Until recently, there have been, to our knowledge, few well-controlled studies examining extraction, recovery, and stability of artemisinin and other compounds in A. annua tea infusion. A systematic study of preparations of A. annua therapeutic tea infusion was performed by van der Kooy et al[27] and showed that nearly 93% of available artemisinin was extracted from dried A. annua leaves, but only under certain conditions. Best preparation method was: 9 g DW leaves/L, for 5 min at 100 °C. Subsequent storage of the tea infusion at room temperature showed that artemisinin concentration was stable for > 24 h, important for malaria-endemic locations where there is no refrigeration. Artemisinin water solubility is approximately 50 mg/L[27], so the amount of artemisinin recovered from hot water tea infusions is reasonable. Other studies using the same extraction protocol also measured extraction and stability of artemisinin and some key flavonoids in the tea. Artemisinin was found to be stable at room temperature for up to 48 h[28]; however, some flavonoids were poorly extracted and not stable at room temperature[29].

Carbonara et al[28] detected an assortment of phenolics, including 0.06 mg/g DW cirsilineol, in an A. annua tea infusion prepared at about a 4-10 fold higher proportion (approximately 38 g DW/L) than that proposed as optimal (9 g DW/L) by van der Kooy et al[27]. Most of the measured phenolics in the tea remained constant at room temperature for 48 h post-infusion. More recently, Suberu et al[19] identified milligram amounts of phenolic acids, flavonoids, and sesquiterpenes in a liter of A. annua tea, all of which demonstrated IC₅₀ values in the micromolar or less range (Table 1). Indeed, the IC₅₀ of the tea infusion itself was 7.6 and 2.9 nmol/L for the chloroquine (CQ)-sensitive HB3 and CQ-insensitive Dd2 strains of P. falciparum, respectively, and better than artemisinin alone suggesting synergism of constituents in the tea mixture. Clearly if a tea infusion is to be a therapeutic option, it must be consistently and reliably prepared and ingested. As suggested by van der Kooy et al[27], ideally a liter of tea infusion would be prepared daily and consumed in equal aliquots of about 250 mL over 24 h for several days.

Ogwang et al[30,31] tested Artemisia tea as a prophylaxis against malaria in 132 adult farm workers, aged 18-60 years, for 12 mo in a randomized clinical trial in Uganda. Tea infusion was consumed once a week at 2.5 g dried leaves per adult infusion dose with 55-100 mg artemisinin/L. Malaria was tracked for 9 mo while adverse clinical effects were tracked for 12 mo. Among those who used Artemisia tea there were 80% fewer fever-related hospital visits. Indeed, some patients reported using A. annua tea for > 7 years with no incidence of malaria and no serious adverse events. Although this study suggested that once weekly consumption of A. annua tea infusion may offer prophylactic protection, there were no children or elderly in the study, so additional clinical trials need to be conducted with different populations and age groups. Authors argued that since a single weekly dose was effective, compounds other than artemisinin may have played the prophylactic role since artemisinin itself has short plasma half-life.

Reports on the efficacy of A. annua (cv. Artemis) tea on human malaria patients by Mueller et al[17,32] and Blanke et al[33] yielded at times conflicting results. Their tea infusions contained 47-94 mg artemisinin/L, but recrudescence was much lower in the quinine-treated control group, so parasite reappearance in the tea-treated patients was ascribed to recrudescence and not re-infection[17]. In the Blanke et al[33] trial that included a placebo tea, recrudescence was consistently lower in the tea patients than in those treated with 500 mg pure artemisinin. More recently, however, De Donno et al[34] showed that 5 g dried leaves in one liter of A. annua tea infusion was effective against both CQ-resistant (W2) and CQ-sensitive (D10) strains of P. falciparum with IC₅₀ values of 5.60 nmol/L and 7.08 nmol/L, respectively, results also consistent with those of Suberu et al[19] as already highlighted. These latter in vitro studies suggested that tea should be efficacious, so why the discrepancy with the earlier human trials? Preparation methodology is crucial for preserving as much biochemical integrity of the plant as possible[27]. The more recent in vitro studies likely used more consistently prepared tea infusions than the earlier human trials, so variations in chemical composition of the infusions and in the plant source material could explain the different responses.

The argument that tea is a monotherapy is unsubstantiated considering the now well-established chemical complexity and related antiplasmodial activity of tea infusions of A. annua and its components. Although data from therapeutic tea trials in animals and in humans correlate well, unfortunately, they do not support use of A. annua tea for treating malaria because animal and human data are comparably negative, the artemisinin dose is not easily controlled, and other potentially synergistic components in the tea are not readily controlled or extracted. Nevertheless, use of the tea could play a role in malaria prophylaxis to reduce incidence of malaria in different communities, or in temporary relief from malaria, mainly in prevention of coma or “to buy time” to enable an infected person from a rural area to travel to a hospital or clinic stocked with ACT.

Recently, Elfawal et al[23] measured parasitemia in mice infected with P. chabaudi that were fed two different doses (0.6 or 3.0 mg artemisinin; 24 and 120 mg/kg) of either pure artemisinin in mouse chow or as pACT. Artemisinin delivered via pACT was at least five times more effective, and with a longer lasting response, than pure artemisinin in reducing parasitemia. Excluding artemisinin there are > 600 phytochemicals that have been identified in Artemisia annua[35], but there is currently a lack of information on the chemistry, effect of the preparation method (harvesting, drying, storage, etc.), and overall bioavailability of these chemicals[36].

Clinical trials using dried leaf A. annua are scarce in the scientific literature and few, other than those in Democratic Republic of Congo by Mueller et al[17,32], are published. Despite the fact that WHO does not encourage either whole plant or tea infusion clinical trials[37], some African universities have been conducting their own trials, many of which have not been published nor results assessed by polymerase chain reaction (PCR) as later done for clinical trials with ACTs (personal comm from C. Kasongo to P. Lutgen). Many of these trials used A. annua infusions, and compared to controls or even other antimalarial drugs, e.g., artesunate-amodiaquine, showed significantly greater sensitivity of the infusion with fewer late therapeutic failures. For example, in Democratic Republic of Congo, 54 malaria-infected volunteers were treated for 10 d with capsules containing powdered leaves of A. annua. Each patient was given 15 g dried leaves containing 15 mg of artemisinin (artemisinin content in leaves = 0.1%[38]). After 2 d all were free of fever and 51 (or 94%) were parasite free after 10 d.

In a study aimed at preventing severe post-operative malaria at Bangui, Central Africa, powdered leaves of A. annua were administered in capsules to 25 patients, 22 of them children aged 1-16 years[24]. Treatment duration ranged from 3-4 d with a dose of 0.4-0.5 g/d of A. annua dried leaves (0.1% artemisinin leaf content) delivering 0.4-0.5 mg/d artemisinin. In spite of the very low administered daily dose of artemisinin, average parasitemia dropped by 62% in the patients with an added benefit of a strong antinociceptive response, especially beneficial to post-operative patients.

The most clinically definitive study to date of pACT efficacy was conducted at the International Centre of Insect Physiology and Ecology (ICIPE) Mbita Field campus, Suba District, in Western Kenya. This was a collaborative project between ICIPE and Kenya Medical Research Institute[20] (Table 2[39]) and was an open-label, non-randomized clinical trial mainly targeted to assess efficacy, safety, and tolerance of increasing doses of pACT delivered as tablets. The tablets were made by a Tanzania-based NGO, Natural Uwemba System for Health, from a hybrid of A. annua grown in the Tanzania highlands (2000-2200 m altitude). Leaves were harvested just before flowering, dried for approximately 3 wk under shade, then crushed, powdered, homogenized, and pressed into 500 mg tablets under ambient temperature. Tablets were robust with no excipient required. Using HPLC with diode array detector, analysis of hexane extracts of randomly selected batches of 100 tablets showed artemisinin content of the tablets was consistent at 0.74% ± 0.06% (i.e., approximately 3.7 mg per tablet).

The four cohorts of the trial each had 12 consenting patients aged 15-56 years (average 23.42) with P. falciparum malaria. Based on Giemsa-stained blood smears counted against 200 wbc, parasitemia was 0.02%-4% and hemoglobin levels > 8 mg/dL. Each cohort received one of four increasing numbers of A. annua tablets, ranging from 2-5 tablets twice on day 1, followed by 1-4 tablets twice daily for the next 5 d (Table 2). A week following the treatments, three patients scattered throughout different cohorts showed re-appearance of parasites in blood smears; however, all doses were effective in clinical and parasitological regression of malaria, with 9%-20% recrudescence at day 28 and no measurable toxicity.

Compared to the usual large pure artemisinin doses of 1000 mg on day 1 followed by 500 mg on each of days 2-7 that were administered to 227 malaria patients[39], patients treated with pACT had generally better therapeutic outcomes (Table 2). The measured pACT cure rate also was comparable to or exceeded other results using pure artemisinin[40,41], and similar levels of artemisinin (artesunate, artemether, etc.[42]). Furthermore, the positive therapeutic response using pACT appeared somewhat independent of dose beyond the second level of dose tested (Table 2[20]). Although oral doses used in the ICIPE[20] trials were far less than any tea studies, levels of recrudescence were much lower than tea and often better than in studies using pure artemisinin[39] (Table 2). Indeed, about 100 total mg of total artemisinin delivered via pACT for a full malaria treatment yielded a better recrudescence rate than the 4000 mg of pure artemisinin used by Giao et al[39] (Table 2). This 40-fold difference correlates well with the early pharmacokinetic studies by Weathers et al[21] that showed 45-fold enhanced bioavailability of the drug when delivered as pACT.

These results suggest that the natural phytochemical blend in pACT is important especially when orally administered as tablets. The results are also consistent with a study in China on mice infected with P. berghei, which compared the effects of pure artemisinin with crude A. annua extracts[43], and the studies by Elfawal et al[23] and Weathers et al[22]. In all three studies the administered products had comparable levels of artemisinin, but crude preparations and pACT were at least 3.5 times more effective in reducing parasitemia than pure artemisinin, suggesting a synergistic role for non-artemisinin constituents in the extracts and orally consumed dried leaves.

When given orally or rectally, dihydroartemisinin showed higher bioavailability in humans than artemisinin in an early pharmacokinetic study by Zhao et al[44]. The C_max, T_max, and T_1/2 for orally delivered dihydroartemisinin were 0.13-0.71 mg/L, 1.33 h, approximately 1.6 h, respectively; for pure artemisinin they were 0.09 mg/L, 1.5 h, and 2.27 h, respectively. Alin et al[45] compared orally delivered artemisinin and artemisinin-mefloquine combination therapy for treatment of P. falciparum malaria. Infected and uninfected patients had similar pharmacokinetic parameters. After a single dose, bioavailability of artemisinin was not altered. Interestingly, pharmacokinetics were similar when comparing treatment failures with successes, suggesting that studies that only measure artemisinin pharmacokinetics were inadequate for predicting therapeutic success[45]. Ilet et al[46] also reviewed artemisinin pharmacokinetics in patients with falciparum malaria and reported a dose of 9.1 mg/kg, which was comparable to that of Alin et al[45]. C_max and T_max values did not differ much from those reported by Alin et al[45].

In the Ilet et al[46] review of pharmacokinetic parameters of artemisinin and its derivatives, oral pure artemisinin doses ranged from about 6-11 mg kg/L in healthy subjects and C_max was 0.15-0.39 mg/L. Dose seemed to have no major effect. An earlier study by Ashton et al[47] compared increasing artemisinin doses of 250, 500, and 1000 mg per person and both C_max and T_1/2 showed dose-dependent increases of 0.21, 0.45, and 0.79 mg/L, and 1.38, 2.0, and 2.8 h, respectively, but T_max remained relatively constant at 2.3-2.8 h.

Diet is an important consideration for any orally delivered drug, and when Dien et al[48] compared artemisinin oral doses given with and without food, C_max values were similar between subjects who fasted and those who did not. Food consumption along with artemisinin did not seem to affect artemisinin absorption. In contrast, a later rodent study by Weathers et al[21] observed that when artemisinin was consumed as part of a complex plant material, pACT, approximately 45-fold more drug entered the serum of mice than orally administered pure drug. Similarly, when pure artemisinin was fed to mice, it was not detectable in the serum after 60 min. However, artemisinin was detected in the serum when consumed in conjunction with mouse chow, which consists of a variety of plant materials including soy, oats, wheat, alfalfa, beet pulp, corn, etc[22].

In a study by Ashton et al[49], artemisinin at 9.1 mg/kg was given daily for 7 d, and measurements taken on days 1, 4, 7, and 21. On day 1 plasma C_max and T_1/2 were similar and comparable to data from other studies using a similar dose. On day 4 and 7, however, C_max decreased, while T_1/2 increased, indicating that although artemisinin was delivered daily for 7 d, it was either not readily absorbed or it degraded after the first dose. After the third dose, C_max fell from 0.31 to 0.11 mg/L, and T_1/2 increased from 3.0 to 4.8 h. These results suggested that either artemisinin was metabolized or accumulated elsewhere in the body.

In the liver, cytochrome P450 (CYP450) enzymes metabolize artemisinin to deoxyartemisinin, deoxydihydroartemisinin, 9,10-dihydrodeoxyartemisinin, and a metabolite named “crystal 7”[50]. Extended artemisinin dosing may not be beneficial as shown by Svensson et al[50] using human liver microsomes where activity of CYP450s, CYP2B6 in particular, correlated with decreasing artemisinin serum levels. In intermittent dosing studied by Ashton et al[49], the P450 levels were allowed to decline for 14 d before delivery of another dose, and C_max rose from 0.11 to 0.20 mg/L, and T_1/2 decreased from 4.8 to 2.7 h. Generally, maximum concentration of artemisinin in the body increased with increasing doses with T_1/2 ranging from about 1.4-4.8 h for reported trials using oral pure artemisinin. Thus, increased and extended artemisinin treatment may reduce recrudescence.

Other than Räth et al[16], there are few reports on the pharmacokinetics of tea infusion artemisinin delivered in humans. In the Räth et al[16] study, artemisinin C_max was 0.24 mg/L at 0.6 h post consumption. Tea infusion containing 94.5 mg artemisinin had a C_max equivalent to a dose of 250 mg pure artemisinin, but at a significantly shorter T_max, 0.6 h vs 2.8 h[47]. Compared to pure artemisinin, the shorter half-life of artemisinin in the tea infusion may account for the observed higher recrudescence. Although tea-delivered artemisinin seemed more bioavailable, its shorter T_1/2 of 0.9 h compared with about 2 h for pure artemisinin, suggested that more than two doses per day may be more beneficial; indeed, four doses a day were recommended.

The unacceptably high recrudescence rates in clinical tea infusion trials were attributed to low plasma concentrations, almost 40% lower than that for traditional doses (500 mg per person of 60 kg or 8.3 mg artemisinin/kg) of pure artemisinin. Although not specified, tea trial doses have been estimated at about 1.5 mg/kg, close to the 1.1 mg/kg dose of pure artemisinin used by Zhao et al[44], which is far below the 8.3 mg/kg that is traditionally accepted as pharmacologically effective. Nevertheless, the C_max of 0.24 mg/L artemisinin for the tea dose is nearly twice that for pure artemisinin (C_max = 0.13 mg/L) as measured by Zhao et al[44]. A. annua tea also showed potent antiplasmodial activity against 40 field isolates of P. falciparum collected in Pikine, Senegal (mean IC₅₀ 0.095 µg/mL[51]).

There are as yet no pharmacokinetic studies of pACT in humans. In a small PK study of healthy mice fed artemisinin there was about 45-fold more artemisinin delivered via pACT than when delivered as the pure drug[21]. More recently, pharmacokinetics of artemisinin and one of its liver metabolites, deoxyartemisinin, were compared over 120 min in healthy and P. chabaudi-infected mice treated with dried A. annua leaves at a 100 mg/kg body weight dose of artemisinin[22]. In pACT-treated healthy mice, the first order elimination rate constant for artemisinin was estimated to be 0.80/h, corresponding to a T_1/2 of 51.6 min. C_max and T_max were 4.33 mg/L and 60 min, respectively. The AUC was 299.5 µg.min/mL. The first order absorption rate constant was estimated at 1.39/h. In contrast, the AUC for pACT-treated infected mice was greater at 435.6 µg·min/mL. Serum levels of artemisinin in the infected mice continued to increase over the 120 min of the study period. As a result, the elimination half-life, T_1/2 could not be determined, so C_max and T_max could only be estimated at ≥ 6.64 mg/L and ≥ 120 min, respectively. Nevertheless, both C_max and T_max of artemisinin were greater in infected than in healthy mice.

Generally, artemisinin concentrations decreased with a concomitant rise in deoxyartemisinin levels only in healthy subjects[22]. In contrast, artemisinin levels in infected mice continued to rise over the study period whilst deoxyartemisinin levels fell and then leveled, so infection seemed to retard the capacity of the mice to process artemisinin into deoxyartemisinin over the two-hour period. Many compounds in A. annua inhibit P. falciparum[52-55] and CYP34A[56]. At the high (100 mg/kg) dose used in the study, nearly equal amounts of artemisinin and deoxyartemisinin were measured in the serum, indicating that an excessive dose of artemisinin was used.

The presence of plant material affected artemisinin pharmacokinetics. At 60 min no artemisinin was detected in serum of mice fed pure artemisinin at 100 mg/kg body weight. When plant material was present, however, as mouse chow or A. annua pACT, artemisinin level in the serum rose to 2.44 and 4.32 µg/mL, respectively, demonstrating that the presence of plant material, even mouse chow, had a major positive impact on the appearance of artemisinin in the blood[22]. To our knowledge, these are the only data available on pharmacokinetics for orally delivered A. annua in animals or humans.

A. annua is rich in essential oils, coumarins, polyphenols, polysaccharides, saponins, terpenes, and flavonoids. The levels of flavonoids and other compounds in A. annua change with developmental growth stage, with some being highest during full bloom[57]. There are > 40 flavonoids[13], and at least 11, including artemetin, casticin, chrysoplenetin, chrysoplenol-D, cirsilineol, eupatorin, kaempferol, luteolin, myricetin, quercetin, and rutin, are reported to have weak therapeutic efficacy against falciparum malaria (Table 1[52-54,58]). Some of these flavonoids were shown to improve the IC₅₀ of artemisinin against P. falciparum in vitro by as much as 50%, suggesting synergy (Table 1[52]). Elford et al[53] also showed that while casticin [5-hydroxy-2-(3-hydroxy-4-methoxyphenyl)-3,6,7-trimethoxychromen-4-one] showed synergism with artemisinin, it did not synergize with chloroquine, suggesting a different interactive mechanism. Combining casticin with artemisinin inhibited parasite-mediated transport systems that control influx of myo-inositol and L-glutamine in malaria-infected erythrocytes. These apparent synergistic actions between flavonoids and artemisinin suggest that flavonoids are likely to be important for efficacious use of A. annua consumed either as whole dried leaves or as tea.

Many flavonoids have antiplasmodial effects and inhibit P. falciparum growth in liver cells in vitro as reported for dietary flavonoids[54]. To our knowledge, there are no reports on pharmacokinetics of A. annua delivered flavonoids. Some flavonoids are reported to have long plasma half-lives; e.g., quercetin, found in A. annua and most fruits, has a plasma half-life of 27 h[59]. Quercetin [2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy-4H-chromen-4-one], also found in garlic, inhibits parasite growth with differential activity against different strains of Plasmodium (Table 1[54,58]). Rutin, which is a rutinose [α-L-rhamnopyranosyl-(1→6)-β-D-glucopyranose] glycoside of quercetin, showed similar results, suggesting that the sugar moiety did not significantly affect antimalarial activity (Table 1[58]). Flavonoids are known to persist in the body for > 5 d; this may explain the once a week dose inducing a prophylactic effect from A. annua tea infusion that was reported by Ogwang et al[30,31]. Many dietary flavonoids inhibit Plasmodium growth in vitro, but amounts in the diets are reportedly insufficient to offer protection against malaria[54]. Plants such as A. annua with high concentrations of flavonoids (e.g., up to 0.6%) may, however, work in concert with artemisinin to prevent malaria when consumed regularly.

The flavone luteolin [2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-4-chromenone] comprises up to 0.0023% DW in Artemisia[14] and has been used for a variety of ailments including cough, diarrhea, dysentery, diabetes, cancer, and malaria. Although luteolin has an IC₅₀ value around 11 µmol/L[54] and is one of the more active antiplasmodial flavonoids found in A. annua, one cannot compare its role between studies as indicated by Ganesh et al[58] (see Table 1). The antimalarial response of different flavonoids seems to be affected by the strain of Plasmodium being tested. Luteolin also prevents completion of a full intra-erythrocytic cycle by inhibiting progression of parasite growth beyond the young trophozoite stage. The mechanism of this antiplasmodial activity seems to be related to the inhibition of parasite fatty acid biosynthesis. These lipids are required by the parasite to detoxify heme into hemozoin[60]. Independent of the human host, apicomplexan parasites use a fatty acid biosynthetic pathway. Enzymes in the pathway, like the NADPH-dependent b-ketoacyl-ACP reductase (FabG), are potential antimalarial targets. Among 30 flavonoids studied, luteolin and quercetin had the lowest IC₅₀ values for the inhibition of these enzymes and also showed in vitro activity in the sub-micromolar range against multiple strains of P. falciparum[60].

Isovitexin {5,7-dihydroxy-2-(4-hydroxyphenyl)-6-[(2S,3R,4R,5S,6R)-3,4,5-trihydroxy-6-(hydroxymethyl) oxan-2-yl]chromen-4-one} is another flavone, the 6-C-glucoside of apigenin, that was found in A. annua tea infusion at > 100 mg/L with micromolar antiplasmodial activity (Table 1[19,28]). Isovitexin inhibits lipid peroxidation and xanthine oxidase activity and protects cells from ROS damage with an overall LD₅₀ > 400 µmol/L[61].

Limonene (1-methyl-4-(1-methylethenyl)-cyclohexene) is part of the “cineole cassette” that includes 1,8-cineole (eucalyptol), limonene, myrcene, α-pinene, β-pinene, sabinene, and α-terpineol[62]; many of these affect particular stages of Plasmodium species. For example, limonene is often present at 7 mg/kg in A. annua[14] and inhibits isoprenoid biosynthesis in Plasmodium[63] and development at the ring and trophozoite stages[64]. Eucalyptol affects the trophozoite stage[65]. Limonene also arrests protein isoprenylation in P. falciparum, halting parasite development within 48 h of treatment[64]. The IC₅₀ against in vitro Plasmodium in these trials was 2.27 mmol/L, more than twice the IC₅₀ of 533 µmol/L measured by van Zyl et al[55]. Limonene and its metabolites remain in the plasma for at least 48 h[66], so the pharmacokinetics is favorable, which is important for elimination of gametocytes and malaria transmission.

The volatile monoterpene α-pinene (4,6,6-trimethylb-icyclo[3.1.1]hept-3-ene) is present in the plant at levels up to 0.05% of dry weight[14]; it has an IC₅₀ of 1.2 µmol/L, in the range of quinine at 0.29 µmol/L[55]. Eucalyptol (1,8-cineole) may comprise up to 30% [0.24%-0.42% (V/DW)] of the essential oil in A. annua[67] and is a strong inhibitor of the pro-inflammatory cytokines tumor necrosis factor (TNF)-α, interleukin (IL)-6 and IL-8[68]. Both chloroquine-resistant and chloroquine-sensitive Plasmodium strains are affected at the early trophozoite stage[65].

Eucalyptol (1,3,3-trimethyl-2-oxabicyclo[2,2,2]octane) is also volatile and rapidly enters the blood when delivered either as an inhalant or orally[69,70]. At an IC₅₀ of 0.02 mg/mL and low toxicity (LD₅₀ of approximately 25 mg/mL), either oral or inhalation delivery is reasonable[65,71]. Indeed eucalyptol concentrations can reach 15 µg/mL in 60 min[69] suggesting its possible use as an antimalarial inhalant.

Artemisia ketone (3,3,6-​trimethyl-​1,5-​heptadien-​4-​one), a major constituent of some cultivars of A. annua, has barely been studied. Other ketones like curcumin[72] have been implicated as inhibitors of β-hematin synthesis, so artemisia ketone may play a similar role and affect hemozoin formation. Although hemoglobin is required for Plasmodium survival and multiplication in merozoites inside the red blood cell, it leaves toxic debris like heme. The parasite subsequently oxidizes Fe²⁺ in heme to Fe³⁺ forming hematin, a nontoxic insoluble polymeric crystal called β-hematin (also known as hemozoin), which also inhibits cell-mediated immunity against the parasite. Water extracts of A. annua inhibit hemozoin synthesis[73].

Essential oils often contain a large amount of monoterpenes that may enhance the antimalarial effect of artesunate and even reverse the observed resistance of P. berghei against artesunate[74]. Monoterpenes tend to be higher in the pre-flowering phase of A. annua[75], but are drastically reduced by high drying temperatures or drying in the sun[13,76] and, of particular concern, during compression of dried leaves into tablets[77]. Although monoterpenes have some antimalarial potential, most are rather volatile and thus they may be therapeutically less important than the nonvolatile flavonoids, phenolic acids, and higher molecular weight sesquiterpenes.

Unlike α-pinene and eucalyptol, camphor (1,7,7-trimethylbicyclo[2.2.1]heptan-2-one) has no reported antimalarial activity, but it may comprise as much as 43.5% of the essential oil of A. annua[78]. Considering camphor is less volatile than either eucalyptol or α-pinene (melting points of 204 °C, 176 °C, and 155 °C, and flash points of 54 °C, 49 °C, and 33 °C, respectively), it may instead play a role in enhanced transport of hydrophobic molecules like artemisinin from pACT across the intestinal wall into the bloodstream[21,22]. Camphor may also affect thymocyte viability and aid in developing malaria immunity through production of T-cells[79]. At 50 μg/mL, camphor increased viability of cultured thymocytes[80].

The sesquiterpene nerolidol (3,7,11-trimethyl-1,6,10-dodecatrien-3-ol) has an IC₅₀ of 0.99 µmol/L and arrests development of the intraerythrocytic stages of the parasite (Table 1[55]). Indians of the Amazon basin in Brazil treated malaria using the vapors of the leaves of Viola surinamensis; nerolidol was identified as the active constituent leading to 100% growth inhibition at the schizont stage[81]. Nerolidol levels vary with the cultivar tested, with one of the highest values found in plants from Ethiopia[82]. There is a greater concentration of this sesquiterpene in stems than leaves of A. annua[83].

Other sesquiterpenes found in the artemisinin biosynthetic pathway were only recently shown to have antiplasmodial activity at µmol/L levels, similar to that of other compounds found in the plant (Table 1[19]). These artemisinic compounds were extracted into A. annua tea infusions and showed varying interactions with artemisinin depending on their relative concentrations and the target parasite strain. For example, arteannuin B showed an additive interaction with artemisinin against the CQ-sensitive Plasmodium HB3 strain, while against the CQ-insensitive Dd2 strain the interaction was synergistic.

Rosmarinic ((2”R”)-2-[[(2”E”)-3-(3,4-dihydroxyphenyl)-1-oxo-2-propenyl]]oxy]-3-(3,4-dihydroxyphenyl) propanoic acid) and chlorogenic ((1S,3R,4R,5R)-3-{[(2Z)-3-(3,4-dihydroxyphenyl)prop-2-enoyl]oxy}-1,4,5-trihydroxycyclohexanecarboxylic acid) acids are strong antioxidants found in a wide variety of A. annua cultivars[56]. In Caco-2 studies, these acids significantly inhibited activity of CYP3A4, one of the hepatic P450s responsible for metabolism of artemisinin to deoxyartemisinin, an inactive form of the drug[50]. These and other phenolic acids are present in A. annua tea infusion[19]. Both phenolic acids have an IC₅₀ of about 65 µmol/L (Table 1) and also significantly reduced secretion of cytokines IL-6 and IL-8, and thus enhanced antimalarial activity while reducing inflammation[56].

Although polysaccharides in other medicinal plants have been more extensively studied, they seem to have been rather overlooked in A. annua, probably because most Artemisia extracts are obtained using organic solvents and polysaccharides are only soluble in water. Polysaccharides extracted from Artemisia iwayomogi showed hydroxyl radical scavenging activity three times stronger than glutathione or caffeic acid, and ROS inhibition was twice as strong as ascorbic acid[84]. In A. iwayomogi, more polysaccharides were found in stems than in leaves and their solubility was also higher from stem than from leaf tissue[84].

The combination of polysaccharides with lipophilic molecules like artemisinin may lead to a higher bioavailability of the antimalarial constituents when delivered via A. annua, which may explain the lower effective therapeutic dose against malaria observed for pACT than for pure artemisinin[20,23,26]. Indeed, Han[85] showed that ginseng polysaccharides had preventive and curative antimalarial activities and synergized with artesunate in malaria-infected mice. Sulfated polysaccharides inhibited the in vitro invasion of merozoites into erythrocytes and interfered with merozoite surface protein[86-88]. Heparin and other sulfated polysaccharides have been shown to inhibit blood-stage growth of plasmodium[89,90]. Some sulfated polysaccharides inhibited the formation of rosettes between infected red blood cells (iRBC) and uninfected RBCs, as well as adhesion of iRBCs to placental chondroitin sulfate A, which is linked to severe disease outcome in pregnancy-associated malaria[91].

Saponins, common in many plants, have an important role in human and animal nutrition and are reportedly present in A. annua, but only as measured in alcoholic extracts using the nonquantitative foaming test[92,93] (Weathers, unpublished). These soap-like amphiphilic (lypo- and hydro-philic) bioactive compounds are mainly produced by plants. Recently, there has been interest in the clinical use of saponins as chemotherapeutic agents[94], and as adjuvants for vaccines[95]. At very low doses saponins are efficient, have hemolytic properties, produce 40-50 Å pores in erythrocyte membranes, and modulate the sodium pump and ATPase[96]. Saponins also have a hypoglycemic effect mainly by inhibiting intestinal permeability and absorption of glucose and may therefore inhibit the growth of P. falciparum, which needs glucose to grow[97]. Better identification, quantification, and investigation into the role of saponins in pACT efficacy are warranted.

The coumarin, scopoletin (7-hydroxy-6-methoxychromen-2-one), also known for its antinociceptive properties[98,99], is commonly found in most Artemisia species at, for example, about 0.2% (w/w) in a Luxembourg cultivar. Known for its anti-oxidant, hepatoprotective, and anti-inflammatory activities, scopoletin scavenging capacity for hydroxyl radical, DPPH, superoxide anion, hydrogen peroxide, and Fe²⁺ chelating activity is almost at the level of α-tocopherol (Vitamin E)[100].

Although not antiplasmodial, scopoletin inhibits TNF-α, IL-6, and IL-8 at millimolar concentrations, and is thus likely one of the major anti-inflammatory and antipyretic constituents of A. annua[101]. Coumarins can activate lymphocytes, thereby stimulating immunological functions[102]. Indeed, scopoletin induced cell proliferation in normal lymphocytes with an immunomodulatory effect[101]. In uninfected erythrocytes internal Na concentration is much lower than external concentration, but the K concentration is higher; in infected blood cells this situation is drastically reversed[103]. Scopoletin significantly stimulated erythrocyte membrane ATPases at 0.1 µmol/L, in particular Na-K-ATPase vs Ca-ATPase or Mg-ATPase[104], so scopoletin may affect malaria infection. A significant hormetic effect was also noticed; stimulation was higher at scopoletin concentrations of 10 µg/mL than at 1 or at 100 µg/mL. In addition scopoletin also inhibited ADP-platelet aggregation at a range of 0.1 to 5 µmol/L and improved blood rheology[105].

Scopoletin may also affect the interaction between malaria and uric acid. Cyclical fevers and high levels of inflammation characterize malaria and this likely aids parasite clearance. Excessive and persistent inflammation, on the other hand, can lead to severe malaria[106]. In the cytoplasm of their parasitophorous vacuole, Plasmodium-infected erythrocytes contain uric acid precipitates that are released upon erythrocyte rupture. Uric acid precipitates are mediators for inflammatory cytokines IL-6, IL-8, and are considered a danger signal for innate immunity. Uric acid is also the causative agent in gout. These precipitates could offer a novel molecular target for anti-inflammatory therapies in malaria. Scopoletin inhibits the activity of xanthine oxidase in hyperuricemic mice after peritoneal administration, and this hypouremic effect is fast and dose-dependent[107].

Although many of the compounds in A. annua have not been tested for their toxicity in, a survey of available MSDS data showed that the LD₅₀ levels for orally administered compounds in rodents ranged from about 160 mg/kg for quercetin to > 8000 mg/kg for nerolidol. The artemisinin LD₅₀ measured via oral dose in a mouse was 4228 mg/kg. Therefore, at the estimated amounts of dried leaves of pACT that may be orally consumed by a malaria patient, most of the compounds reported thus far in A. annua are at concentrations that are orders of magnitude below their LD₅₀ toxicity values.

Toxicology of the dried leaf tablets used in the Kenyan human trial measured the following components: serum levels of urea, serum proteins, creatinine, γ-glutaryl transferase, serum glutamic pyruvic transaminase, serum glutamic oxaloacetic transaminase, or alkaline phosphatase levels, hemoglobin, and pre- and post-electrocardiograms[20]. Compared to levels prior to treatment with pACT, there was no significant change post-treatment.

Because production costs are usually closely held secrets, there are few cost estimates that are publicly available to compare pACT production with extracted artemisinin. However, costs can be estimated from a study by de Vries et al[108] where they reported a 1 kg recovery of artemisinin from A. annua containing 0.6% artemisinin. Downstream processing costs and product losses increase with increasing number of unit operations (unit ops), a fact often not generally appreciated[109]. Indeed for biotechnology processes, recovery can be anywhere from 9%-51%[110]. As an example, if each step of a 4 step process is 95% efficient, then the overall process has a final efficiency of about 81%, while a single step process at 95% efficiency has a 95% overall recovery. The described process steps for extracted artemisinin (eAN) vs pACT-AN are shown in Figure 2. From the point of harvested dried leaves to material ready for packaging or conversion to the delivered drug (e.g., artesunate or artemether), pACT has one unit op and eAN has eight[108]. Extraction solvents and other chemicals are clearly no longer part of the cost. Because there is one vs eight unit ops for eAN and at least two of the eAN unit ops involve significant amounts of heat, pACT energy cost is significantly reduced by at least 90%. Costs for labor, interest, depreciation, and maintenance are all also affected by the number of unit ops[109], so we estimated that with seven fewer unit op steps those costs would reduce by approximately 88%. Although better extraction processes may be in play[111], using the de Vries et al[108] analysis our estimate of cost reduction for producing pACT is about 30% less than the cost of producing eAN. Data provided by de Vries et al[108] was based on 0.6% artemisinin content, so if a higher producing cultivar was harvested, costs would drop proportionately. Moreover, cost drops again because with pACT there is no need to convert artemisinin to artesunate or artemether; those conversions were necessary because they have higher bioavailability than pure artemisinin, which is not an issue with pACT[21,22].

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Comparison between plant-based artemisinin combination therapy production and extracted artemisinin from dry harvested leaves to product ready either for packaging (plant-based artemisinin combination therapy) or conversion to artemether or artesunate (extracted artemisinin). AN: Artemisinin; eAN: Extracted artemisinin; pACT: Plant-based artemisinin combination therapy.

The de Vries et al[108] process cost estimation focuses on a production yield of 1 kg of artemisinin from 500 kg dried leaves, so per Giao et al[39] that amount of pure artemisinin would treat only 250 patients. Based on the data shown in Table 3 from Kenyan or WPI A. annua at 0.7 and 1.4% artemisinin, 15 and 7.5 g DW leaves, respectively, are required for a total adult pACT treatment; so from 500 kg leaves, 33300 and 66600 patients could be treated, respectively. This represents more than a 130-fold increase in patients treated compared to pure artemisinin with proportionate reduction in price.

A. annua dry leaf production varies around the globe. “In East Africa yields average 2.5 T/ha (range = 0.75-4.2)...”[112]. Based on our field trials[113], the reported average A. annua leaf production in E. Africa[112], and the doses used in the Kenyan human trial[20], one can estimate the amount of dry leaf production, and depending on the amount of artemisinin in the biomass, estimate possible number of adult patients that could be treated with pACT (Table 3).

Using the dosing information obtained from the Kenyan human malaria trial[20], each adult needs about 100 mg artemisinin total over 6 d for a malaria treatment, so for A. annua leaves with 0.7% artemisinin, 15 g of dried leaves would be needed for a 6 d treatment course. At 2 ton of dried leaves harvested per hectare, 127260 adult patients could be treated for malaria (Table 3). For leaves containing 1.4% artemisinin, only 7.5 g of dried leaves are required, so from a hectare of land producing 2 tons of leaves twice as many patients could be treated (Table 3). Clearly choosing cultivars that have higher levels of artemisinin in their leafy biomass will dramatically increase the number of patients that can be treated from 1 ha.

According to Roll Back Malaria, from one ton of purified artemisinin current ACT therapy can provide 1.76 million adult malaria treatments using artemether/lumefantrine, and 2.5 million adult treatments using artesunate/amodiaquine[114] (Table 4). Using the same one ton artemisinin equivalent, but delivering the drug via pACT with 0.7% artemisinin content, one would have harvested about 142.8 tons of dried A. annua leaves. Assuming 15 g dried leaves per patient from the dosing data in the Kenyan human malaria trial (Table 2[20]), 8.64 million adult patients could be treated, about a four-fold increase over either of the current ACT drugs. The actual cost of pACT, therefore, mainly depends on the cost of the dried leaves and their artemisinin content.

As yet unpublished data from the Rich and Weathers labs demonstrated that pACT prevents emergence of artemisinin drug resistance; the plant itself seems to function as its own ACT (pACT). This would obviate the need for inclusion of a co-drug as used in currently administered ACTs. The co-drug costs at least as much as the artemisinic portion of the drug[6]. Consequently, elimination of the added co-drug could result in at least an additional 50% reduction in cost, so that the final pACT cost reduction is conservatively estimated to be far below that of a current course of ACT therapy.

Considering that A. annua is nontoxic and safe to consume orally, dose may not have to be adjusted for children. On the other hand, the leaves taste bitter, so masking the taste, perhaps with sugar, should help with pediatric treatment. Our recent simulated digestion study showed that adding table sugar (sucrose) to pACT did not significantly alter the amount of artemisinin released after digestion, with the added benefit of doubling the amount of flavonoids released[115].

There are at least three other emerging antimalarial therapeutic technologies: synthetic artemisinin[116], semi-synthetic artemisinin (SSA) production from genetically engineered microbes[117], and a single dose drug, OZ439[118]. In early 2013, Sanofi/PATH Drug Development Programme, announced they would have the capacity to produce up to 60 MT of SSA in 2014 at about $400/kg, depending on quantity; Sanofi now has WHO prequalification for its SSA[119]. Although not much cheaper than the current price of about $550/kg[120], supply would be more or less unlimited. Despite what might seem as an advantage to large amounts of SSA production, there are also some serious disadvantages, and comparison of some advantages and disadvantages for each of these new synthetic antimalarial drugs and pACT is noted in Table 5.

The traditional and least costly method for cultivating A. annua uses seeds and in developing countries farmers prefer to save seeds from one growing season to the next. However, seed generated plants of A. annua will vary widely from generation to generation even with high quality starting stock (see review by Ferreira et al[10]). Stem cuttings of A. annua readily root in about two weeks, so clonal propagation via rooted cutting is recommended to eliminate this variability. Although this method of propagation is not cost effective for large plantations, it would work for a few hectares or for controlled environment agriculture. Given the large numbers of patients that could be treated from growing just a few hectares of A. annua (Table 3), clonal propagation by rooted stem cuttings is recommended. Since pACT therapy involves the direct consumption of the dried leaves of the plant, harvested leaf material must be kept clean, which is easiest to do in controlled environment agriculture and following Good Agricultural Procedures[121], particularly as applied to fresh produce[122]. However, controlled agriculture would probably result in loss of agricultural jobs, a concern to be assessed locally. Alternatively, great care must be taken during field harvest and post-harvest storage, so as not to affect the quality of the product. WHO has established good agricultural practices specifically for A. annua for purposes of artemisinin extraction[123], for general medicinal plants[124], and to minimize contamination of herbal medicines[125].

To deliver a reliable dose of therapeutics to a patient, the dried leaves of harvested A. annua must have a reliable and consistent composition. Clonal propagation provides the required consistency. Recently we showed that of 10 crops harvested from vegetative and early flowering plants grown over three years under diverse conditions in the lab, field, and home garden, the artemisinin content of a single clone of A. annua (SAM) was 1.38% ± 0.26% (w/w)[77]. Thus, despite variations in culture and environmental conditions, a consistent level of the main therapeutic constituent can be achieved. Moreover, the content of harvested leaves is certainly not a guarantee of finished product, e.g., compressed leaf tablets. Analyses by Weathers et al[77] showed that although artemisinin content was very stable after tablet compression, other constituents vaied significantly. For example, although flavonoids increased with tablet compression, the more volatile monoterpenes decreased substantially. Thus, it is critical to monitor the composition profile of both incoming harvested material as well as the final product.

Complex and expensive analytical procedures have been used to analyze the many products found in A. annua, but they are not necessary to measure and assure product quality. Artemisinin is easily extracted and then can be quantified using a variety of thin layer chromatography (TLC) methods and visualized with p-anisaldehyde stain[126,127]. Other key constituents like the flavonoids are also readily separated using TLC and visualized under either UV ± AlCl₃ reagent[128]. Total flavonoids also can be quantified using inexpensive visible spectroscopy via the AlCl₃ method with quercetin used as an inexpensive standard. To our knowledge no inexpensive, reliable spectrophotometric assay is available to measure artemisinin in complex plant extracts.

Artemisinin and its derivatives are also effective against a number of viruses[129], a variety of human cancer cell lines[130-133], and several neglected tropical diseases including schistosomiasis[134], leishmaniasis[135,136], trypanosomiasis[137], and some livestock diseases[133,138].

Although they rank below malaria in terms of public health importance, schistosomiasis, leishmania, and trypanosomiasis result in estimated annual infections of about 240 million, 1.3 million (0.3 visceral and 1.0 cutaneous), and 30000, respectively[139]. These diseases along with many others respond to treatment with artemisinins. Although the IC₅₀ is about 1000-fold greater than for Plasmodium sp., the greater apparent bioavailability of artemisinin via oral pACT[20-22] would likely reduce the amount of drug required for treatment. At present, pACT has not been tested in vivo for diseases other than malaria.

Malaria treatment is further complicated for Human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) patients. Malaria and HIV co-infection represents a major health burden in Africa mainly because it is now “well established that HIV infection results in a higher incidence and more severe manifestations of malaria”[140]. With a weakened immune system, AIDS patients are more susceptible to malaria and also respond slower to malaria therapy[140-142]. Furthermore, in a meta-analysis by Tusting et al[143], socioeconomic development strongly correlated with better malaria therapeutic outcomes. Recently, A. annua has demonstrated anti HIV activity[126,144] and thus oral consumption of the dried leaves of this herb will not only treat malaria, but should also enhance the well-being of HIV/AIDS patients.

A. annua is grown in more than 75 countries[145]. In 2011 about 163 MT of artemisinin were extracted from plantations and small stakeholder farms mainly located in China, Vietnam, and Eastern Africa including Madagascar; value was about $550/kg[120]. With the advent of the production of semi synthetic artemisinin by Sanofi, 60 MT were projected for 2014 with an anticipated price of about $400/kg[119]. As this new source of artemisinin becomes available, the Netherlands Royal Tropical Institute projected that the market for natural Artemisia will significantly destabilize, undermining the security of farmers. The Tropical Institute was further concerned that “pharmaceutical companies will accumulate control and power over the production process; Artemisia producers will lose a source of income; and local production, extraction and (possibly) manufacturing of ACT in regions where malaria is prevalent will shift to the main production sites of Western pharmaceutical companies”, disrupting the fragile economics of these already impoverished countries[120]. The average small stakeholder crop area is about 0.2 ha in China and Africa[120], so while implementation of pACT may not require as much agricultural land as for extracted artemisinin, it could still help provide small stakeholders with a source of income. We have estimated that localized micro manufacturing plants could be constructed for < $50000 USD, and produce quality-controlled pACT tablets with readily verifiable contents. Our overall approach, schematically illustrated in Figure 3, leads to local control of malaria and possibly other artemisinin susceptible diseases while also improving the socioeconomic status of the populations.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Overall scheme for plant-based artemisinin combination therapy production. pACT: Plant-based artemisinin combination therapy; TLC: Thin layer chromatography.