Histone deacetylase inhibitors with high in vitro activities against Plasmodium falciparum isolates collected from Gabonese children and adults

Histone deacetylase (HDAC) enzymes are targets for the development of antimalarial drugs with a different mode of action to established antimalarials. Broad-spectrum HDAC-inhibitors show high potency against Plasmodium falciparum, but displayed some toxicity towards human cells. Inhibitors of human HDAC6 are new drug candidates with supposed reduced toxicity to human cells and favorable activities against laboratory P. falciparum strains. We investigated the potency of 12 peptoid-based HDAC-inhibitors against asexual stages of P. falciparum clinical isolates. Parasites representing different genetic backgrounds were isolated from adults and children with uncomplicated malaria in Gabon. Clinical studies on (non-HDAC-inhibitors) antimalarials, moreover, found lower drug efficacy in children, mainly attributed to acquired immunity with age in endemic areas. Therefore, we compared the in vitro sensitivity profiles of adult- and child-derived isolates to antimalarials (HDAC and standard drugs). All HDAC-inhibitors showed 50% inhibitory concentrations at nanomolar ranges with higher activities than the FDA approved reference HDAC-inhibitor SAHA. We propose peptoid-based HDAC6-inhibitors to be lead structures for further development as antimalarial chemotherapeutics. Our results further suggest no differences in activity of the tested antimalarials between P. falciparum parasites isolated from children and adults.

use 7,[9][10][11][12] . Recent data suggest HDACi as attractive drug candidates against parasitic diseases caused by Trypanosoma, Toxoplasma, Schistosoma, Leishmania, and Plasmodium [13][14][15][16][17] . The natural product apicidin was the first HDACi to be tested against P. falciparum and exhibited broad-spectrum antiprotozoal activity in vitro and in vivo in mice 18 . SAHA (suberoylanilide hydroxamic acid, vorinostat), romidepsin, belinostat, and panobinostat are all clinically approved HDACi used for cancer treatment and affect growth of various Plasmodium species including drug resistant P. falciparum strains 15 . Notably, HDACi were shown to be active against multiple life-cycle stages of P. falciparum including liver stages and gametocytes 12,[19][20][21] . HDACi are promising lead structures for antimalarial drug development, but their use might otherwise be limited due to concomitant toxicity to human cells. This problem could be mitigated by developing inhibitors with relative or complete specificity towards plasmodial HDACs. In P. falciparum, up to six HDACs have been identified that share only a certain degree of sequence identity with human HDACs 22 . Little knowledge about structure and function of these enzymes in P. falciparum limits structure-based design of new inhibitors 23 . An alternative approach is to expand on human HDACi molecules, which are known to be less harmful to mammalian cells and drive their development towards parasite selectivity as well as anti-plasmodial activity. Selective inhibitors of human HDAC6 (hHDAC6), a class II enzyme, exert lower levels of cytotoxicity to human cells compared to HDAC class I inhibitors 24 . hHDAC6 targets in particular non-histone proteins (alpha-tubulin, Hsp90) and class II homologues that are also present in P. falciparum (PfHDAC2 and 3) [25][26][27] .
Based on this assumption, a series of peptoid-based HDACi were developed 5,6 . These compounds are classical HDAC inhibitors that have a cap-linker-zinc binding group structure with a peptoid-based cap group (N-alkyl glycine derivatives). Preclinical screens of these candidates identified potent activity against blood stages of P. falciparum laboratory strains 3D7 and Dd2 and against P. berghei liver stages with promising parasite selectivity indices 5,6 .
In vitro activity assessment of candidates against clinical P. falciparum isolates in early drug development can inform about the drug´s potency against parasite strains circulating in the target population in malaria endemic areas. Plasmodium parasites sampled from malaria patients are genetically very different from laboratory strains of P. falciparum that have been in in vitro culture for decades 28 . Additionally, the natural P. falciparum population is constantly exposed to host factors including antimalarial drug pressure and is therefore genetically highly diverse, and parasites may be intrinsically heterogenous in their susceptibility towards the molecule 29,30 . An additional layer of complexity results from clinical trials reporting different drug efficacies (of non-HDACi) against P. falciparum infections in adults and children [31][32][33] . These differences are mostly attributed to the partial immunity that is developed by the populations living in malaria endemic regions after multiple P. falciparum infections 34,35 . However, it has not been investigated if the parasites themselves isolated from children or adults show different drug susceptibility profiles in in vitro assays. Age-dependent immune responses that cause a difference in the number of P. falciparum strains co-infecting a single individual, also known as multiplicity of infection, could be one factor that provokes different susceptibility profiles in vitro.
Amongst the panel of previously published peptoid HDAC inhibitors, we selected 12 candidates with an IC50 below 100 nM against the laboratory strain 3D7 and different toxicity profiles for in vitro potency testing against P. falciparum isolates collected from infected individuals in Gabon, a country highly endemic for malaria 5,6,[36][37][38] . We furthermore investigated the susceptibility of P. falciparum parasites isolated from children and adults towards standard antimalarial compounds and compared their activity profile.

Results
In total, 85 clinical isolates were collected from 52 children and 33 adults with uncomplicated P. falciparum malaria in Gabon. Clinical isolates were tested for their susceptibility to 12 HDACi candidates, 1 approved HDACi cancer drug as comparator and 8 known antimalarial compounds. Of the 85 assays, 53 (33 from children, 20 from adults) tests fulfilled strict quality criteria for successful growth and were included into further analysis of the inhibitor concentrations. The median age (IQR) of children and adults included was 3 years (2-4 years) and 21 years (19-50 years), respectively. The median parasitemia (IQR) in children and adults was 25,000 parasites/µl (9,120-62,192 p/µl) and 3,933 parasites/µl (1,193 p/µl), respectively.
In vitro activity of peptoid-based HDAC inhibitors against laboratory and clinical P. falciparum isolates. We assessed in vitro activity of 12 peptoid-based HDACi candidates against P. falciparum isolates obtained from children and adults. The panel includes molecules from two generations of synthesis, no. 1 series (1a, 1d, 1g, 1h, 1i, 1u, and 1v) and no. 2 series (2c, 2g, 2h, 2i and 2m) (see Supplementary Fig. 1) 5,6 . Compounds were also tested against 3D7 laboratory strains to confirm activity of new compound production lots (Table 1). Compound 1 u was the most active HDACi candidate with a molecular activity of approx. 13 nM against P. falciparum strains isolated from both, children and adults (see Table 1). Compounds 1a, 1d, 1h, 1v, 2g, 2h, and 2i had good antiplasmodial activities with IC50 values in the double digit nanomolar range in children and adults. The IC50 values of drug candidates showed a wide range of activity ( Supplementary Fig. 2). Some HDACi (1a, 1g, 1h, 1v, 2c, 2g, and 2h) were at least 2-fold more active against the parasite strains obtained from children compared to adults' strains, but differences did not reach the level of statistical significance.

Comparing drug susceptibility of P. falciparum strains obtained from children and adults.
To investigate if parasites obtained from semi-immune adults are less fit and more susceptible to drug testing in vitro, assays were done with standard and well-characterized antimalarial compounds tested against P. falciparum isolates from children and adults. All compounds confirmed potency against Gabonese parasite strains including chloroquine ( Table 2). No significant differences in median IC50s between isolates from children and adults were observed for any of the drugs. www.nature.com/scientificreports www.nature.com/scientificreports/ determine the stability of a compound at study end. Drug instability was not observed in any of the standard antimalarials or SAHA. Except for 1h, 2g and 2m, all HDACi candidates were at least 1.6-fold less active at the study end when controlled with 3D7 ( Supplementary Fig. 3). Strains from children and adults were equally sampled over time (data not shown).

Discussion
The development of new drugs and particularly of those with novel targets and modes of action is urgently needed to compete with the development of resistance by P. falciparum to current antimalarials. So far, mainly broad-spectrum HDACi have been investigated as potential antiplasmodial drugs, since several human HDAC www.nature.com/scientificreports www.nature.com/scientificreports/ homologues have been characterized in P. falciparum. Anti-cancer HDACi, already FDA approved, have been tested against P. falciparum and have been shown to effectively kill the parasites in vitro and also showed high activity against P. berghei in in vivo mouse models at sub-micromolar concentrations, but are known to cause a certain level of toxicity to human cells 6,18,20 . HDACi affect multiple eukaryotic cell functions, including non-histone related pathways where class IIb HDACs (e.g. hHDAC6) are involved 24 . Homologues of hHDAC6 proteins have been identified in P. falciparum and targeting class II HDACs might be a possibility to circumvent toxicity in humans 24 .
We tested the 12 most promising peptoid-based HDACi candidates against clinical P. falciparum isolates from Gabon and confirmed previous results in laboratory strains which showed the HDAC6 inhibitor 1u, N-(2-(Cyclohexylamino)-2-oxoethyl)-N-(4-(hydroxycarbamoyl)benzyl)-4-isopropylbenzamide, to be the most potent HDACi with a mean IC50 of 4 nM and high selectivity towards the parasites 5 . Here, our results also showed compound 1u to be the most active HDACi against clinical P. falciparum isolates. The high drug activity of 1u may have been obtained by replacement of the N, N-dimethylamino group, from molecule 1h, with a less polar isopropyl group 5 . The results of our investigation in clinical isolates resemble those previously obtained in laboratory strains 6,12 , confirming that these compounds are equally active against a parasite population of high genetic diversity under selection pressure by the currently used antimalarial drugs 30,37 . All compounds were more active than the comparator broad-spectrum HDACi SAHA (vorinostat), which we found to have an IC50 similar to that reported in the literature 15,39 . HDACi candidates, including potent compound 1u, seem to suffer from chemical instability in solution that is less pronounced for SAHA. Loss in activity over time may also explain the observed wide range of IC50 data, that otherwise could be interpreted as suggesting towards intrinsic parasite resistance.
Differences in in vivo drug activities against P. falciparum infections between children and adults are usually attributed to the well-known acquisition of immunity after repeated infections in high endemic regions [32][33][34] . There is evidence that multiplicity and diversity of strains differs between adults and children [40][41][42] . However, to the best of our knowledge, there has been no study that specifically looked at differences in antiplasmodial in vitro drug activity between children and adults. Overall, our results suggest no significant differences in in vitro activity of HDACi and standard antimalarials in P. falciparum strains collected from children and adults. We found only amodiaquine to potentially show such a difference, a drug to which the parasite population is constantly exposed, although this finding did not hold under correction for multiple testing. Whether this difference is a true finding has to be further examined in future investigations. Overall, however, our results suggest no significant differences in the in vitro activity of either HDACi or standard antimalarials against P. falciparum strains collected from adults and children.
The standard antimalarials we tested were all highly active against clinical P. falciparum isolates from Gabon with no observed resistance except in chloroquine, to which 13 outliers showed an IC50 greater than 40 nM. First line treatment in Gabon changed from chloroquine to the artemisinin combination therapies (ACTs) artesunate-amodiaquine or artemether-lumefantrine in 2003 43 . Despite this change, chloroquine resistance appears to remain high, but a tendency towards declining chloroquine resistance can be extrapolated from the in vitro data [43][44][45][46] . This might indicate that the full reversal of chloroquine sensitivity can be observed in the future, as has occurred in other malaria endemic regions 47 . Only a few outliers were present for the other standard antimalarials, which may not necessarily be attributed to resistance. Amodiaquine and lumefantrine are first-line partner drugs to artemisinin-derivatives in Gabon, but fortunately signs of resistance cannot be detected 48 .
Our analysis of clinical P. falciparum isolates from Gabon confirm the results from previous work showing HDACi candidate 1u to be highly active and indeed more potent than the comparator cancer drug SAHA. These inhibitors targeting class II HDAC proteins are potential lead structures for further development as antimalarial chemotherapeutics with promising selectivity towards Plasmodium parasites, but require improvement of chemical stability. Differences in the immune status to malaria between adults and children seemed not to affect the observed drug potency against P. falciparum.

Methods
Clinical isolate sampling. In total, 85 participants with uncomplicated malaria were enrolled between October 2017 and June 2018 residing in Lambaréné and surrounding villages in Gabon. Informed consent was obtained from study participants or the legal representative, if minor. The study was approved by the Institutional Ethics Committee of the Centre de Recherches Médicales de Lambaréné (CERMEL) with the number CEI-CERMEL015/2015. Inclusion criteria were written informed consent, age either 1 to 5 years (children cohort) or 18 years and older (adult cohort), and P. falciparum monoinfection with a parasitemia above 1000 parasites/ µl assessed by Giemsa-stained thick blood smear. To obtain clinical P. falciparum isolates, a venous blood sample was taken in a lithium heparin tube/EDTA tube and processed in the in vitro drug sensitivity assay within approximately 6 hours. All methods were performed in accordance with relevant guidelines and regulations.
Drug sensitivity assay. Drug sensitivity assays were performed according to standard procedures 46 .
Briefly, 96 well-plates were pre-dosed with a threefold serial dilution of the respective drug in complete culture medium to obtain the following range of concentrations: chloroquine, 1.2 to 1000 nM; lumefantrine, 0.4 to 1000 nM; amodiaquine, 0.4 to 100 nM; piperaquine, 0.3 to 250 nM; pyronaridine, 0.02 to 50 nM; ferroquine, 0.08 to 200 nM; mefloquine, 0.7 to 500 nM; atovaquone, 0.02 to 20 nM; 1a, 4.1 to 10000 nM, 1d, 1.2 to 3000 nM; 1g, 12.3 to 10000 nM; 1h, 0.34 to 750 nM; 1i, 12.3 to 10000 nM; 1u, 0.34 to 750 nM; 1v, 2.7 to 2000 nM; 2c, 12.3 to 15000 nM; 2g, 0.34 to 750 nM; 2h, 1.2 to 1000 nM; 2i, 0.41 to 300 nM; 2m, 6.9 to 5000 nM; SAHA, 12.3 to 15000 nM. Ring-stage parasites from the laboratory strain 3D7 and clinical P. falciparum isolates were adjusted to a parasitemia of 0.05% with 0 + erythrocytes and the hematocrit was set to 1.5% in a total volume of 225 µl per well. After 72 hours of incubation at 37 °C, plates were freeze-thawed three times and analyzed by measurement of P. falciparum histidine-rich protein 2 (HRP2) with an enzyme-linked immunosorbent assay (ELISA) 51 . Only assays with successfully grown and propagating parasites reflected by a 1.5 OD increase between full and no inhibition within the 72 hours were included in the IC50 analysis. All experiments were done in duplicates. To control quality and stability, all compounds were additionally tested against the 3D7 laboratory strain once before and three times after testing of clinical isolates.

Statistics.
Individual inhibitory concentrations were determined by non-linear regression analysis of log-concentration-response curves using the drc-package v3.0-1 of R version 3.4.2. Data for the clinical isolates was presented using the median 50% inhibitory concentration (IC50) and the interquartile range (IQR). Correlations between IC50 values of clinical isolates of children and adults of the different standard antimalarials and HDACi were calculated using the Mann-Whitney U (nonparametric) test in JMP v14.0.0 software. IC50 values, calculated by R, to be over the highest tested drug concentration were substituted with the highest tested concentration value if parasites were dead at the highest tested drug concentration (indicated by a low OD value of confirmed dead parasites from the same patient).

Ethics approval and consent to participate. The study was approved by the Institutional Ethics
Committee (CEI) of CERMEL with the number CEI-CERMEL015/2015.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.