Discovery of Sustainable Drugs for Neglected Tropical Diseases: Cashew Nut Shell Liquid (CNSL)‐Based Hybrids Target Mitochondrial Function and ATP Production in Trypanosoma brucei

Abstract In the search for effective and sustainable drugs for human African trypanosomiasis (HAT), we developed hybrid compounds by merging the structural features of quinone 4 (2‐phenoxynaphthalene‐1,4‐dione) with those of phenolic constituents from cashew nut shell liquid (CNSL). CNSL is a waste product from cashew nut processing factories, with great potential as a source of drug precursors. The synthesized compounds were tested against Trypanosoma brucei brucei, including three multidrug‐resistant strains, T. congolense, and a human cell line. The most potent activity was found against T. b. brucei, the causative agent of HAT. Shorter‐chain derivatives 20 (2‐(3‐(8‐hydroxyoctyl)phenoxy)‐5‐methoxynaphthalene‐1,4‐dione) and 22 (5‐hydroxy‐2‐(3‐(8‐hydroxyoctyl)phenoxy)naphthalene‐1,4‐dione) were more active than 4, displaying rapid micromolar trypanocidal activity, and no human cytotoxicity. Preliminary studies probing their mode of action on trypanosomes showed ATP depletion, followed by mitochondrial membrane depolarization and mitochondrion ultrastructural damage. This was accompanied by reactive oxygen species production. We envisage that such compounds, obtained from a renewable and inexpensive material, might be promising bio‐based sustainable hits for anti‐trypanosomatid drug discovery.


Introduction
Neglected tropical diseases (NTD) are ag roup of 17 highly debilitating and potentially fatal poverty-related diseases.T hese include protozoan,b acterial, and helminthic infections that prevaili nt ropical ands ubtropical areasi n1 49 countries. Communities living in poverty,l acking accesst ob asic sanitation and in close contact with infectious disease vectors, domestic animalsa nd livestock,a re those worst affected. [1] Notwithstanding the recent re-emergence of interesti nN TD, continuous research efforts are neededt os ustain any drug development pipeline in the medium and long term. [2] Human African trypanosomiasis (HAT) is one of the mostn eglected tropical diseases,e ndemici ns ub-Saharan Africa. Trypanosomab rucei rhodesiense (East and SouthernA frica) and T. b. gambiense (West and CentralA frica) are the causative protozoanp arasites, which are transmitted to humans by tsetse flies that are found only in Africa. [3] This disease,w hich is disabling and fatal if left untreated, is am ajor cause of rural underdevelopment and severely affects economiesa nd communities. The control ande limination of HAT, which are declared goals of the WHO, [1] would be am ajor step in the reductiono f the overall burden of tropical disease that continues to limit development in sub-Saharan Africa. [4] Although HATt ransmission is limited to the tsetse belt, comprisingm uch of sub-Saharan Africa, the risk of HATi nt ravelers and migrants, albeit low,cannot be overlooked. [5] Moreover, animal African trypanosomiasis (AAT) has an enormous impact on African agriculture and food security.T his condition is caused by relatedt rypanosome speciesi ncluding T. b. brucei, T. congolense, T. evansi,a nd In the search for effective and sustainable drugs for human African trypanosomiasis (HAT),w ed eveloped hybrid compounds by merging the structural features of quinone 4 (2-phenoxynaphthalene-1,4-dione) with those of phenolic constituents from cashew nut shell liquid (CNSL). CNSL is aw aste product from cashew nut processing factories, with great potentiala sa source of drug precursors. The synthesized compounds were tested against Trypanosoma brucei brucei,i ncluding threem ultidrug-resistant strains, T. congolense,a nd ah uman cell line. The most potent activity was found against T. b. brucei,t he causativeagent of HAT. Shorter-chain derivatives 20 (2-(3-(8-hy-droxyoctyl)phenoxy)-5-methoxynaphthalene-1,4-dione) and 22 (5-hydroxy-2-(3-(8-hydroxyoctyl)phenoxy)naphthalene-1,4dione) were more active than 4,d isplaying rapid micromolar trypanocidal activity,a nd no human cytotoxicity. Preliminary studies probing their mode of action on trypanosomes showedA TP depletion, followed by mitochondrial membrane depolarization and mitochondrion ultrastructural damage. This was accompanied by reactive oxygen species production.W e envisage that such compounds, obtainedf rom ar enewable and inexpensive material, might be promising bio-based sustainable hits for anti-trypanosomatid drug discovery.
T. vivax,a nd, not necessarily being dependent on tsetse flies, has spread to much of South America, South Asia, and the Middle East. [6] HATt reatment today relies on five drugs (pentamidine (PMD), suramin, melarsoprol, nifurtimox,a nd eflornithine;s ee Supporting Information Figure S1 for chemical structures), which suffer from toxic side effects, lack of efficacy,a nd development of resistance. [1] The managemento fp atientsu sing these drugs is complex and risky,r equiring the support services of aw ell-trained staff. [7] Moreover,d espite the fact that almosta ll HATc ontrol programs subsidize the cost of drugs and hospitalization, [8] the availability of quality medicinal agents on as ustainable basis is an increasingly appreciated public health care concept. Consequently,l owering the costs of therapy by developing new drugs based on inexpensive resourcesisavaluable approachtob epursued.
Based on the above considerations, as well as on our continuous interest in the NTD field, we explored the possibility of using cashewn ut shell liquid (CNSL) as as ustainable, low-cost startingm aterialf or the developmento fnew drugs against trypanosomiases. CNSL, which is obtained as the by-product of cashew nut processing, has provent ob eone of the mostv ersatile food wastes for the production of functional materials and chemicals. [9] However,i ts potential as ap recursor of drugs has been relativelyu nderexplored. Being an inedible waste material, it presents clear environmental, financial and ethical advantages over synthetic drugs and even overn atural products derived from crops grown for that purpose. [10] In addition, the fact that East Africa (Tanzania, Kenya, and Mozambique) and WestA frica (Benin, Guinea-Bissau, Ivory Coast,and Nigeria) are among the largest CNSL-producing countries, opens up the exciting possibilities to engagee ndemic countries as crucial actors in NTD drug discoverya nd development. [11] On this basis, we have developedanew chemical library of CNSL-derived hybridsa nd investigated their anti-trypanosomal potential. In particular, the compoundsw ere evaluated against wildtype (WT) and multidrug-resistant African trypanosomes; T. b. brucei is extremelyc losely relatedt ot he human-infectious species, and T. congolense is the principal agentc ausing AATin Africa. Some of the compounds displayed low micromolar activity against T. b. brucei and absence of toxicity on ah uman cell line. We therefore investigated the mechanism by which this compound class exerts its trypanocidal activity.

Results and Discussion
Design rationale CNSL mainly consists of phenolic lipids, that is, anacardic acids (1 in Figure1)( 71.7 %), cardanols (2)( 4.7 %), and cardols (18.7 %) (3). [12] The pentadecyl alkyl side chain of 1-3 may be saturated, mono-olefinic,d i-olefinic or tri-olefinic with ah igh percentage of the components having one or two double bonds (Figure 1), depending on the productionm ethod. [12] Al-thoughC NSL components have been reported to possess a wide range of biological activities,i nm any cases they are not potent enough to be drug candidates. [9] To overcome this limi-tation, their use in combination with standard drugs, and the design of new semi-syntheticd erivatives have been exploited. [9] Along these lines, we decided to develop as eries of CNSLbased hybrid compounds. In particular,b uildingo nt he strategy that the combinations of two different fragments into one covalently linked hybrid compound can convey synergy and increasep otency, [13] we combined the chemical features of CNSL derivatives with those of ap reviously developed anti-trypanosomal hit compound (4 in Figure 2). [14] Intriguingly,b oth 4 [14b] and am ixture of anacardica cids, [15] isolated from Brazilian CNSL, have been reported to inhibit trypanosomalg lyceraldehyde-3-phosphate dehydrogenase( GAPDH), an essential glycolytic enzymeand avalidated anti-trypanosomatid target. [16] Furthermore, thanks to the presence of an aphthoquinone moiety, 4 was shownt og enerate reactive oxygen species  (ROS), am echanism that may further contribute to its multitarget trypanocidal activity. [14b] In fact, 4 exhibitedh igh potency against T. b. rhodesiense (STIB 900 strain) (IC 50 = 80 nm)a nd a promising selectivity index (SI) of 74, with respect to L6 mammalian cells. [14a] In particular, we anticipated that overlapping 4 with CNSL derivatives 5 and 6 ( Figure 2) through their commonp henoxy moietyc ould lead to hybrids with an improveda nti-trypanosomal profile and an improved sustainability.I na ddition, considering that the presence of the long alkyl chain (C 15 )might limit drug-likenessd ue to excessive lipophilicity (see predicted physicochemical properties in Ta ble 1a nd S1) and might give rise to surfactant properties and nonspecific activities, we also turned our attention to the shorter-chain (C 8 )C NSL derivative 7.F ollowing this design strategy,t he small combinatorial library of 8-22 was generated ( Figure 2a nd Scheme 1).
In vitro activity against T. b. brucei wild-type and resistant strains and T. congolense wild-type strain The parent compound 4 and the CNSL-based hybrid derivatives 8-22 were tested for effects on cell viability against bloodstream trypomastigotes of the standard drug-sensitive T. b. brucei strain 427WT.T he most active compounds were further tested against the multidrug-resistant strain B48 and the drug-transporter deletion mutanta qp2/aqp3-KO ( Table 1). The current first-line drug PMD was used as the reference compound. It should be noted that the EC 50 valuesr eportedh erein were produced using ad ifferent species (T. b. brucei versus T. b. rhodesiense)a nd ad ifferentr esazurin-based protocolf rom that used in the previous report, [14a] in which the EC 50 value of 4 against T. b. rhodesiense was found to be 80 nm.T he very robust protocol used here uses am uch higherc ell density and consequently resultsi ns ubstantially higherE C 50 values. [20] Using this protocol, only compounds 18-22,w hich carry the shorter (C 8 )a liphatic chain, showeda nti-trypanosomal activity in the micromolar range (5.0-40.5 mm). Importantly,u nder these conditions, 18-22 had highera ctivity than 4,w hich displays an EC 50 value of 48.7 mm.C rucially,c ompounds 18-22 showedn os ign of human cytotoxicity:n one of the compounds displayed any effects on either the viability or growth of the human foreskin fibroblast (HFF) cell line at the highest tested concentration (200 mm). Notably, 18-22 displayeds tatistically identical activity (p > 0.05) against the aqp2/aqp3-KO cell line from which the well-characterized drug transporter HAPT1/TbAQP2 has been deleted, [21] resulting in am oderated level of PMD resistance (Table 1). Indeed, there was no cross-resistanced etected even in the very highly multidrug-resistant cell line B48, although it displayed 178-fold resistance to PMD in this series of experiments (Table 1). Interestingly,compounds 18-22 were up to 12-fold more active against all strains than the starting compound 4,s uggesting an improvement of the activity in at least some of the newly synthetized CNSL-based hybrids. This shows that the shorter alkyl chain with ap rimary alcoholice nd function, together with the regioisomerics ubstitution on the naphthoquinone moiety, enhances the anti-trypanosomal activity of the test compounds (see below).
The two most active compounds 20 and 22 andt he parent compound 4 werea lso tested against T. congolense IL3000 cell line andc ompared with the standard drug against animal trypanosomiasis, diminazenea ceturate (DA in SupportingI nformation Figure S1). All three compounds were less active than DA, with EC 50 values in the mid-to-high micromolar range, as well as severalfold less active than against T. brucei. Despite the small number of compounds tested, the results with T. congolense suggest that this species may be systematically less sensitive to this scaffold than T. brucei is. Similarly, T. congolense is less sensitivet han T. brucei to suramin [6] and PMD, [22] reflecting speciesdifferencesind rugt arget or accumulation.

Structure-activity relationship (SAR) of trypanocidalactivity
Our goal was to design and synthetize as mall libraryo fC NSLbased hybrids, aiming for asynergistic inhibition of energy metabolism in T. b. brucei,t argeting mitochondrial functions and GAPDH inhibition. Along the line of this rationale, we designed the library of CNSL-based hybridss tartingf rom quinone 4 and longer (5 and 6)a nd shorter (7)C NSL derivatives. Intriguingly, only C 8 compounds 18-22 showed ap romising anti-trypanosomala ctivity,w hereas 8-17 (C 15 )w ere not effective up to a concentrationo f2 00 mm.Notably,asignificantly higher efficacy was detected for 18-22 in comparison with 4,s uggesting that, at least fort his subset, the proposed hybridization strategy was successful. As expected by the predicted logP values (Table 1), the nature of the chain dramatically affected activity, probablyb ym odulating cell partitioning.T his might be due to the insertion of the alkyl chain on the phenoxy ring that might positivelym odulate the lipophilicity of the test compounds in-creasingc ell viability.I nf act, significant differences can be appreciated among the longer-a nd the shorter-chain subsets: the CNSL-based derivatives 8-12 and 13-17 showedn oa ntitrypanosomal activity against T. b. brucei up to 200 mm.

Anti-trypanosomal profile and time-to-kill determination
To determine whether 18-22 inhibited growth or cell division rather than causing cell death of T. b. brucei,c ell growth curves were performed by treating 427WT trypomastigotes with concentrations corresponding to 0.5 ,1 ,a nd 2 the EC 50 value, using untreated cells as ac ontrol.I ncubation with test compounds at half the EC 50 ( Figure 3A)c aused mostly some delayed growth phenotype,w ith rates increasing after 10 or 20 h. At 1 their respective EC 50 ( Figure 3B), the hybrid compounds inducedaconsistent, rapid-onset decrease in cell growth rate over 48 h. Compounds 18 and 19 showed trypanocidal activity at this concentration,s terilizing the culture in 2a nd 8h,r espectively.C ompound 22 appeared similarly to rapidly decrease the cell density but, at this concentration, killed only a proportion of the cell population before stabilizing; 20 and 21 decreased the growth rate substantially,w ith an early trypano-static effect over the first 12 h( Figure3B). At double the concentration ( Figure 3C), all the hybrids cleared the culture between the 4h and 8h time points,w itht he cell populations rapidly decliningb etween 2a nd 4h.T he rapid time to kill displayed by 18-22 (and the lack of cross-resistance with existing chemotherapy) is ac lear advantaget oward any (pre)-clinical development of this scaffold,a nd encourages furtherh it optimizationefforts.

Activityoft he compounds against T. brucei GAPDH
Based on the reported activity of starting compound 4,w hich displays an IC 50 of 7.25 mm against GAPDH, [14b] we first tested inhibition of this enzyme by 8-22,a tafixed concentration of 10 mm,f ollowing ap reviously reportedp rotocol. [14c] However, none of the compounds displayed significant inhibition of GAPDH activity (< 15 %d ecreaseo bserved, data not shown). One explanation for this lack of inhibitory activity might be that the current series of hybrids is partially or completely prevented from binding the active site of the trypanosomal glycolytic enzyme because of steric hindrance by the long alkyl chain. Limitationso fs olubility prevented us from testing the series at higherc oncentrations. However,i tm ustb en oted that we cannote xclude the possibility of al ow-affinity inhibition of GAPDH contributingt ot he trypanocidal effect,a sc ompounds may accumulate to relativelyh igh concentrations within the parasite. Indeed, examples where ah igh level of accumulation include almost all the first-line trypanocidal agents (Supporting Information Figure S1), including DA, [23] PMD, [24] melarsoprol, [25] and suramin [26] and highly activee xperimental therapies, including ar ecently described series of bisphosphonium compounds [27] that strongly accumulate in the T. b. brucei mitochondrion.

Mode of action studies
Because the drug designo fo ur series of hybridsi sb ased on the naphthoquinone framework that is recognizeda sap rivileged structure for the modulation of the mitochondrial functions [28] and 4 itself acted at mitochondrial targets through production of ROS, [14b] we had reasont of urther investigate their mode of action at the mitochondrial level in T. b. brucei. Furthermore, the chemical structures of the newly CNSL-based hybridsr esemble that of ubiquinone( Supporting Information Figure S2), which is an essential carrier in the electron-transport chain via ar edox reaction inner the mitochondrial membrane.A ccordingly,w ei nitially aimed foras ynergistic inhibition of the energy metabolism for our hybrids, targeting the mitochondrial membrane enzymesi nvolved in electron transport (e.g.,t rypanosome alternativeo xidase (TAO), F o F 1 -H + ATPase), in addition to glycolysis (GAPDH, located in the glycosome). [29] This because inhibition of both of the essential arms of trypanosomal energy metabolism would be expectedt od eliver synergistic effects.D espite the factt hat we were not able to demonstrateaGAPDHi nhibitory activity of our hybrids, we lookeda tt he effects of the two most active compounds 20 and 22 on ATPc ontent and on the mitochondrial membrane potential( MMP), as relevant parameters for an action on the trypanosomal energy metabolism. [30] ATPa nd MMP determinationa sr elevant evidence for a mitochondrial mode of action T. b. brucei 427WT BSF were incubated with 20 and 22,a ta pproximately 0.5 EC 50 ,a nd the ATPc ontent was determineda t differentt ime points ( Figure 4). The F o F 1 -ATPase inhibitor oligomycin (see Supporting Information Figure S2 for structure)w as used as positive controla nd untreated cells as negative control. We found that both compounds rapidlyd ecrease the ATP content,w ith even 30 min inducing ah ighly significant reduction in cellular ATPlevels (p < 0.01), stabilizing at approximately 50 %o fu ntreated [ATP] after 4-6 h( p < 0.001). The response to 20 and 22 was similar not justt oe ach other,c onsistent with an identicalm ode of action, but also highly similar to that of oligomycin, although the latter depressed the ATPl evels even furthera tt he respectivec oncentrations used (Figure4). It should be noted that the observed decrease in ATPd oes not correlatew ith cell death, which even at 2 EC 50 does not occur in large number until after 4h.T he observed decrease in ATP levels could be the result of mitochondrial functions and therefore we next investigated the timing of the same concentrations of 20 and 22 on the mitochondrial membrane potential Y m ,u sing flow cytometryw iththe fluorescent probe TMRE. Valinomycin (SupportingI nformation Figure S2) was used as the control for depolarizationa nd troglitazone (Supporting Information Figure S2), as the control for hyperpolarization ( Figure 5). [31] Both test compounds decreased Y m slightly over the first hour (p > 0.05), an effect that was stronger at the 4h(p < 0.05 for 20)a nd subsequent time points (p < 0.01 for both compounds). The Y m is expressed as the percentage of cells displaying af luorescence of ! 500 arbitrary units (calibrated at 50 %f or untreated cells at t = 0), with ashift toward lower fluorescence indicating am itochondrial membrane depolarization (Supporting Information Figure S3). It is clear from the time dependency and magnitude of the effects that cellular ATPd epletionp recededt he partial depolarization of the mitochondrial inner membrane. We thus conclude that the depolarization is likely to be the result of the decreased availability of ATPf or the F o F 1 -ATPase, which maintains the membranep otential of the mitochondrion of bloodstream T. brucei,u sing the ATPt o pump protons out of the mitochondrial matrix. [32] The primary cause of the ATPd epletion may be the inhibition of an essential mitochondrial function, and/or as tep in the glycolysis.
However,w er e-tested compounds 18, 19, 20 and 22 against4 27WT cells, in the presence and absence of 5mm glyceroli nt he medium, an addition that sensitizes the cells to TAOinhibitors such as salicylhydroxamic acid (SHAM) and ascofuranone (Supporting Information Figure S2). [33] Indeed, we found that the cells were significantly sensitized to SHAM in the presence of glycerol(p < 0.001), but not to the hybrid compounds (Table 2). Interestingly,c ompound 20 displayed significantly (p < 0.001)l ess activity in the presence of glycerol, whereas the EC 50 values for the other test compounds was unchanged. We conclude that 18, 19, 20 and 22 do not act via direct inhibition of TAO.
CNSL-basedhybridsi ncrease the production of reactive oxygen species (ROS) in T. b. brucei To determine whether our hybrids display ar edoxa ctivity in T. b. brucei 427WTB SF (similarly to what has been demonstrat-   (Figure 6). [34] The resultss how that under normal culture conditions trypanosomesg enerate asteady amount of ROS, which is greatly increasedi nt he presence of H 2 O 2 .T he level of ROS was dose-dependently increased over an incubation period of up to 2h with 20 and 22 ( Figure 6A,B, respectively). As suggested previously for 4, [14b] 20 and 22 could be substrates of the electrontransport chain, leading to cycle of reduction by glycerol-3phosphate dehydrogenase (G3PD) followed by reaction with molecular oxygen and the production of ROS. In normal conditions G3PD works like as huttle between glycosomes and mito-chondria, keepingt he redox balance and feeding the respiratory chain for the ATPp roduction, [35] with TAOa cting as electron acceptor, oxidizing the ubiquinol pool formed in consequence of G3PD activity. [36] Accordingly,t he observed overproduction of ROS is consistentwith an ubiquinol-like reactivity for our hybrids. [37] Thisw ould cause both moleculara nd structural damage to T. b. brucei mitochondria (see sectiono nT ransmission Electron Microscopy,below). [38] CNSL-basedhybridsc ause damages of the mitochondria but not the kinetoplast Transmission electron microscopy (TEM) was used to study the trypanosomes' ultrastructure after exposure to compound 22 at 1 EC 50 (7.6 mm)i no rder to visualize any damage to the mitochondrion. Based on the information obtainedf rom the ATP and Y m determinations, TEM samples were taken after 4a nd 8h exposure to test compound, as "early" and "late" time pointst od efine the effects on cellular ultrastructure ( Figure 7). After 4h of incubation with 20,s ome mitochondriap resented an irregular shape. Moreover,m embranous, electron-light structures had appeared in manym itochondria, resembling vacuoles( Figure 7, second rowofi mages);n ootherultrastructur-   al changes were evident, appearing to confirmt he mitochondrion as am ain target for CNSL-based hybridsi nt rypanosomes. Indeed, 20 caused further ultrastructure abnormalities in the mitochondrial structure at 8h,d isplaying damage to the mitochondrial membrane, and the presenceo fm embranous and dense vesicles within the matrix (Figure 7, thirdr ow). However, the treated cells do not display any morphological alteration to kinetoplasts,w hichc ompletely maintain the original disklike structure. [39] Accordingly, this evidences tands for an exclusive metabolic effect at mitochondrial level for the test compounds,w hich do not interferew ith both functioning and replication of kinetoplasts.
To confirm whether the CNSL-based hybrids exclusively target the mitochondrial energetic pathway rather than the mitochondrial kDNA, we decided to test compounds 4, 20 and 22 against the isometamidium-adapted ISMR1 cell line of T. b. brucei in parallel with the parental cell line 427WT.A s ISMR1 is ad yskinetoplastic cell line, and thus highly resistant to the drug isometamidium( ISM;S upporting Information Figure S2) and other kinetoplast-targeting drugs, [32] ap rimary effect on energy metabolism rather than mitochondrialk DNA should not resulti nr esistancet ot he test compounds. Indeed, compounds 4, 20 and 22 didnot show as ignificant increase in the EC 50 values relative to the standard wild-type strain 427, tested in parallel ( Figure 8). These data exclude kDNA as ap otential target for our test hybrids, with the availabled ata all consistentwith amitochondrialt arget.
Althoughw eh ad observed no toxicity of thesen aphthoquinone hybridst owardh uman cells, we next assessed whether they might damage mitochondriai nh uman cells. We therefore determined the ATPc ontent in treated and untreated HFF cell line with ad eliberately high concentration (200 mm)o f22.T he compound had only av ery minor effect on cellular ATPl evels, even at this concentration,b eing approximately 26 trypanocidal EC 50 ,a lthough the effect became significant at 8h of incubation (p < 0.05). Oligomycine xhibited as trong effect even at 4h(p < 0.0.1) (Figure 9).
We also used TEM after exposuret ot he same concentration of 22 (200 mm)a tt wo different time points( 4a nd 12 h) in order to analyze whether ultrastructural changes similart o those in T. brucei could be observed. The TEM images reveal that no morphological damage occurs to mitochondria in human cells after treatment with even high concentrations of 22,n or were any other ultrastructural changes in the exposed HFF cells visible (Figure 10). These resultsc learly show the absence of metabolic and structuralt oxic effects at the human mitochondrial level, indicating ah ighly species-specific mode of action for our CNSL-based hybrids.

Metabolomics analysis
Metabolomics hass uccessfully been used to elucidate the mode of action of drugs against T. brucei. [40] We therefore used am etabolomics approacht of urther investigate the mode of action of 22. T. b. brucei 427WT BSF were incubated with and without0 .5 EC 50 of 22 for 2h to identify specific metabolic changes causedb ym oderate exposure to the compound (early effects). Samples were instantly cooled down to block the intracellular metabolic activity, and processed to extract the pool of intracellular metabolites. Analysis by LC-MS showed al arge number of metabolite differences between the treated and untreated cells, butw ithout ac lear pattern desig-   nating ap articulare nzyme function or pathway, or indeed a particularc lass of metabolites being affected apart from some sphingolipid precursors which accumulatedi nt he exposed cells (Table 3), possibly indicating some direct or indirect interferencei nt his pathway.T hese results are consistentw ith a multi-target mode of action (as reportedf or 4) [14b] or the result of general damage such as caused by ROS, affecting many cellular functions at once.

Conclusions
In this report, we explore the use of aw aste product of food production,t he CNSL, as ac heap and abundant source of new anti-infectivea gents for use against NTD. The required chemistry is simple, accessible and easily scaled up, making the entire production process highly affordable and local. Taken together, the collectedd ata suggest that, in the case of compounds 20 and 22,t his CNSL hybridization strategy led to an increased trypanocidal activity up to 10-fold relative to the parentc ompound 4.T his did not apriori signify that the new compounds necessarily act similarly as 4,a st he fusion with the CNSLbased fragments affects both the pharmacokinetic and pharmacodynamic features of the synthesized hybrids. In addition to the contribution in target recognition,t he long alkyl chain clearly affects the physicochemical properties, which are determinantfor uptake by both target and host cells, delivery to mitochondria, and interaction with molecular targets.
We present clear evidencef or mitochondrial targeting in trypanosomes, especially the TEM pictures are convincing, with mitochondrial damage but no other ultrastructural changes observed after just 4hincubation with 1 EC 50 of 22.I tisq uite possible that this localized damage is linked to the observed production of ROS. However,A TP depletion precedes mitochondrial abnormalities and the depolarization of Y m can be observedw ithin 30 min, and we propose that the hybrids, as intended, exert more than one trypanocidal effect, consistent with the complex return of the metabolomics experiment,( althoughi tw as not possible to derive any clear conclusion from the metabolomics data, at this point). All this does not exclude at arget insidet he mitochondrion, but it is not (principally) TAOo rg lycerol-3-phosphate dehydrogenase( G3PDH), which together enable the re-oxidationo fg lycolysis-produced NAD + via the cyanide-resistant, non-protonmotive oxidation of ubiquinol by TAO. [41] This is because the activity of the hybrids was not enhanced in the presence of glycerol. The target linked to rapid ATPd epletion could also be in glycosomes but we saw no evidence of glycosomal damagei nt he TEM images, nor consistenti nhibition of GAPDH. Thec ombination of cellular damageb yR OS, while depleting the ATPr equired for damage repair,i sl ikely to contribute to the rapid cell death observed at just 1-2 EC 50 .
It is probablyb ecause the kinetoplast is not involved in the trypanocidal action of these hybrid compounds that there is no cross-resistance with diamidines( PMD and DA) and phenanthridine (ISM, ethidium)d rugs. As these are the mainstay of human (PMD) and especially veterinary trypanosomiasis treatment, this is an important "plus" for this compound series. Even more important,t he compounds displayed excellent in vitro selectivity over human cells;i na ddition to the lack of effect on growth rate of humanf ibroblasts, no effects on mitochondrial ultrastructure or ATPc ontent,e ven at very high concentrationso fc ompound were observed. This is especially encouraging in light of the severet oxicity of current HATchemotherapy (Supporting Information FigureS1). [42] In conclusion, while further rounds of optimization are required before these molecules can be turnedi nto valuable leads for trypanosomiasis, we have successfully demonstrated that CNSL hybridss howp romisea sat emplate for the development of anti-infectived rugs on as ustainable basis. Importantly,t he use of CNSL and its components for combating trypanosomiasis could be of high importance and economically feasible in developing countries that cultivatea nd process cashewa nd, at the same time, represent endemica reas. Further studies will focus on extending SAR, optimizingf or T. congolense,a nd testing on veterinary trypanosome species that are more closely related to T. brucei: T. evansi and T. equiperdum,w hich, being dyskinetoplastic, are not treatable with DA or ISM, andn ot geographically restricted to the tsetse belt of Africa. [6] Experimental Section Chemistry General:A ll of the commercially available reagents and solvents were used as purchased from Sigma-Aldrich and Te dia without further purification. The cardanol mixture (2)w as purchased from Resibras (Fortaleza, Brazil). Reactions were followed by analytical thinlayer chromatography (TLC), performed on precoated TLC plates (layer 0.20 mm silica gel 60 with af luorescent indicator UV254, from Merck and F 254 Silicycle plates). Developed plates were airdried and analyzed under aU Vl amp (UV 254/365 nm) or visualized by exposure to iodine stain. The oxidative cleavages were per-
The system was maintained at reflux for 4h.T hen, the collecting flask was changed and an ew extraction was carried out under the same conditions. In both cases, the solvent was evaporated under reduced pressure, providing 60 go fn atural CNSL as ab rown liquid in 40 %yield relative to the nutshell mass.
Mixture of anacardic acids (1) from natural CNSL:I naflask were added 30 go fn atural CNSL ( % 86.08 mmol), 15 gC a(OH) 2 (202.44 mmol) and methanol (180 mL) and water (30 mL). The reaction system remained at reflux and agitation for 3h.Then, the mixture was cooled to room temperature and filtered. The solid obtained was washed with ethyl acetate to remove the other components of the CNSL. The calcium salts formed were treated with 50 %h ydrochloric acid solution to pH 1.0 to liberate the mixture of anacardic acids, which were extracted with ethyl acetate (3 50 mL). The combined organic phases were washed with brine and dried over anhydrous sodium sulfate. After removal of the solvent under reduced pressure, the mixture was purified on as ilica gel chromatography column, eluted in am ixture of hexane and 20 % ethyl acetate, affording the mixture of anacardic acids (1)a sa brown oil in 70 %yield. 3-(8-Hydroxyoctyl)phenol (7):AnErlenmeyer flask containing asolution of 12 go fm ixture of cardanols 2 (monoene, diene and triene) (39.4 mmol) distilled acetic anhydride (12 mL) and phosphoric acid (12 drops) was placed inside an unmodified household microwave oven and irradiated for 3min (3 1min) at ap ower of 400 W. After,t he residue was extracted with ethyl acetate (3 15 mL) and the combined organic fractions washed with solution of 5% sodium bicarbonate (20 mL), 10 %h ydrochloric acid solution (20.0 mL), brine (20 mL), and dried over anhydrous sodium sulfate. After evaporation of the solvent at reduced pressure, the reaction mixture was purified by chromatography on as ilica gel (dichloromethane) affording the desired compound in 73 %y ield. Then, 10.00 go ft he mixture of acetylated cardanols was diluted with dichloromethane (20 mL) and methanol (20 mL) in ao zonolysis flask of 250 mL. The flask was adapted to the ozonator with as tream of ozone for one 1.5 h, in bath of dry ice/acetone. Next, the secondary ozonide was reduced with 5.9 go fs odium borohydride (158.7 mmol) in 60 mL of methanol. At the end of addition of sodium borohydride, the reaction remained for 6h under stirring. Then, the mixture was acidified with concentrated hydrochloric acid to pH 3, and it was extracted with ethyl acetate (3 30 mL). The combined organic fractions were washed with brine (30 mL) and dried over sodium sulfate. After evaporation of the solvent, the product was purified by chromatography on silica gel (dichloromethane/chloroform, 5:5a nd then chloroform/ethanol, 9:1) General procedure for the synthesis of CNSL-based hybrids (8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22):T oas olution of the proper phenol derivative (5-7)( 0.1 mmol and 0.3 mmol for 7)i nd ry DMF (0.04 m), K 2 CO 3 (0.1 mmol and 0.3 mmol for 7)w as added. The resulting mixture was stirred for 0.5 ha tr oom temperature. For phenol 7,t he reaction was conducted at 0 8Ct oa void the formation of side products. Subsequently,t he suitable 2-bromo-1,4-naphtoquinone (23-27)w as added (0.1-0.3 mmol) to the reaction mixtures, which were stirred at room temperature for further 2-3 h. Then, LiCl 5% solution was added (10 mL), whereas the reaction mixtures involving 26 and 27, were previously acidified with asolution of 2 n HCl (pH 5), until the chemical toning changes from blue-green to orange. Next, the obtained mixtures were extracted with ethyl acetate (10 mL 3). The organic extracts were collected, dried over Na 2 SO 4 and the solvent was evaporated under vacuum. The crude residue was purified by chromatography on silica gel and/or purified by crystallization, when required.  Methyl 2-((8-methoxy-1,4-dioxo-1,4-dihydronaphthalen-2yl)oxy)-6 pentadecylbenzoate (9):T he title compound was obtained according to the general procedure using 5 and 24,a nd purified by chromatography on silica gel with am ixture of ethyl acetate/petroleum ether/toluene (6:3:1), as eluent. Compound 9 was obtained as ay ellow waxy solid. Yield:6 0%. 1     Methyl 2-((8-hydroxy-1,4-dioxo-1,4-dihydronaphthalen-2-yl)oxy)-6-pentadecylbenzoate (11):T he title compound was obtained according to the general procedure using 5 and 26,a nd purified by chromatography on silica gel with am ixture of petroleum ether/ ethyl acetate/toluene (7:2:1), as eluent. Compound 11 was obtained as an orange-yellow waxy solid. Yield:6 5%. 1

8-Methoxy-2-(3-pentadecylphenoxy)naphthalene-1,4-dione (14):
The title compound was obtained according to the general procedure using 6 and 24,a nd purified by chromatography on silica gel with am ixture of petroleum ether/ethyl acetate (6:4), as eluent. Compound 14 was obtained as ay ellow waxy solid. Yield:6 5%.             (20):T he title compound was obtained according to the general procedure using 7 and 25,a nd purified by chromatography on silica gel with am ixture of ethyl acetate/n-hexane/toluene (7:2:1). Compound 20 was obtained as an orange-yellow waxy solid. Yield:3 4%.C ompound 20 was also synthesized from 7 (0.3 mmol), 25 (0.3 mmol), and K 2 CO 3 (0.3 mmol) in dry DMSO (0.04 m), and purified according to the above-reported procedure. Yield:3 1%. 1 potential PAINS, due to the presence of the quinone sub-structure and the long alkyl chain. However,d espite the potential for these compounds to interfere with non-cellular assays, [44] we found no activity against isolated TbGAPDH protein. Furthermore, the fact that the trypanocidal activity is limited to quinones with C 8 alkyl chain (18)(19)(20)(21)(22)p oints to ah igh degree of specificity and interaction with well-defined target(s) rather than an onspecific interaction. In addition, 20 and 22 showed severalfold lower trypanocidal activity when evaluated against T. congolense (IL3000 WT). These observations led us to believe that the current subset does not behave as PAINS.

Biological evaluation
Organisms and culture media:O nly bloodstream trypomastigotes of T. b. brucei were used throughout this study.T he drug-sensitive wild-type strain Trypanosoma brucei brucei Lister 427 (427WT) [25a] was used alongside three multidrug-resistant strains:B 48, ISMR1, and aqp2/aqp3-KO. B48 was created from 427WT after deletion of the TbAT1 gene, encoding for the P2 drug transporter, [25b] and adaptation to increasing concentrations of PMD. [45] The aqp2/aqp3-KO strain was generated from wild-type T. b. brucei 2T1 cells after knockout of the locus encoding for the aquaglyceroporin 2a nd 3 channels, resulting in melarsoprol-pentamidine cross-resistance. [21,46] ISMR1 is an ISM-resistant clone obtained from T. b. brucei 427WT that lost their kinetoplast DNA and express an F o F 1 -ATP synthase mutation. [32] All the T. b. brucei strains were cultured as described [47] in standard HMI-9 medium, supplemented by 10 %o f heat-inactivated fetal bovine serum (FBS), 14 mLL À1 of b-mercaptoethanol, and 3.0 gL À1 of sodium hydrogen carbonate (pH 7.4). Parasites were cultured in vented flasks at 37 8Ca nd 5% CO 2 atmosphere and they were passaged every 3days. Bloodstream forms of T. congolense savannah-type strain IL3000 were cultured in basal MEM medium, supplemented by 10 %f resh goat serum, 14 mLo fb-mercaptoethanol, 800 mLo f2 00 mm glutamine solution, and 10 mL of penicillin/streptomycin solution per liter of medium (pH 7.3). [48] T. congolense were cultured in six-well plates at 34 8C and 5% CO 2 .
In vitro drug susceptibility assay:T he drug susceptibilities of bloodstream-form trypanosomes of T. b. brucei 427WT,B 48, aqp2/ aqp3-KO and ISMR1 strains were determined by using the resazurin (Alamar blue) viability indicator dye following ap reviously described protocol, [49] slightly adapted for T. congolense. In brief, the assays were performed in 96-well plates with 2 10 5 cells per well for T. b. brucei or 5 10 5 cells per well for T. congolense,i nt heir respective culture media. First, 200 mLo ft est compounds' solutions (400 mm,i nT. brucei or T. congolense medium as appropriate) was added to the first well of each 12-well row,f rom which doubling dilutions were conducted over one row per test compound. Trypanosomes (100 mLi ne ach well) were added and the plates were incubated for 48 ha t3 7 8Ca nd 5% CO 2 ,f ollowed by addition of 20 mLA lamar blue solution (125 mg mL À1 of resazurin sodium salt (Sigma-Aldrich) in phosphate-buffered saline (PBS)) followed by 24 ho fi ncubation at 37 8Ca nd 5% CO 2 .P MD (Sigma-Aldrich) (for 427WT,B 48, aqp2/aqp3-KO), DA (Sigma-Aldrich;f or T. congolense) and ISM (gift from Merial France) (for ISMR1) were used as trypanocidal positive controls. Fluorescence was detected using aFLUOstar Optima (BMG Labtech, Durham, NC, USA) at wavelength of 540 nm (excitation), 590 nm (emission). EC 50 values were calculated by nonlinear regression using an equation for as igmoidal dose-response curve with variable slope using Prism 5.0 (GraphPad Software Inc., San Diego, CA, USA).
Growth curve: T. b. brucei 427WT cells were grown to mid-logphase in standard HMI-9/FBS medium distributed in in six-well plates at 1 10 5 cells mL À1 and incubated with two different concentrations of test compounds (EC 50, 2 EC 50 )f or 48 ha t3 78Ca nd 5% CO 2 .U ntreated parasites, used as control, were grown in parallel. The cells were counted by using hemocytometer cell counter (cell count/mL 10 4 )a t0 ,2 ,4 ,8 ,1 2, 24, 28, 32, 36, 48 ho fi ncubation. Each experiment was performed as two independent replicates, and each sample was counted at least twice.
Cytotoxicity assay on human foreskin fibroblast (HFF):T oxicity of test compounds to mammalian cells was carried out using the resazurin assay previously described [31] with slight modifications as follows. HFF cells were grown in ac ulture medium containing 500 mL of Dulbecco'sm odified Eagle's medium (DMEM;S igma), 50 mL newborn calf serum (NBCS;G ibco), 5mLp enicillin/streptomycin (Gibco), and 5mLo fl-Glutamax (200 nm,G ibco), at 37 8C and 5% CO 2 in vented flasks and passaged at 80-85 %o fc onfluence. For the cytotoxicity assay,c ells were suspended at 3 10 5 cells mL À1 and 100 mLa liquots were added to each well of a9 6well plate. The plate was incubated at 37 8Ca nd 5% CO 2 for 24 h to allow cell adhesion. Serial test compounds dilution was performed in ad ifferent 96-well plate and 100 mLo fe ach dilution was transferred to each well containing cells, resulting in exposure of cells to 200-0.2 mm of test compound. Phenylarsine oxide (Sigma-Aldrich) was used as positive control and drug-free incubation as negative control. The plates were incubated at 37 8Ca nd 5% CO 2 for an additional 30 h, at which point 10 mLo fr esazurin solution (125 mg mL À1 in PBS) was added, followed by af inal incubation for 24 h. The plates were read and the data analyzed as described above for the trypanosome assay.
ATPa ssay on T. b. brucei and HFF cells:C hanges in cellular ATP levels due to the exposure of trypanosomes to the test compounds (0.5 EC 50 )w ere monitored using the Molecular Probes ATPDetermination Kit (A22066, Invitrogen Detection Te chnologies), based on the luciferin-luciferase bioluminescent enzymatic reaction. Bloodstream trypomastigote cultures of T. brucei s427WT were incubated with and without test compound, and, at each predetermined incubation time, 10 7 cells of each sample were transferred into am icrofuge tube and centrifuged at 2000 g for 10 min at 4 8C. The pellet was washed twice with 1mLo f5 0mm Tris·HCl (pH 7.4) containing 0.1 mm DTT,a nd then resuspended in 200 mL of the same buffer.T he cells were lysed by sonication on ice (twice for 10 ss eparated by 30 s), using aS oniprep 150 (MSE) at 8 mm amplitude. The samples were centrifuged at 10 000 g for 10 min at 4 8Ca nd the supernatant was collected, instantly frozen in liquid nitrogen, and stored at À80 8C. Oligomycin (2.0 mgmL À1 )w as used as positive control. ATPl evels were quantified using the contents of the kit following the manufacturer's instructions;9 0mLo fs tandard reaction solution was added to each well of a9 6-well plate and the background luminescence was recorded in aF LUOstar OPTIMA fluorimeter;1 0mLo fe ach sample was then added to each well. The plate was incubated at 28 8Cf or 15 min and the luminescence was measured, including as tandard curve with an ATPc oncentration ranging from 1nm to 1 mm to allow the calculation of the ATPc oncentrations in each sample. The ATPc ontent was measured at 0, 0.5, 1, 2, 4, 8a nd 12 ho fi ncubation with test sample.
The same procedure was used to determine the ATPc ontent of HFF cells, using 200 mm of the test compounds for incubation times of 0, 4, and 8h,1 10 6 HFF cells per sample and ac entrifugation speed for pelleting the cells of 800 g for 5min. Oligomycin (2 mgmL À1 )was used as apositive control as described. [50] ChemMedChem Mitochondrial membrane potential assay on T. b. brucei: Changes in mitochondrial membrane potential (Y m )a fter incubation of trypanosomes with the test compounds were determined using fluorescence-activated cell sorting (FACS) with the indicator dye tetramethylrhodamine ethyl ester (TMRE), as described [34] with minor modifications. Cell suspensions of T. b. brucei 427WT trypomastigotes were incubated at 0.5 EC 50 with test compound and 1 10 7 cells were transferred at each time point (0, 0.5, 1, 2, 4, 8 and 12 h) into am icrofuge tube and centrifuged at 2000 g for 10 min at room temperature. The pellet was washed once in 1mL of PBS (pH 7.4) and resuspended in 1mLo fP BS containing 200 nm TMRE;the cells then were incubated at 37 8Cfor 30 min and subsequently placed on ice for another 30 min before analysis on a Becton Dickinson FACSCalibur using aF L2-height detector,a nd CellQuest and FlowJo software. Valinomycin (Sigma-Aldrich; 100 nm)a nd troglitazone (Sigma-Aldrich;1 0 mm)w ere used as controls for mitochondrial membrane depolarization and positive hyperpolarization, respectively. [51] Transmission electron microscopy assay on T. b. brucei and HFF cells:T EM of bloodstream trypomastigotes (427WT) and of HFF was performed essentially as previously described. [52] Briefly,c ell cultures were adjusted to 2.5 10 6 cells mL À1 and incubated in the presence or absence of test compounds at EC 50 .C ells were fixed overnight at 4 8Ci n2 .5 %g lutaraldehyde and 4% paraformaldehyde in 0.1 m phosphate buffer (pH 7.4). Samples were washed with 0.1 m phosphate buffer (pH 7.4), post-fixed in 1% osmium tetroxide for 1h on ice, and washed with the phosphate buffer. Cells were next incubated in 0.5 %u ranyl acetate solution for 30 min, washed with distilled water and dehydrated in increasing concentrations of acetone (30,50,70, 90 and 100 %). Cells were embedded in epoxy resin;t hin sections of 50-60 nm were observed in aT ecnai T20 (FEI) at 200 kV.
Reactive oxygen species (ROS) assay on T. b. brucei:T he production of ROS when T. b. brucei 427WT trypomastigotes incubated with different concentration (2.5 EC 50 ;1 .25 EC 50 ;0 .3 EC 50 )o f test compound was assessed using the indicator dye 2,7-dichlorodihydrofluorescein diacetate (DCFH-DH;S igma-Aldrich). This assay was performed in a9 6-well black-bottomed well plate;2 00 mLo f each test compound solution, at 5 EC 50 in assay buffer pH 7.3, was added to the well in the first column of the plate and doubling dilutions were carried out across the row,a fter which 3 10 6 cell in 100 mLo fa ssay buffer were added to each well, immediately followed by 2 mLo f1m m DCFH-DH, under minimal light conditions. The plates were incubated in aF LUOstar OPTIMA fluorimeter at 37 8C5 %C O 2 and the fluorescence was monitored at 485 nm for the excitation, 520 nm for the emission for 3h,t aking readings of each well every 2min. Three wells were included in the plate as controls:1 )3 10 6 cells per well in assay buffer without test compound;2 )3 10 6 cells per well in assay buffer with 2 mLo f1 0mm H 2 O 2 ;3 )assay buffer without cells and test compound.
Metabolomics assay in T. b. brucei:T he experiment was performed to detect changes in the intracellular metabolic pathway in T. b. brucei 427WT BSF after 2h incubation with the most active test compound, at ac oncentration of 0.5 EC 50 ,a nd compared with aD MSO drug free control. Parasites were grown to 1.5 10 6 cells mL À1 in order to have am id-log-phase culture at the end of the incubation time. At the end of the incubation, the cell density was adjusted and 1 10 8 cells were transferred to a5 0mLc entrifuge tube. The samples were quenched by rapidly cells cooling to 4 8Ci nadry ice/ethanol bath and centrifuged at 1250 g for 10 min at the same temperature. 10 mLo fs upernatant (spent medium) was collected, 200 mLo fC MW (chloroform/methanol/ water 1:4:1) was added and the mixture stored at À80 8Cu ntil LC-MS analysis. The rest of the supernatant was discarded;t he cell pellet was transferred to am icrofuge tube and centrifuged at 4500 rpm for 5min at 4 8C. The pellet was washed once in 1mL PBS at 4 8Ca nd resuspended in 200 mLo fC MW extraction solvent at 4 8C. The samples were shaken at 4 8Cf or 1h to break up the pellet and allow the complete extraction of intracellular metabolites, followed by centrifugation at 13 000 rpm for 10 min at 4 8C. 180 mLo fs upernatant was collected in an LC-MS vial. For each biological replicate, 15 mLf rom each sample were all combined in the same MS vial;t his pooled sample served as aq uality control. All samples were gassed with Ar before sealing and stored at À80 8C. All the samples were analyzed with LC-MS. The experiment was conducted in three independent biological replicates, each analyzed in duplicate (two technical replicates).
LC-MS analysis and data extraction:S amples were randomly placed in the autosampler tray and the LC-MS experiment was performed on an Accela 600 HPLC system combined with an Exactive (Orbitrap) mass spectrometer from Thermo Fisher Scientific (Hemel Hempstead, UK). In separate runs, 10 mLo fs ample was injected onto aHiChrom Ltd. (Reading, UK) ZIC-pHILIC column (150 4.6 mm, 5 mmp article size). The LC-MS system was run in binary gradient mode;aflow rate of 0.3 mL min À1 was used and samples were kept in av ial tray set at 3 8C. The gradient conditions were as follows:( A) 20 mm ammonium carbonate pH 9.2, (B) acetonitrile; 0min 80 %B;3 0min 20 %B;3 6min 20 %B ,3 7min 80 %B;4 6min 80 %B .T he ESI interface was operated in positive and negative ion switching mode, with + 4.0 kV of spray voltage for positive mode and À3.5 kV for negative mode. The temperature of the ion-transfer capillary was 270 8Ca nd sheath and auxiliary gas were set at 57 and 17 arbitrary units, respectively.The full scan range of both positive and negative modes was set at 75 to 1200 m/z with AGC target and resolution as Balanced and High (1E6 and 50 000), respectively.P rior to analysis, mass calibration was performed for both ESI modes using the standard Thermo Calmix solution. The mass spectrometry data was extracted by using m/z Mine 2.20 [53] and the masses were searched against an in-house database.