the sacrificed mice, washed in PBS solution (1 ×), and dried with filter paper. The dried tissues were cut into small pieces after weighing. Each 100 mg tissue fragment was placed into a 2 mL centrifuge tube, rapidly frozen in liquid nitrogen (15 min), and stored in an ultra-low-temperature refrigerator at -80℃.

The samples were thawed following a programmed temperature increase: from − 80℃ to -20℃ (2 h), then to 4℃ (1 h), and finally to room temperature (30 min) and they were processed prior to analysis. Briefly, 1 mL of 80% methanol was added to the tissue samples, which were then homogenized in an Automatic Sample Rapid Grinder (Shanghai Jingxin Industrial Development Co. Ltd., China) for 1 min until no debris remained. The mixture was fully vortexed for 1 min and centrifuged at 4℃ for 15 min (15,039 × g) after being sonicated in an ice bath for 15 min. Then, 200 µL of acetonitrile was added into 50 µL of the extracted supernatant and left to stand for 3 min. The supernatant was vortexed for 2 min and centrifuged at 4℃ for 15 min (15,039 × g). The residue was dried in a Nitrogen Blowing Concentrator and submitted for LC-MS/MS analysis.

Simultaneous detection method for the targeted amino acids by LC-MS/MS

The method validation was carried out according to our previous work24. Detailed information and steps are presented as follows.

LC-MS/MS analytical conditions

An Agilent 1200 High-Performance Liquid Chromatograph system coupled with Agilent 6460 triple quadrupole mass spectrometer (Agilent, USA) was employed and the chromatographic separation was performed on a Dikma Inspire Hilic column (2.1 × 100.0 mm, 3 μm). Positive-ion mode under an MRM model provided a satisfying response signal. The injection volume was 2 µL, and the flow rate was 0.3 mL/min. The column temperature was set at 40 °C. The mobile phase consisted of water containing 5 mM ammonium acetate (A, pH = 4.0) and acetonitrile (B). The gradient elution program was set as follows: 0–9 min, 10%A; 9–11 min, 50%A; 11–12 min, 10%A; 12–20 min, 10%A. The total analysis time was 20 min.

An electrospray ionization (ESI) source was used to ionize the chemicals. The optimized MS conditions were as follows: dry-gas temperature, 300℃; gas flow, 10 L/min; nebulizer gas pressure, 15 psi; sheath-gas temp, 250℃; sheath-gas flow, 7 L/min; capillary voltage, 40,000 V; nozzle voltage, 500 V. The screened fragmentor (V) and CE (eV) for different metabolites, as well as other parameters, are shown in Table S1.

Calibrated linearity and method validation

50 µL of stock solutions (1 mg/mL) for amino-acid standards including Gly, Ala, Ser, Pro, Val, Thr, Ile, Lys, Gln, Phe, and Trp, as well as 25 µL of TMAO and Tau, were added to 400 µL of BSA substrate solution to obtain a mixed standard solution at a concentration of 50 µg/mL (Gly, Ala, Ser, Pro, Val, Thr, Ile, Lys, Gln, Phe, and Trp) or 25 µg/mL (TMAO and Tau). Repeatedly, the mixture was diluted by BSA solution to obtain a serial of final concentrations of 25,000, 12,500, 6,250, 3,125, 1,562.5, 781.25, 391.95, 97.7, 48.8, 24.41, 12.21, 6.10 ng/mL and processed according to the sample-treatment method. The linear range of standard curves was 0.61–10000.00 ng/mL.

The amino acid standard solutions were gradually diluted into high-(HQC), medium-(MQC), and low-concentration quality control working solution (LQC) and then treated according to the preparation method of the standard curves. The final concentrations were 5000.00, 1250.00, and 312.50 ng/mL, respectively. The solutions were stored at 4 °C in the refrigerator before use.

The calibrated standard curve was freshly prepared to obtain the linear regression equation and correlation coefficient (R). Intra- and inter-day precision, the lowest limit of detection (LLOD) or the lowest limit of quantification (LLOQ), and stability along with recovery and matrix effects, were routinely investigated to evaluate the method reliability.

Multivariate statistical analysis

The concentrations of the targeted metabolites were uploaded into the MetaboAnalyst 5.0 system for multivariate statistical analysis. First, the mass-spectrometry data were processed in the Agilent Qualitative Analysis B.03.00 Tools software, and the concentration of each metabolite (ng/mL) was calculated and visualized using GraphPad Prism 9. After normalization, multivariate statistical analysis was performed, combining unsupervised principal component analysis (PCA) and supervised orthogonal partial least squares-discriminant analysis (OPLS-DA). The variable importance in the projection (VIP) > 1, |p(corr)|>0.58, and p < 0.05 were used as screening criteria, with the Fold change (FC) set to 1 or 2 according to the difference multiplier, to obtain differential metabolites. The intersection for altered metabolites between different groups was depicted by a Venn diagram, and the disturbed metabolic pathways were mapped using a mouse-sourced KEGG database in the MetaboAnalyst system (Kyoto Encyclopedia of Genes and Genomes, https://www.kegg.jp/).

Results

Body weights and organ indexes of mice infected with sensitive or resistant malarial parasites

As clearly illustrated in Fig. 1-I A, over a 7-day observation period, the body weights of healthy mice demonstrated a consistent upward trjectory. In marked contrast, the body weights of all infected groups underwent a decline, with the sensitive-infected groups showing a more substantial decrease. As presented in Fig. 1-I B, the administration of medications did not bring about a statistically significant improvement in the weight loss of the infected mice. When compared with the vehicle-treated group, both the sensitive and resistant groups exhibited an increase in spleen indexes in Fig. 1-II C-D, while their intestine indexes decreased significantly as shown in Fig. 1-II E-F. In Fig. 1-II G-J, both ART monotherapy and its combination treatment promoted the liver, spleen, and intestine indexes in mice infected with sensitive parasites when compared to the control group. Specifically, within the sensitive-infected groups, the combination treatment led to lower spleen index but a higher intestine index compared to ART monotherapy. In the resistant-infected groups, the spleen and kidney indexes were alleviated under the treatment of either ART monotherapy or its combination as compared to their corresponding sensitive-infected counterparts. However, there were no significant differences between the two treatment modalities in the resistant-infected groups, as depicted in Fig. 1-II H and I. Additionally, ART monotherapy increased the intestine index of resistant-infected mice compared to sensitive-infected ones, as shown in Fig. 1-II J.

Fig. 1
figure 1

Body weights (I) and organ index (II) of the healthy (C) and infected mice with the sensitive (S) and resistant (R) malarial parasites in the presence and absence of ART monotherapy (S-ART or R-ART) and ART-CHR combination (S-1:2 or R-1:2). C, G. Liver; D, H. Spleen; E, I. Kidney; F, J: Intestine.

Full size image

Varied amino acids and their metabolic pathways in sensitive and resistant groups with or without ART monotherapy and its combination

The typical total ion chromatogram (TIC) diagrams of the biological samples, which were shown in Fig. S1. The LLOQs were comprehensively evaluated. In terms of accuracy, the values ranged from − 12.59 to 7.89%, and for precision, they ranged from 1.77 to 14.00%. Moreover, the inter- and intra-day precision, matrix effect, recovery, and stability all met the acceptable standards. Regularly, QC samples were employed to monitor the applicability of the method for detecting analytes, with the relative standard deviation (RSD%) maintained at less than ± 15%. Consequently, the targeted amino acids in serum, liver, spleen, kidney, and intestine samples were precisely quantified, and the results were listed in Table S2-S4.

PCA and OPLS-DA analysis

For the serum samples, the PCA scores plot in Fig. 2-I A accounted for a total of 69.3% of the variables, with PC1 contributing 48.3% and PC2 contributing 21%. The sample distributions between the healthy and resistant groups were close and were located on the left-side of PC1. In contrast, the samples from the sensitive groups exhibited excellent intra-group aggregation and distinct inter-group dispersion. The OPLS-DA analyis, as shown in Fig. 2III-A, achieved a remarkable separation through pairwise comparison. Similarly, for liver, spleen, and kidney specimens, the PCA scores plot (Fig. 2-I B, C, and D) and OPLS-DA (Fig. 2-III B, C, and D) clearly demonstrated a good separation between any two subgroups. However, in the case of intestines, as shown in Fig. 2-I E, the PCA scores plot showed an overlap between the sensitive and resistant groups, suggesting their similar metabolic profile. Nevertheless, OPLS-DA still managed to provide an ideal separation, as depicted in Fig. 2III-E.

Fig. 2
figure 2figure 2figure 2

PCA (I andII) and OPLS-DA analysis (III and IV) of healthy (C) and infected mice with the sensitive (S) and resistant (R) malarial parasites in the presence and absence of ART monotherapy (S-ART or R-ART) and ART-CHR combination (S-1:2 or R-1:2). PCA analysis among healthy control (C, red) and sensitive (S, purple) or resistant groups (R, green) without drug treatments (A. serum; B. liver; C. spleen; D. kidney; E. intestine); PCA analysis among infective control (S in purple color or R in green color) and medication groups by ART alone (S-ART in green color or R-ART in purple color) and ART-CHR combination (S-1:2 or R-1:2 both in red color) (F. Sensitive groups; G. Resistant groups). OPLS-DA analysis between healthy control (C, red) and sensitive (S, purple) or reistant groups (R, green) without drug treatments (H. serum; I. liver; J. spleen; K. kidney; L. intestine); OPLS-DA analysis between infective control (S in purple color or R in green color) and medication groups including ART monotherapy (S-ART in green color or R-ART in purple color) and ART-CHR combination (S-1:2 or R-1:2 both in red color); M. S vs. S-ART; N. S vs. S-1:2; O. S-1:2 vs. S-ART; P. R vs. R-ART; Q. R vs. R-1:2; R. R-1:2 vs. R-ART.

Full size image

The PCA scores plot in Fig. 2-II failed to achieve a satisfactory separation among the three drug-treatment groups across different bio-samples. The OPLS-DA analysis within the sensitive groups showed a good separation when comparing S and S-ART or S-ART and S-1:2, but not between S and S-1:2. In the resistant groups, distinct separations were observed only in the following pairwise comparisons: S vs. S-ART in the serum, liver, and intestine samples; S vs. S-1:2 in the kidneys and intestines; S-1:2 vs. S-ART in the serum, livers, and intestines; R vs. R-ART in the livers, kidneys, and intestines; R-ART vs. R-1:2 in the serum and liver samples, as presented in Fig. 2-IV. This indicated that the genotypic difference of Plasmodium berghei under different drug pressures led to discrepant metabolic phenotypes characterized by spatial variation.

Differential amino acids and the interfered pathway network

The heat maps in Fig. 3-I and II effectively visualized the contents of differential analytes and clustered them at the multivariate levels through pairwise comparison. For instance, in the serum samples (Fig. 3-I A), when compared with the vehicle group (C vs. S), a decrease in Lys, Phe, Gln, Tau, TMAO, Trp, Val, and Ile along with an increase in Ala, Pro, Gly, and Ser was observed in the sensitive groups. Similarly, in the resistant groups (C vs. R), Thr, Tau, TMAO, Lys, Val, and Gln were found to be reduced. When comparing the resistant groups with the sensitive groups (S vs. R), an increased in Ile, Phe, Gln, Lys, Trp, Ala, Pro, TMAO, Tau, and a decrease in Val, Thr, Gly, and Ser were identified. The differential analytes in the tissues were screened out in a similar fashion, as shown in Fig. 3-I B, C, D, E. Briefly and importantly, compared with those in the sensitive groups, only Thr showed a consistent downward trend in all biological samples from the mice infected with resistant parasites, while Phe was up-regulated in all tissues except the intestine tissue. The levels of other metabolites were inconsistent and exhibited significant tissue-specific differences. Furthermore, Fig. 3-II (F-J) presented the interfered metabolites under the treatment of ART and its combination. Notably, compared with ART monotherapy, the most pronounced effects of the ART-CHR combination were observed in the resistant intestines. As shown in Fig. 3-II J-6, compared to ART monotherapy, the combination led to a reduction in Lys, Thr, Phe, Ile, Trp, and Pro, and an increase in Gln, Gly, TMAO, and Tau.

Fig. 3
figure 3figure 3

Heat map analysis. (I) Heat maps of serum (A), liver (B), spleen (C), kidney (D), and intestine (E) samples from the sensitive (S) or resistant (R) parasite infected mice without medication; (II) Heat maps analysis of serum (F), liver (G), spleen (H), kidney (I), and intestine (J) samples from the sensitive (S) or resistant (R) parasite infected mice under ART alone or ARH-CHR combination; 1. S vs. S-ART; 2. S vs. S-1:2; 3. S-1:2 vs. S-ART; 4. R vs. R-ART; 5. R vs. R-1:2; 6. R-1:2 vs. R-ART.

Full size image

The S-plot analysis, depicted in Fig. S2-I and II, clearly visualized the relationship between the differential metabolites and the calibrated principal components. The general screening criterion was set at |p(corr)|>0.58. VIP score plots in Fig. S2-III and IV effectively displayed the influence of each variable and metabolites with VIP > 1 was selected for further analysis. Volcano maps, presented in Fig. S2-V and VI, intuitively illustrated the statistical and biological significance through T-test analysis and FC value calculation. Venn diagrams in Fig. S3-I and II precisely indicated the number of specific and shared differential metabolites between any two subgroups. The integrated application of these methods, therefore, convincingly demonstrated the significance of these metabolites at the multivariate statistical level.

The accurate contents of the screened metabolites with statistical significance are presented in Fig. 4-I (A-E) and -II (F-K). Through pairwise comparisons, it was evident that, as presented in Fig. 4-I A-E (R vs. S), resistant parasites consistently increased Phe levels in the serum, livers, spleens, and kidneys when compared to sensitive parasites. The same pattern was observed for Trp in the serums, spleens, and kidneys, and for Ile in the serums and kidneys. Gly levels decreased in both the serums and livers. Meanwhile, there was an increase in Lys levels in the serums and a decrease in TMAO levels in the livers. As depicted in Fig. 4-II K-1 and − 5, compared to ART monotherapy, the ART-CHR combination reversely reduced Phe levels in the serums and intestines, and inhibited the elevation of Trp and Ile in the intestines of the resistant groups. Although significant differences exist among other metabolites, they either exhibit changes consistent with those in the resistant group or have no differences when compared with the resistant group. These metabolites are, therefore, not within the scope of our evaluation.

Fig. 4
figure 4figure 4

Histograms of differential metabolites in serum and tissues from the mice infected with sensitive (S) and resistant (R) Plasmodium berghei K173. (I) Comparison between sensitive and resistant groups (A. serum; B. liver; C. spleen; D. kidney; E. intestine). (II) Histograms of differential metabolites in serum and tissues from the mice infected with sensitive (S) and resistant (R) Plasmodium berghei K173 under the treatments of ART alone or ART-CHR combination; F. S vs. S-ART; G. S vs. S-1:2; H. S-1:2 vs. S-ART; I. R vs. R-ART; J. R vs. R-1:2; K. R-1:2 vs. R-ART. (1) Serum; (2) Liver; (3) Spleen; (4) Kidney; (5) Intestine.

Full size image

The visual bubble diagrams of the enrichment pathways and their spatial metabolic network were plotted based on KEGG (Kyoto Encyclopedia of Genes and Genomes) content25. The visual bubble diagrams of the enrichment pathways and their spatial metabolic network were plotted in Fig. S4-I and -II and Fig. 5-I and -II), respectively. In Fig. S4-I and -II, it is clear that the larger the pathway impact, the higher the enrichment degree. The color of the bubbles, ranging from yellow to red, indicates a successive decrease in the P-value. The larger the bubble size, the greater the number of metabolites enriched in the corresponding pathway. Among these pathways, phenylalanine, tyrosine, and tryptophan biosynthesis was significant in most of the subgroups. The mutual and specific metabolites between the sensitive and resistant subgroups or between ART alone and its combination were clearly visualized in Fig. 5-I and -II. In conclusion, phenylalanine metabolism, phenylalanine, tyrosine, tryptophan biosynthesis, tryptophan metabolism, and valine, leucine, isoleucine biosynthesis were identified as the potential metabolic targets of CHR as an inhibitor of ART resistance. These pathways and their upstream and downstream molecules, thus, potentially play crucial roles in the mechanisms of CHR against ART resistance.

Fig. 5
figure 5

Spatial metabolic pathway network of differential amino acids. (I) Comparison between mice infected with sensitive (S) and resistant (R) parasites; Metabolites labeled in red and blue respectively represented up-regulated and down-regulated. Metabolites labeled in orange suggested that there may be oppositely regulated between the groups. The superscript “S” or “R” means the variations original from sensitive or resistant groups. (II) Comparison between mice infected with sensitive (S) and resistant (R) parasites under the treatments of ART monotherapy and ART-CHR combination. The superscript arabic numbers next to the letter S or R represents 1S vs. S-ART or R vs. R-ART, 2S vs. S-1:2 or R vs. R-1:2, 3S-ART vs. S-1:2 or R-ART vs. R-1:2.

Full size image

Discussion

In our prior research employing1H NMR-based untargeted metabolomics, we hypothesized that an infection with P. berghei could disturb the host’s amino-acid metabolism. Specifically, we predicted perturbations in the metabolites of amino acids such as Ala, Val, Ile, Tyr, Gly, and TMAO, as detailed in reference16. Moreover, we observed that both ART-sensitive and -resistant parasites manifested overlapping as well as distinct metabolic characteristics within the murine hosts17. Building upon these findings, the present study utilized a LC-MS/MS targeted metabolomic approach to quantitatively analyze twelve selected endogenous amino acids and TMAO, a metabolite closely associated with gut microbiota. These analyses were conducted on diverse biological samples obtained from hosts infected with the parasite. Our results revealed that amino-acid metabolism is indeed disrupted during Plasmodium infection and in the development of drug resistance. Notably, CHR was found to potentially act as a restorative agent for amino-acid metabolism homeostasis, particularly in relation to ART resistance in Plasmodium berghei K173.

Phe, an essential amino acid isolated from Escherichia coli, also functions as an antagonist of the α2δ subunit of voltage-dependent Ca2+ channels, with a dissociation constant (Ki) of 980 nM26. Phe hydroxylase, the key enzyme responsible for the conversion of Phe to tyrosine (Tyr), often malfunctions during inflammatory conditions or malignancies27. Biochemically, Phe and Tyr can be metabolically processed to generate L-DOPA, a precursor of dopamine28. In the context of the malaria-vector interaction, quorum sensing-activated metabolic conversion of Phe to acetyl-CoA drives the biogenesis of outer membrane vesicles. This process enhances the resistance of mosquito commensals to Plasmodium colonization29. Additionally, four dipeptides, namely Phe-Arg, Arg-His, Trp-Arg, and Tyr-Arg, have been identified as promising candidates for antimalarial development30. Trp metabolism also played a crucial role in the malaria life-cycle. Molecules derived from Trp catabolism have been shown to synchronize the P. faciparum life cycle30. Furthermore, DPY19-medicated Trp C-mannosylation ensures the quality control of thrombospondin-repeat proteins at two critical life-cycle stages, which is essential for the successful transmission of human malaria parasites31. Ile, an essential branched-chain aliphatic amino acid prevalent in numerous proteins, is also a metabolite either found in or produced by Escherichia coli. As an isomer of leucine, Ile plays a pivotal role in hemoglobin synthesis and in regulating blood-sugar and energy levels. Ile can be synthesized through multiple pathways, including the deamination of Thr or the catalysis of aspartic acid by Thr deaminase. However, the precise role of Ile in ART resistance and the mechanism by which CHR reverses this resistance remain elusive and thus merit further in-depth investigation.

Conclusion

This study aimed to screen out the potential amino acids associated with ART resistance and to explore how CHR regulated these metabolites. The metabolic pathways of Phe, Trp, and Ile metabolic pathways along with their upstream or downstream molecules, are likely to be of the greatest interest for elucidating the precise mechanism in our future research.