Final Report Summary - A-PARADDISE (Anti-Parasitic Drug Discovery in Epigenetics)
Executive Summary:
The aim of the A-ParaDDisE project was to feed the drug development pipeline for Neglected Tropical Diseases (NTDs) and in particular leishmaniasis, Chagas disease, malaria and schistosomiasis. These diseases collectively infect more than one billion people, mainly in developing countries, cause hundreds of thousands of deaths annually and represent a major economic burden. Treatments for all these diseases face problems due to developing resistance, severe side-effects and/or prolonged dosing schedules. The development of new drugs is a priority and we extended and expanded a strategy applied to schistosomiasis drug development in a previous EC-funded project, SEtTReND. This involved targeting enzymes and other proteins involved in the epigenetic processes that control gene transcription and in particular the modification of histones via acetylation/deacetylation and methylation/demethylation: histone modifying enzymes (HMEs). These enzymes are acetyl- (HAT) or methyltransferases (HMT), deacetylases (HDAC) or demethylases (HDM) and bromodomain-containing proteins that are readers of acetylated histones and recruit multi-protein complexes involved in transcriptional regulation. These enzymes are increasingly targeted in the therapy of a variety of pathologies including cancer, which allows us to benefit from the expertise and inhibitors already developed as starting points for developing anti-parasitic drugs.
We therefore aimed to identify HME inhibitors that will be lead compounds for drug development. We employed two complementary strategies: phenotypic screening using focused HME inhibitor libraries and a structure-based approach involving validated target enzymes and high-throughput or in silico screening. Bioguided optimization of the hits identified fed into lead development, basic ADMET and in vivo testing to generate lead compounds. The main achievements of the A-ParaDDisE project are as follows:
All the HMEs encoded in the genomes of the parasites studied have been identified. Phylogenetic analyses and molecular modeling allowed the choice of HMEs showing significant differences, particularly within the catalytic pocket, from their human counterparts, likely to facilitate the development of selective inhibitors. A total of 37 HMEs from Schistosoma mansoni, Leishmania braziliensis, Plasmodium falciparum and Trypanosoma cruzi were cloned.
- Transcriptomic studies (RNASeq) allowed the identification of gene expression signatures in parasites either treated with HME inhibitors or invalidated for a target HME gene/transcript.
- Gene knockout studies for 8 T. cruzi HMEs has shown that three (two HDACs and one HMT) are essential for parasite growth and survival. Knockout of one other HDAC does not affect viability but leads to a metacyclogenesis defect.
- Recombinant proteins have been obtained for a large number of HMEs. Notably two T. cruzi HMEs, TcHDAC2 (essential for parasite survival) and TcHAT1 have been produced as active enzymes following extensive protein engineering. Diffracting crystals of the former enzyme have been produced and structural analysis is ongoing. The S. mansoni arginine methyltransferase SmPRMT3 has also been produced as an active enzyme and crystallized.
- High throughput screening (HTS) of SmPRMT3 has been carried out. HTS using TcDAC2 has been submitted to the European Lead Factory and screening will start shortly.
- Bioguided inhibitor optimization was carried out for the essential S. mansoni HDAC8. A combination of enzyme inhibition assays, co-crystallization with the enzyme and phenotypic screening yielded compounds with increased selectivity and potency.
- Phenotypic screening was carried out on all the parasites studied in the project using focused libraries of HME inhibitors supplied by the chemist groups in the project. More than 500 compounds were tested and a large number (> 100) of hits identified.
- Pharmacokinetic and toxicity studies were carried out on the most promising hits and in vivo testing was carried out in mouse infection models for Plasmodium, T. cruzi and S. mansoni infections. Three compounds with significant in vivo activity have so far been identified and others are continuing to be tested.
The A-ParaDDisE project has fulfilled its principal objectives and the results obtained validate the strategy of targeting parasite epigenetic process for drug development.
Project Context and Objectives:
The disease burden imposed by parasitic infections worldwide is huge, and disproportionately affects poor countries in tropical and subtropical regions. Their impact is not only due to the mortality that they provoke, but also, more insidiously, to disease morbidity, which affects overall well-being of the populations concerned and is an underlying cause of persisting poverty [1]. The aim of our project was to develop novel drug candidates against four major parasitic diseases, schistosomiasis, leishmaniasis, Chagas disease and malaria, by targeting the enzymes responsible for epigenetic processes in the parasites. These diseases are among those defined as “Neglected Tropical Diseases” (NTDs) in the London declaration of 2012 “Uniting to Combat Neglected Tropical Diseases” (http://unitingtocombatntds.org/resource/london-declaration) which put the control or elimination of NTDs in the vanguard of global efforts to “chart a course toward health and sustainability among the world’s poorest communities to a stronger, healthier future”. A key element of the objectives set out concerns the development of new drug treatments in order to combat developing resistance against some treatments and to replace other treatments that are unsatisfactory due to significant side-effects, prolonged treatment regimens or are not effective at all disease stages. The drug development process is challenging and the attrition rate for new compounds entering the pipeline is high, adding to the cost of new drugs that successfully negotiate the difficult and expensive road from the initial identification of a “hit” during the screening process via lead optimization, pre-clinical testing and clinical trials. In order to produce more drugs, a large number of “hits” are required to feed the pipeline and this has been the fundamental aim of the A-ParaDDisE project.
The need for new drugs to treat NTDs
Our project focuses on parasitic diseases that all present problems for their control in endemic countries. There are no effective vaccines currently available and although some alternative control strategies exist, for example the use of insecticide-impregnated bednets in the case of mosquito-borne infections, drug treatments remain in the front line of disease control. There is only one drug available for the treatment of schistosomiasis, praziquantel, which is used for mass chemotherapy in sub-Saharan Africa, and besides increasing evidence for treatment failure, concerns have been raised about the real efficacy of this drug [2]. The treatment of choice for systemic leishmaniasis, liposomal Amphotericin B, has severe side-effects and is costly. Miltefosine has been used as an oral leishmaniasis treatment, but in India resistance to the drug is now a significant problem. The treatment of Chagas disease using benznidazole or nifurtimox requires long treatments, causes side-effects and it has recently been shown that benznidazole has no effect on the chronic, most clinically important, form of the disease. Finally, the increasing level of resistance to Artemesinin and Artemesinin-based combination therapies in south-east Asia, as well as reports of treatment failure in patients infected in Africa, has rendered urgent the introduction of alternative therapies.
Targeting epigenetic processes in parasites
The term “epigenetics” encompasses several types of heritable changes in gene expression that are linked to structural changes in chromatin, but without any modification of the DNA sequence. They include methylation of the DNA itself, variant histones, chromatin remodeling factors, non-coding RNAs and reversible, post-translational modifications of histone “tails”. The latter are numerous and include acetylation, methylation, phosphorylation, SUMOylation, ubiquitinylation, which, together with other “histone marks” constitute a complex “histone code”. This has a central role in regulating chromatin architecture, the binding of protein cofactors, the regulation of gene expression and its dysregulation in pathology. The enzymes that effect these modifications, “writers” and “erasers” of histone marks, and, increasingly, “reader” proteins that interact with them, are privileged targets for chemotherapeutic interventions. In particular, the writers, readers and erasers associated with histone acetylation and methylation have been under active investigation as potential drug targets for some time, particularly in cancer. Five FDA-approved inhibitors of histone deacetylases (HDAC) are now used in cancer treatment and a large number of other inhibitors of histone modifying enzymes (HME) or of bromodomain proteins, which interact with acetylated lysine residues, are in clinical trials.
Parasites are eukaryotic organisms and, as such, share common histone marks and epigenetic mechanisms with humans [3]. Moreover, HMEs have been shown to have vital roles in the growth and survival of parasites [4] and are therefore valid therapeutic targets. In the malaria parasite, Plasmodium falciparum, HDACs, histone acetyltransferases (HAT) and methyltransferases (HMT) have roles in the control of gene expression and cell cycle progression, and particularly in the control of variable surface gene expression involved in immune evasion by the parasite. In Trypanosoma brucei, a protozoan parasite of the Kinetoplastid family, which includes the agents of leishmaniasis (Leishmania sp.) and Chagas disease (Trypanosoma cruzi), both HDACs and HATs have been shown to be essential for growth and survival using gene knockout. Moreover, during a previous EC-funded project (SEtTReND) transcript knockdown by RNA interference was used to identify three HMEs, HDAC8, the lysine demethylase KDM1 and the arginine methyltransferase PRMT3 as valid targets in Schistosoma mansoni, the flatworm parasite causing intestinal schistosomiasis.
Inhibitors of HMEs, particularly HDAC inhibitors, have been shown to be effective against various parasites maintained in culture. For example, Trichostatin A (TSA) and the approved cancer drug Vorinostat are effective against P. falciparum and S. mansoni, but less so against Leishmania sp. [reviewed in 4]. However, these are pan-HDAC inhibitors that are active against all the isoforms of HDACs and inhibit human HDACs very effectively. The key to using HME inhibitors as a basis for new drugs against parasites is to develop compounds that are selective for parasites: i.e. that selectively inhibit a particular HME isotype and which also show species selectivity. We consider these criteria essential in order to develop candidate drugs that show minimal side-effects when used in humans. Moreover, the feasibility of this approach has been demonstrated by the development of HDAC inhibitors selective for P. falciparum that were effective in the low nM range against the parasite in culture and with selectivity indices >600 when tested on a macrophage cell line [5]. Similarly, selective inhibitors of class III histone deacetylases, the sirtuins, have been developed that are selectively active against L. infantum.
During the SEtTReND project, which focused on S. mansoni, we developed a target- and structure-based strategy to develop selective inhibitors as drug precursors. The most striking example of this strategy is the development of selective inhibitors of S. mansoni HDAC8 (SmHDAC8). This involved the crystallization of the recombinant protein produced in Escherichia coli, the solution of its 3D structure by X-ray crystallography, in silico screening of the catalytic pocket structure by docking compounds in a virtual library and the testing of selected compounds for their capacity to inhibit the recombinant SmHDAC8 in an enzymatic assay and for their ability to kill parasite larvae in a phenotypic assay [6]. This strategy was complemented by high-throughput screening of a compound library using recombinant SmHDAC8 and this latter strategy was also successfully used to identify inhibitors of S. mansoni sirtuin 2 (SmSirt2). Target and structure-based screening was also carried out utilizing in silico homology modeling of P. falciparum HDAC1 [5] to develop selective inhibitors that are promising drug precursors.
Objectives and strategy of the A-ParaDDisE project
The A-ParaDDisE project drug discovery strategy expanded the successful approach developed for the flatworm S. mansoni during the SEtTReND project to three protozoan parasites, P. falciparum, Leishmania sp. and T. cruzi. It combined structure-based approaches on selected, validated HME targets, with phenotypic screening using focused libraries of HME inhibitors. The “hits” identified by these screening steps then underwent a bioguided optimization process in order to improve their selectivity and activity against the parasite. The best compounds identified during this process were then subjected to toxicological studies and basic pharmacokinetic analysis before testing in vivo in mouse infection models. This strategy is summarized in Fig. 1. (see report attached).
This central strategy was supported by the specific principal objectives:
- The identification of HMEs from all four parasite species by data mining of the respective genome sequences. This was followed up by molecular and functional characterization of selected targets, target validation using gene knockout and transcript knockdown.
- Transcriptomic studies were carried out on parasites treated with HME inhibitors or knocked out for selected HMEs in order to determine the pathways affected by the treatment and to validate inhibitor selectivity.
- Phenotypic screening of all four parasites using focused HME inhibitor libraries.
- Production of recombinant parasite HME proteins for selected therapeutic targets in order to allow high-throughput screening (following enzyme assay development) crystallization and structure determination using X-ray crystallography.
- Optimization of inhibitor structures by chemical synthesis based on molecular modeling and structural studies. Bioguided feedback was used to direct optimization of inhibitor potency, selectivity and biological activity.
- Pharmacological and toxicological studies (in vitro and in vivo) were carried out on the best hit compounds derived from the screening/optimization process. Inhibitors showing high selectivity indices, satisfactory PK characteristics and absence of toxicity in mice were then tested in vivo in mouse infection models.
The A-ParaDDisE project has used an innovative and comprehensive strategy for drug discovery pooling resources and expertise garnered against different parasites by experts in their domains. (Project members listed in report attached).
The project partners brought different areas of expertise to the project including bioinformatics and target identification (2), biochemical characterization and transcriptomics (1, 4, 5, 7, 11, 15, 16), parasite phenotypic screening (1, 7, 11, 15 and 16), chemical synthesis (3, 6, 9, 12, 13 and 14), protein engineering and structural analysis (2), molecular modeling and in silico screening (3, 12), enzyme assay development (6, 9) high-throughput screening (9), pharmacokinetics, toxicology and in vivo testing (1, 7, 9, 10, 15) and project management (8). All of these aspects contributed to the overall success of the project.
Project Results:
(SEE FIGURES AND TABLES IN REPORT ATTACHED)
The drug discovery strategy of the A-ParaDDisE project involved both the direct screening of compound libraries composed of HME inhibitors (phenotypic screening), and the screening of individual HMEs, either in silico by docking compound structures into homology models of parasite HMEs, or by high-throughput screening of compound libraries. Both strategies have yielded compounds that are active against parasites maintained in culture that were fed into the drug development pipeline. However, this process involved a variety of steps necessary to the overall success of the strategy. This summary describes these steps, and the principal results obtained.
4.1.3.1. Structure-based screening strategy.
Two working hypotheses underpinned the possibility of targeting individual HMEs. First, individual HMEs would be valid therapeutic targets. Since these enzymes are members of protein families, the different isoforms may have overlapping functions, this requires verification. The second hypothesis is that such HMEs would have structural differences, particularly in the catalytic pocket, that would allow the development of inhibitors that are both isoform and species selective. The choice of individual HMEs to be screened as therapeutic targets was informed by several complementary approaches. These are summarized in Fig. 2
Bioinformatic genomic data mining was used to identify all the possible targets in the genomes of the studied organisms and molecular modeling of catalytic domains identified those that showed significant differences compared to the human counterpart. Gene or transcript knockdown (depending on the parasite studied) allowed the identification of HMEs that are essential to parasite development and survival, particularly in life-cycle stages that occur within the human host. Gene expression signatures, corresponding to the transcriptomes of parasites either treated with HME inhibitors or by HME gene knockout, characterize the biological functions of the parasite that are altered and define the molecular mechanism of action (MMA). Finally, phenotypic screening with HME inhibitors defined classes of HMEs that are promising targets.
4.1.3.1.a. Identification of the targets: mining parasite genomes.
In order to identify the possible HME targets in the parasites 6 species of kinetoplastids, which include 4 species of Leishmania, the Chagas disease agent, Trypanosoma cruzi and the agent of sleeping sickness, Trypanosoma brucei as a reference organism, the malaria parasite Plasmodium falciparum and 3 species of schistosome (Schistosoma mansoni, S. haematobium and S. japonicum) were included in the analysis. In the case of the schistosomes, improvements in genome assembly and annotation allowed an update of the analysis previously carried out during the SEtTReND project. In each case the predicted proteome database for each parasite was analyzed using InterProscan 5, which combines different protein signature methods into a single resource and allows functional analysis of proteins via family classification, domain and site prediction. The mass of data recovered in this way necessitated the construction of the A-ParaDDisE SQL (structured query language) database to allow data storage and recovery. This then became the main source for data-mining with a single table for each species proteome. To identify the HME proteins, significant matches with Hidden Markov Models (HMM) built for the catalytic domains of histone acetyltransferases (HAT) histone deacetylases (HDAC), histone methyltransferases (HMT) or histone demethylases (HDM) were searched using SQL queries. The final results for this search for the protozoan parasites are shown in Fig. 3.
This analysis shows that numbers of HDACs are similar for the protozoans (4 in kinetoplastids and 3 in P. falciparum, and this is also true for the GNAT class HATs with 10-13 in kinetoplastids and 9 in P. falciparum. However, the latter parasite has many fewer HMTs: 7 compared to more than 20 in the kinetoplastids and only one histone demethylase. This suggests that these enzymes may be promising therapeutic targets since their functional redundancy is likely to be reduced.
In addition to this analysis, the proteins containing bromodomains, which interact with acetylated lysine residues, were also identified in all the parasites studied using the same methodology. These turned out to be numerous (21-26 in Leishmania sp., 22 in T. cruzi, 29 in P. falciparum and 22 in S. mansoni) complicating their exploitation as therapeutic targets.
In addition to this enumeration, a phylogenetic analysis was carried out for the Zn2+-dependent class I and II HDACs and the NAD+-dependent class III HDACs, known as sirtuins, in order to identify the orthology relationships of parasite enzymes to their human counterparts and to determine which of them were the most divergent and therefore most likely to allow the development of selective inhibitors. Overall, as illustrated by Fig. 4, which shows the class IIa HDACs, schistosome enzymes were the most conserved. This is expected due to the relative evolutionary closeness of the metazoan platyhelminth parasite to humans, compared to the unicellular protozoans. Nevertheless, whilst certain schistosome HDACs, like the class I enzymes SmHDAC1/2 and SmHDAC3 are highly conserved, SmHDAC8 and all the class II enzymes are more distant. In contrast, all the kinetoplastid class II HDACs and some of the class I enzymes, are on long branches that diverge early in the phylogenetic tree and have no clear orthologous relationships with the human isoforms. This suggests that they may be promising targets.
The analysis of the class III HDACs, which bear no structural or phylogenetic relationship to the class I and II enzymes, and which have an NAD+-dependent catalytic mechanism, showed similar results in terms of the relative phylogenetic distances separating protozoan sirtuins from human isoforms. It also highlighted the absence of class IV sirtuins (Sirt6 and 7) from the kinetoplastids and of class II sirtuins (Sirt4) from schistosomes.
These analyses are the subject of a publication currently in revision [7]. Further to this work an in-depth phylogenetic analysis of Leishmania HMEs has also been carried out.
4.1.3.1.b. Construction of homology models of parasite HDACs
The first X-ray structure of a parasite HDAC, SmHDAC8, was obtained during the SEtTReND project. However, information was lacking on the structures of the HDACs from other parasites. For this reason, homology models of HDACs from Leishmania, T. cruzi and P. falciparum were prepared and compared to HDACs from S. mansoni, other flatworms and humans [8]. In addition, docking and molecular dynamics studies were carried out using known HDAC inhibitors. The general structure of class I HDACs is shown in Fig. 5.
Docking studies showed that, overall, the residues coordinating the Zn2+ ion are perfectly conserved in all the HDACs studied, and residues coordinating the warheads of inhibitors are also conserved. Differences are observed in the lower part of the substrate-binding tunnel and at the entrance of the so-called foot pocket (Fig. 6), which is not always present, and particularly in the walls of the binding pocket tunnel and the entrance of the side-pocket. In this way, a catalogue of differences has been compiled that could be exploited in the design of selective inhibitors.
Docking studies with HDAC inhibitors with reported activity against the parasites supported the mechanism of action of these compounds as acting selectively on HDACs. For example, HDAC inhibitors such as SAHA, Trichostatin A and the selective HDAC6 inhibitor Tubastatin A, as well as other compounds, could be docked to homology models of L. braziliensis and L. major HDACs [8]. However, this was not always the case. The HDAC8 selective inhibitor PCI-34051 could be docked into the homology models of Leishmania HDACs 2 and 4 (Fig. 7), but not HDACs 1 and 3. A homology model of the T. cruzi HDAC2 (TcDAC2) was used for an in silico docking screen to identify potential inhibitors (see below).
Interestingly, all inhibitors developed against SmHDAC8 could be docked into platyhelminth HDAC8 homology models and the docking poses were similar due to the high level of similarity of the binding pockets. This suggests that the inhibitors developed against the schistosome enzyme might also inhibit the orthologues in other schistosome species, as well as those in Clonorchis sinensis and in the Echinococcus, Taenia and Hymenolepis genera.
4.1.3.1.c. Target validation and characterization
In order to validate individual HMEs as therapeutic targets, we have carried out gene or transcript knockdown studies, characterized in detail the functional activity of certain HMEs and investigated the effects on gene transcription in parasites either treated with HME inhibitors or knocked down for an HME in order to elucidate molecular mechanisms of action.
i) Target validation
During the previous SEtTReND project, transcript invalidation of a large number S. mansoni HMEs using RNA interference (RNAi) showed that three were essential to the survival and/or development of the parasite within its mammalian host. These were SmHDAC8, the histone demethylase SmLSD1 and the arginine methyltransferase SmPRMT3. Studies on these three HMEs have continued during the A-ParaDDisE project.
Of the 50 HMEs encoded in the T. cruzi genome (Fig. 3), 8 were selected for gene knockout studies. These include the four HDACs, two HATs the methyltransferase DOT1A and a JmjC family histone demethylase. Gene knockdown was carried out principally using a classical genetic recombination strategy. Briefly, an initial knockout in the first allele was done using a construct containing flanking regions of the targeted gene, intergenic regions of two genes and a selectable marker for hygromycin resistance. Confirmed heterozygous knockout parasites were then transfected with a second construct carrying a different selectable antibiotic resistance marker. Homozygous knockouts were selected using both antibiotics.
In this way it was demonstrated that two HDACs, TcDACs 1 and 2, as well as the histone methyltransferase TcDOT1A, were essential for parasite survival and development: it was not possible to obtain viable homozygous knockout parasites. In the case of a further HDAC, TcDAC4, homozygous knockout parasites were viable. However, they did show morphological defects during metacyclogenesis (Fig. 8), indicating that TcDAC4 does fulfill a distinct specific role in T. cruzi. The viability of the TcDAC4-/- parasites allowed the investigation of their transcriptome using RNASeq (see below).
Following the publication of a methodology allowing the application of CRISPR/Cas9 gene invalidation to T. cruzi [9] this is also being applied to obtain knockouts of HME genes. This work will be completed after the end of the A-ParaDDisE project.
Whilst similar studies were not done on Leishmania parasites, a drug target deconvolution study has been undertaken for a compound that was a hit in the screening assays. This compound is tested at increasing concentrations on L. donovani promastigotes transfected with a homologous cosmid library from virulent hamster-derived parasites. Resistant parasites have been isolated and selected cosmids sequenced. They are now under analysis for the identification of the potential target.
ii) HME functional activity
The role of histone modifications in kinetoplastid parasites is not known, particularly since these organisms carry out genome-wide transcription producing polycistronic mRNAs, which are then divided by trans-splicing. The control of protein expression seems to occur mainly at the post-transcriptional level. However, histone marks do exist and as part of the A-ParaDDisE project a mass-spectrometry-based proteomic study was carried out to identify them in T. cruzi [10]. A total of 13 different chemically-distinct modifications were characterized, including some novel and unusual marks such as lysine acylations, Ser/Thr acetylation and N-terminal methylation. Some of the marks correlate in nature and position to those in other eukaryotic organisms, suggesting the existence of similar regulatory mechanisms. Others are unique to trypanosomatids and the enzymes responsible may be drug targets.
Functional characterization of T. cruzi HMEs has involved their subcellular localization using fusions with fluorescent proteins and tags, or immunolocalization. These analyses showed that most are nuclear as expected. However, a cytosolic localization for one HAT was detected, suggesting that it acts on targets other than histones. The analysis of the phenotypes of the T. cruzi HME mutants also gives insights into their functions. In the case of TcDAC2, the alterations include cell division defects, and DNA accumulation.
The functional study of S. mansoni HMEs was focused on SmHDAC8 and SmLSD1, both of which had been shown to be essential to the parasite by RNAi treatment. The use of two LSD1 inhibitors showed that these induced tegumental and muscle abnormalities as well as the presence of apoptotic and necrotic vesicles and granules. Adult schistosomes treated with these inhibitors also showed extensive alterations to their reproductive systems. Death of the parasites after treatment was due to the induction of apoptosis as confirmed by the expression of caspases 3 and 7 and the TUNEL staining of DNA double-strand breaks. That these alterations were due to the inhibition of LSD1 was shown by the resulting hypermethylation of histone 3 at lysine 4 (H3K4), the specific target of this enzyme. This study was completed by RNASeq analysis of inhibitor-treated parasites.
The precise biological roles of HDAC8 remain enigmatic. Relatively few partner proteins and substrates have been identified. We therefore undertook a study to identify partner proteins of SmHDAC8 using a yeast two-hybrid (Y2H) strategy. Using full-length SmHDAC8 as bait a large number of potential partners were identified corresponding to biological processes ranging from cytoskeleton organization to oxidative stress. The interaction of human HDAC8 with actin in the cytoskeleton of smooth muscle cells has already been described [11] and we are therefore investigating more closely the interaction of SmHDAC8 with cytoskeletal regulatory proteins identified in the Y2H experiment. A further function for HDAC8 that has been described concerns the deacetylation of the cohesin complex component SMC3 [12]. The cohesin complex is involved in mitosis, gene expression and DNA repair, forming a “lasso” around DNA strands, and the deacetylation of SMC3 is essential to the release of DNA strands by the complex. SmHDAC8 also deacetylates SMC3 and interacts with another complex component, suggesting that it is an integral part of the complex.
iii) Gene expression signatures
Transcriptional analysis (RNASeq) has been used in order to determine gene expression signatures in three parasites, S. mansoni, T. cruzi and L. braziliensis, under different conditions. The principal objective is to elucidate the processes and signaling pathways affected by treatment of the parasites with HME inhibitors and to determine their molecular mechanisms of action. RNASeq has also been performed following the invalidation of a T. cruzi HDAC with the aim of determining its function.
At the start of the project an RNASeq analysis pipeline was developed in order to analyze the RNASeq data obtained from S. mansoni worms treated with HME inhibitors. The pipeline applies the de novo assembly of transcripts with Trinity tools, without aligning the reads to the genome. Using unpublished S. mansoni male, female and egg RNASeq data previously acquired, gene expression signatures were obtained for these life-cycle stages. These results, highlighting 238 novel protein coding genes and extending the sequences of hundreds of predicted genes already annotated in the S. mansoni genome, were published [13]. Further to this, a local genome browser database was implemented at the University of Sao Paolo (http://schistosoma.usp.br/folders/geneNetwork/) incorporating the genome sequence, annotated genes, transcripts and ChIPseq histone marks and made public.
RNASeq analysis methodology was further improved and automated in order to introduce the use of different quantification tools. It is considered necessary to run different tools on the same dataset and consider only genes listed by all the tools (the main intersection in a Venn diagram) in the final analysis. After exhaustive tests of different programs, the final pipeline is composed of the following steps (some tools used are in parentheses): data download (ftp), data checking (md5sum), data decompression (gunzip and cat), metadata preparation, statistics about the reads (FastQC), adapter and quality trimming (Trimmomatic), read mapping (Tophat2), visualization of the mapped reads (UCSC Genome Browser), counting indexing, counting quantification (Cuffquant, Sailfish, and Kallisto), differential expression analysis (Cuffdiff, EdgeR, and Sleuth), comparison between sets of DEGs, gene enrichment analysis (Cytoscape with Bingo, Revigo, and Ontologizer) and finding normalizer genes. It is important to emphasize that for counting quantification and differential expression analysis steps three sets of tools were used: Metagenes (consolidated transcripts from publicly available RNASeq reads which represent a more complete transcriptome than the Sanger Centre Smp database), the new genome browser and hierarchically clustered heatmaps of differentially expressed genes.
These tools were applied to S. mansoni transcriptomes analyzed after treatment with the pan-HDAC inhibitor, Trichostatin A (TSA), the histone methyltransferase inhibitor GSK343, the SmHDAC8 selective inhibitor TH65 and the LSD1 inhibitor TB7/36.
In schistosomes treated with TSA, the subject of an article currently in press [14], histone acetylation at H3K9 and H3K14, indicating transcriptional activation, was increased while H3K27me3 levels (transcriptional repression) were unaffected. The effect of TSA on gene expression in schistosomula was tested at three different time points and found a marked genome-wide change in the transcription profile. ChIPSeq public data was combined with the RNASeq data and showed that differentially expressed genes with the H3K4me3 (activating) mark at their promoter regions generally showed transcription activation on TSA treatment, whereas those without this mark were mostly down-regulated. The affected processes include an upregulation of DNA replication genes, a down-regulation of 20 out of 22 genes involved in the accumulation of reactive oxygen species and changes in the expression levels of dozens of genes encoding proteins with histone reader motifs. These included SmEED, an orthologue of a member of the PRC2 transcriptional regulatory complex. This complex contains the histone methyltransferase EZH2 and we therefore tested the highly selective EZH2 inhibitor GSK343 for its activity on the parasite. This inhibitor showed a synergistic effect in combination with TSA: a sublethal dose of TSA (1 µM) combined with a sublethal dose of GSK343 (20 µM) led to significant schistosomulum mortality after 3 days in culture. Moreover, when tested alone, GSK343 caused substantial damage to the adult male and female worm teguments and eggs released by treated female worms showed eggshell fissures and were not viable.
Transcriptomes of adult worms and schistosomula treated with the sub-lethal dose of 20 µM GSK343 showed a sex-specific response. Three analysis tools were used and only genes identified by all three as showing expression variations were analyzed. Whilst 607 female and 580 male transcripts were downregulated, and 412 and 532 were upregulated in females and males respectively (Fig. 9), only 54 downregulated genes and 86 upregulated genes were in common. In male worms genes related to membrane constituents, cell homeostasis and drug transport were downregulated. In females, genes involved in egg formation were downregulated, coherent with the effects of the drug on eggs. Interestingly, EED was also downregulated in female worms, as was the case after TSA treatment, as was a key gene involved in stem cell differentiation. These data underline the potential of targeting EZH2 in schistosomes.
tified by all three methods (cutdiff, Sleuth and EdgeR) were retained in the analysis.
In order to help determine why SmHDAC8 is essential for schistosome development and survival in the mammalian host, the selective inhibitor TH65 (see below) was used to treat adult worms and their transcriptomes were then analyzed. At a dose of 20 µM, this inhibitor caused both tegumental damage, including blebbing and loss of tubercles, and effects on the reproductive organs: notably release into the medium of vitelline cells. Due to these severe effects, three doses of TH65 were tested: 5, 10 and 20 µM. Table 1 shows the dose-dependent increase in the numbers of differentially expressed genes.
SmHDAC8 seems to regulate common pathways in both sexes since about 50% of the differentially expressed genes are common between females and males. In female worms genes related to egg development and transmembrane ion transport are the most striking, corroborating the phenotypic effects of TH65 treatment on egg morphology and viability. In males, genes encoding tegument surface proteins are downregulated, again in line with the phenotypic effects of drug treatment. On the other hand, upregulated genes in both sexes included those related to nucleic acid binding, ribosome biogenesis and RNA processing, flagging SmHDAC8 as an important actor in transcriptional and post-transcriptional control of expression. Interestingly, genes encoding proteins identified as partners or substrates of HDAC8 were not affected.
RNASeq analysis of adult worms treated with the LSD1 inhibitor TB7/36 showed that 2709 genes were differentially expressed in males and 938 in females. In addition, 270 genes were differentially expressed in schistosomula treated with the inhibitor. Downregulated genes in male worms included those involved in drug pumps and cytoskeleton formation, whereas in females transmembrane transporter genes were downregulated. Notch signaling pathway genes were upregulated in female worms and schistosomula. Notch pathways have multiple roles including in neuronal function and development, but also in the regulation of cell fate decisions in various human tissues and the role of HDAC8 in maintaining Notch1 stability was recently reported [15].
At this stage of the analysis of the results obtained with different HME inhibitors in schistosomes it is too early to draw hard and fast conclusions. However, some common effects of inhibitors of different enzymes can be identified:
- All the inhibitors studied have striking effects on the morphology of female reproductive tissues and the eggs. This may be partly due to the fact that, in adult worms, these tissues are the most metabolically active and show the most intense cell proliferation. However, these effects are correlated with the downregulation of genes involved in eggshell formation and transmembrane transporters.
- Effects on signaling pathways and genes involved in cell differentiation and cell fate determination are affected. This suggests that HME inhibitors affect parasite development pathways. The effects of TH65 on gene transcription can be correlated with the effect of SmHDAC8 knockdown on schistosome development and survival in mice.
In addition to transcriptomic experiments undertaken on schistosomes, studies were also carried out on HDAC4-/- T. cruzi and on L. braziliensis parasites treated with an HME inhibitor.
RNASeq results indicate that several genes are differentially expressed when TcHDAC4 is not present. More than 50 different genes are up-regulated in T. cruzi HDAC4 null mutant epimastigotes and epimastigotes under nutritional stress, suggesting that, in these stages, histone deacetylation by TcHDAC4 is important to maintain the expression of some genes at a low level. Later on in the metacyclogenesis process (attached differentiating epimatigote cells), lacking of TcHDAC4 causes both up- and down-regulation of almost 2000 genes. Preliminary analysis also indicates that a number of surface protein genes are differentially expressed in T. cruzi HDAC4 null mutant cells. As these surface proteins play a direct role in the T. cruzi infection process, our data indicate that HDAC4 should also be considered as target for T. cruzi drug design to block expression of surface proteins. These results also show that an individual HDAC controls the expression of a specific gene subset, although whether this control is exerted at the transcriptional or post-transcriptional level, via effects on mRNA stability for example, remains to be determined.
We have adopted the Dual-RNA-Seq approach [16] in order to evaluate the effect of two selected HME inhibitors on the gene expression of both Leishmania braziliensis (M2904/GFP strain) and murine macrophages (RAW 264.7). Macrophages were cultivated for 24 hours before infection with Leishmania parasites. After 24h of infection, treatments were performed for 48 hours with either the HDAC inhibitor TH85, the HMT inhibitor BSF2 or with vehicle (DMSO) for the control. After treatment, Leishmania-infected macrophages were collected, total RNA from the cells was extracted, quantified and checked for quality and integrity. The experiments were performed with 4 biological replicates each for a total of 12 samples. The extracted RNAs were of high quality and were used for the generation of Illumina cDNA libraries. The sequencing was very balanced, such that approximately 60 million reads were obtained for each sample. Preliminary mapping analysis showed that around 1.5% of the reads from each sample map onto the genome of Leishmania (approx. 1 million reads for each sample). Differential gene expression analyses are being performed now and will help to elucidate the role of histone modifying enzymes on gene expression programs in Leishmania.
4.1.3.1.d. Production of recombinant HMEs and structural determination
Recombinant HMEs have been produced with two main objectives: to supply enzymatically active proteins to allow high-throughput inhibitor screening and to produce protein crystals for X-ray crystallography and the solution of the 3D structures of HMEs alone or in complexes with inhibitors. This in turn would permit the in silico screening of virtual compound libraries. Moreover, the study of crystal structures of inhibitors bound to the active site of SmHDAC8 has been instrumental in the development of selective inhibitors of this enzyme.
i) Production of recombinant HMEs
During the project the production of a variety of HME recombinant proteins from all the parasite species studied has been attempted. These include 7 HDACs, one sirtuin (class III HDAC), one HAT, two PRMTs and two HDMs. Some of these proteins proved difficult to produce in quantity or in an intact form, for instance the S. mansoni LSD1 protein. The production of bromodomain-containing proteins was put on hold in the absence of their validation of targets. In addition, the possibility of reconstituting complexes between target HMEs and regulatory subunits has been investigated. The main motivations for this were: (i) the stabilization of HMEs through complex formation and (ii) the reconstitution of a biological context for the exploration of HME activity and inhibition. However, a major hurdle has been the correct identification of the partner subunits, for which clear orthologues of human counterparts may not exist, particularly in the case of protozoan parasites. Stabilization of HMEs has also been explored through protein engineering: using shortened constructs, point mutations and the removal of insertions. Many of the parasite HMEs contain insertions within their catalytic domains.
The principal HME proteins produced were:
Arginine methyltransferases (PRMTs). Two S. mansoni PRMTs have been studied, SmPRMT1 and SmPRMT3. The latter was validated as a therapeutic target using RNAi transcript knockdown and was produced in a catalytically active form. This protein was then used for high-throughput inhibitor screening. The production of diffracting crystals proved difficult, however, and current efforts are directed at the coexpression of SmPRMT3 with a putative protein partner, the ribosomal protein RPS2.
Lysine demethylases. Following difficulties in obtaining SmLSD1 in an intact form, another schistosome histone demethylase, SmUtx was studied. This enzyme is the only schistosome demethylase capable of demethylating H3K27me3, a key repressive histone modification, and this enzyme may represent a uniquely sensitive target in the parasite. This has been successful (Fig. 10), but it has not so far been possible to measure the enzyme activity of the recombinant protein, perhaps due to the incompatibility of the human substrate used.
Histone acetyltransferases. The T. cruzi histone acetyltransferase TcHAT1 has been produced in a stable form following protein engineering to remove some insertions that form unstructured loops. The resulting recombinant protein is enzymatically active and has been used for tests of potential inhibitors.
Histone deacetylases. Production of the enzymatically active S. mansoni sirtuin SmSirt2 has continued with the aim of exploring the selective inhibition of this enzyme. Concerning the zinc-dependent class I and II HDACs the main strategy has been to expedite the search for anti-parasitic drugs using FDA-approved scaffolds for designing selective drug leads. With this aim the production of large quantities of several HDACs has been carried out for work both within and outside the A-ParaDDisE consortium.
A major effort has gone into producing L. braziliensis HDAC3. The full-length protein expressed on its own poses problems of degradation and aggregation. The removal of a large insertion within the catalytic domain solved some of the problems of degradation and increased yields. However, its propensity to aggregate was not diminished. Putative protein partners (SANT-domain containing proteins) were identified and cloned, but their coexpression with LbDAC3 did not solve the problem and work on this protein was stopped.
Work on the selective inhibition of SmHDAC8 has continued and intensified during the A-ParaDDisE project. Very large quantities of SmHDAC8 have been produced and have been used for biophysical and crystallization studies, as well as for inhibitor testing (determination of IC50 values). The enzyme has also been distributed to groups outside the consortium in collaborations on the selective inhibition of SmHDAC8. Mutational analysis has also been carried out in order to determine the eventual importance of catalytic site loops. It should be noted that the work on the selective inhibition of SmHDAC8 has required the comparison with human HDAC8 (hHDAC8). Large amounts of the latter have also been produced for work within and outside the consortium.
A potential therapeutic target of major interest identified during the A-ParaDDisE project is HDAC2 from T. cruzi (TcDAC2). TcDAC2-/- parasites are not viable, showing that this enzyme is essential to survival. It was initially produced in as a full-length protein in an enzymatically active form (Fig. 11). This protein was used for initial studies of its activity and inhibition but its crystallization proved difficult to reproduce. A systematic study combining protein engineering with biophysical studies (see below) have yielded a protein with markedly improved stability that has been used successfully for reproducible crystallization.
ii) Biophysical studies
The simple use of SDS-PAGE analysis of recombinant proteins to assess degradation and yields is insufficient as a predictor of their enzymatic activity, inhibition and stability with a view to their crystallization. Various biophysical methods were therefore brought to bear and, although these methods sometimes generated different results, the integration of the analyses enabled choices of constructs and inhibitors for further structural characterization:
- Dynamic Light Scattering (DLS). This has been used to assess the aggregation state of HMEs and their possible stabilization through the use of small molecules such as inhibitors.
- Thermal Shift Assay (TSA). TSA provides similar, but more precise information than DLS and has been use to assess the stability of different HME constructs produced. It has also been used extensively to determine the (de)stabilization of HMEs by mutations and inhibitors.
- Isothermal Titration Calorimetry (ITC). This has been used to determine thermodynamic parameters of the binding of inhibitors to HMEs (Fig. 12). The Kd values obtained by ITC could be compared to IC50 values measured in order to show up similarities or dissimilarities. ITC was employed both for mutated and non-mutated proteins, providing a fine measure of the effects of the mutations on inhibitor binding. This technique is complex to set up, however, meaning that only a subset of HMEs, mutants and inhibitors was considered.
iii) Structural studies of HMEs and their inhibitors
The structural characterization of HMEs and HME/inhibitor complexes has been a major objective of the A-ParaDDisE project. This aspect has depended on the success of protein production and the support of biophysical studies.
SmPRMT3. This enzyme has been produced as a full-length protein, but the best results were obtained using a catalytic core construct. Although crystals were obtained using both recombinant proteins, they failed to diffract adequately for structural determinations (Table 2).
The co-crystallization of SmPRMT3 with its protein partner SmRps2 is under study with the aim of improving crystal quality.
TcHAT1. This protein has a tendency to aggregate, but this was reduced by adding the enzyme cofactor acetyl-CoA. However, crystals could not be obtained, whatever the construct used. Attempts will be renewed when a substrate of the enzyme has been developed to enable co-crystallization experiments.
SmSirt2. This enzyme has also proved to be refractory to crystallization. The development of selective inhibitors should allow co-crystallization studies to go ahead in the near future.
SmHDAC8. The crystal structure of this enzyme was solved during the previous SEtTReND project [6]. During the A-ParaDDisE project a large number of inhibitors of SmHDAC8 have been generated within the consortium (see below), particularly using a strategy of bioguided optimization. The best compounds, as assessed by their low nM IC50 values and/or their effect on the parasite in culture, were used to solve the structures of numerous SmHDAC8/inhibitor complexes.
In particular, the solved crystal structure of SmHDAC8 bound to an inhibitor has revealed the mode of binding of these inhibitors to the enzyme via a schistosome-specific clamp [17] (Fig. 13). This structure also explains the strongly increased selectivity of these inhibitors for SmHDAC8 and their loss of activity against most human HDAC isoforms, apart from hHDAC8. Nevertheless, selectivity against the latter has also been achieved (up to about 8-fold) and much of the structural analysis has aimed at achieving a better understanding of the determinants for these observations.
The work on smHDAC8 selective inhibition has also been carried out by using known human HDAC8-selective inhibitors. A large amount of biochemical and biophysical data has been collected on these inhibitors as well as structural data (manuscript in preparation). Interestingly, these data further refine our view on HDAC8 selective inhibition and open new avenues towards the selective inhibition of other HDACs as well as the comprehension of their catalytic mechanisms and target recognition.
In addition, many other compounds have been received from within and outside the consortium. Structural data was also collected on many complexes of these inhibitors with SmHDAC8, providing a wealth of further structural data (manuscripts in preparation).
TcDAC2. Large quantities of enzymatically active full-length TcDAC2 were prepared, but this protein failed to crystallize reproducibly. Major protein engineering involving the partial removal of insertions, backed up by biophysical methods to determine the stability of the resulting constructs in the presence of inhibitors allowed a significant increase in protein stability. This allowed the growth of large, well-diffracting crystals. Structure determination is underway.
The process and methodologies leading to this latter result are highly original and are the subject of a patent application under preparation. No further details can be given here.
4.1.3.1.e. In silico and high-throughput screening of HMEs
i) Virtual screening of HME inhibitors.
In silico docking studies were done using generated homology models:
- Hydroxamic acid-containing SmHDAC8 inhibitors previously reported [6, 17] were docked into platyhelminth HDAC8 homology models. The docking protocol established for SmHDAC8 was used. The catalytic pockets of the HDAC8 enzymes from other schistosome species, from other trematodes and from cestodes were highly similar and all the inhibitors could be docked into the homology models with similar docking poses. The hydroxamic acid residue was always coordinated by the zinc ion and a conserved tyrosine and two histidine residues. The derived data suggest that SmHDAC8 inhibitors may also inhibit the enzymes of these phylogenetically related species and that a drug targeting SmHDAC8 would also be effective in these other species.
- The same compound libraries were docked into HDAC homology models of TcDAC2, L. braziliensis LbDAC2 and the three P. falciparum HDACs, PfDACs 1, 3 and 3 with the object of selecting compounds for phenotypic screening. Virtual screening was also carried out on a bromodomain from a recently-described essential P. falciparum protein [18]. Again, hits were subjected to phenotypic screening.
- The homology model derived for TcDAC2 was used to explain the binding modes of inhibitors tested in an enzyme assay. The results suggested starting points for chemical optimization of inhibitors. These screens also suggested that human HDAC inhibitors currently in clinical trials for cancer treatment might represent starting points for lead optimization.
- A homology model of SmSirt2 was also generated for docking studies. In order to take into account the conformational plasticity of this protein, several available conformations/structures of human Sirt2 were used in model generation. Several compound libraries including the GSK-PKIS kinase inhibitor data set, the GSK Malaria Box, KinetoBox and hits from HTS screening during the SEtTReND project, were docked and interaction/binding modes evaluated. The derived models suggest that active compounds interact with a large cavity known as the extended C-site and selectivity pocket.
ii) Enzyme assay development and enzyme inhibition-based screening
Enzyme assays have been developed for a variety of HMEs produced as recombinant proteins. These have been utilized to screen inhibitors identified from virtual compound screening and high-throughput screening and in the following iterative process where optimized inhibitors have been developed by means of synthetic chemistry.
- In vitro assays for SmHDAC8, TcDAC2 and SmSirt2 were developed and large numbers of inhibitors screened. In the cases of TcHAT1 and SmUtx assay development is ongoing.
- It has recently been demonstrated that some sirtuins are fatty acylases in addition to their deacetylase activity [19]. For SmSirt2 an assay was developed showing that the enzyme deacylates substrates with long-chain fatty acids linked to a fluorogenic substrate. This assay will be used to determine SmSirt2 substrate preference: demyristoylation over deacetylation for example.
- Screening campaigns using the FDA approved drug library and the GSK-PKIS kinase inhibitor set allowed the identification of some hits for SmSirt2 in the low µM range. Two hits from the kinase inhibitor set are now used in optimization campaigns.
Assay development adapted to High-Throughput Screening (HTS) was also undertaken:
- For SmPRMT3 screening an assay for primary screening was developed based on the Cisbio EPIgeneous Methyltransferase assay, the enzyme activity being measured by the conversion of the S-adenosylmethionine (SAM) methyl donor to SAH in a homogeneous time-resolved FRET assay (HTRF; Fig. 14). Dose-response curves were determined using both the Cisbio EPIgeneous Methyltransferase assay and another commercial assay: the Perkin-Elmer LANCE assay. Both assays had to be adapted to use in 384-well HTS plates.
- For TcDAC2 HTS an assay was developed and formatted to 384-well plates using the recombinant TcDAC2 protein. The assay measures fluorescence intensity using a tripeptidic trifluoro-lysine substrate with a coumarin fluorophore (Ac-Leu-Gly-(TFA)Lys-AMC). Further optimization resulted in an assay protocol with excellent statistics to meet the requirements for HTS of the European Lead Factory (see below) who will carry out the HTS. This assay was also used in order to evaluate HDAC8 inhibitors against TcDAC2 and to test different TcDAC2 protein constructs.
iii) High-throughput screening of HMEs
HTS was performed against SmPRMT3 by screening a diversity set of 79696 distinct compounds at 20 µM concentration. The HTS went technically well (Z’-factor 0.73) and yielded 114 primary hits, although many were very weak. Follow-up assays of the primary hits including quality control and determination of dose-responses showed that many of them interfered with the assay read-out and were therefore false-positives. Three compounds showed dose-responses in both available SmPRMT3 assays. One was pentamidine, an old drug previously used against leishmaniasis but which has severe side-effects. Small chemistry programs around the other two hits were initiated and 27 analogues were tested, but none were active. The SmPRMT3 chemistry program was therefore stopped. In addition 66 potential PRMT inhibitors generated within the consortium were tested, but none were active.
TcDAC2 is a priority therapeutic target for the A-ParaDDisE consortium: it is essential for the survival of T. cruzi, the enzyme has been produced in quantity and in an enzymatically active form, an HTS-compatible assay has been developed and diffracting crystals have been produced, indicating that the detailed catalytic site structure will be available in the near future. Since SmHDAC8 had already been screened against the compound diversity set available within the consortium, it was decided to submit a project to the European Lead Factory (ELF) in order to screen TcDAC2 against the extensive ELF library derived from a variety of academic and Pharma industry sources. This project was accepted and is currently underway. The ELF will supply a qualified hit list of a maximum of 50 of the best inhibitors for testing against the parasite, docking studies, compound optimization and lead development.
HTS campaigns against SmHDAC8 and SmSirt2 performed in the SEtTReND project provided hits which during the A-PARADDISE project have been optimized in an iterative process involving synthesis of >500 compounds, biochemical assays, phenotypic assays, in-vivo PK studies and an in-vivo efficacy study.
4.1.1.1. Phenotypic screening of focused inhibitor libraries
Phenotypic screening using focused inhibitor libraries, supplied by the “chemist” partners in the A-ParaDDisE project, formed a major part of the overall drug development approach. More than 500 compounds were tested against S. mansoni (in two screening centers), P. falciparum (two screening centers), Leishmania sp. (two centers) and T. cruzi. In general, compounds were initially screened at 10 and 20 µM against cultured parasites and compounds giving >50% growth inhibition or mortality were further tested to determine EC50 values and, in the case of S. mansoni, against different life-cycle stages. The compounds tested included inhibitors of HDACs, sirtuins, HATs, HMTs, HDMs and also DNA methyltransferase (DNMT) inhibitors and bromodomain inhibitors. Inhibitors were designated as ‘hits’ according to the ‘Katsuno criteria’ [20]. For kinetoplastid parasites, for instance, this implies an EC50 value <10 µM and a selectivity index (for toxicity on cell lines) >10. For P. falciparum these criteria are more stringent (EC50 < 1µM, SI>100). For schistosomes criteria were not defined, but we have decided to use the same criteria as for the kinetoplastids.
4.1.3.2.a. Phenotypic screening against S. mansoni
Phenotypic screening against schistosomes is well known to present problems, with different methods sometimes giving radically different results [21]. For this reason, S. mansoni phenotypic screening was carried out in two centers using different methods. Primary screening against the schistosomula larvae using a fluorescence-based method measuring mitochondrial function via the reduction of resazurin (Alamar blue) to resafurin (fluorescent) and a secondary visual screen measuring the maintenance of adult worm pairing and egg laying was carried out in one center. In another center, primary screening on schistosomula was carried out with a propidium iodide exclusion assay and secondary screens measuring lactate production by schistosomula and adult worm motility were carried out in a second center.
In the first center, 36 compounds were selected for supplementary screening on adult worms, of which 11 have so far been found to induce complete separation of worm pairs at 20 µM. Of the compounds for which EC50 values were determined, four were <10 µM and are considered the best candidates for in vivo testing.
Comparison of the results of the tests on schistosomula and adult worms in the second center showed that, of a total of 132 compounds testing positive in one or other of the tests carried out, only 33 were positive in all tests (Fig. 15). Of these, 8 compounds had selectivity indices >20 and were considered the best candidates for in vivo testing. Only two of these were in common with the compounds selected in the first center.
4.1.3.2.b. Phenotypic screening against Leishmania
Due to the variability between different Leishmania species, phenotypic assays were carried out on L. amazonensis and L. donovani in one center and on L. braziliensis in a second center. Tests were carried out on both free-living promastigotes and on intramacrophagic amastigotes.
In the first center, tests on L. amazonensis and L. donovani promastigotes were done using an Alamar blue viability assay and on L. amazonensis amastigotes using a high-content cell-based assay allowing simultaneous evaluation of effects on parasite load and on macrophage viability. Tests on promastigotes yielded 47 compounds active at 10 µM on promastigotes of both species. In contrast, only 22 compounds were active on intracellular amastigotes. Of these, 8 compounds were active both on promastigotes at 10 µM and on amastigotes at 1 µM. Five compounds were putative DNMT inhibitors and one other was an HDM inhibitor. Analogues of active compounds were synthesized and supplied for testing, yielding further active compounds. The best compounds showed activity in the low µM range.
Tests on L. braziliensis promastigotes and intramacrophagic amastigotes yielded fewer compounds that were active against the parasite and simultaneously showed low toxicity to the macrophages. Only three compounds, both HDAC inhibitors, were selected as candidates for in vivo studies. Two were also selected for studies on their effects on the parasite transcriptome.
4.1.3.2.c. Phenotypic screening against T. cruzi
Compounds were tested using the Tulahuen strain of T. cruzi engineered to express E. coli β–galactosidase. Infective trypomastigotes were used to infect primary cultures on L929 fibroblasts in microtiter plates. Parasite viability after 4 days of drug treatment was measured by quantifying the β–galactosidase activity in the culture lysate. L929 cell viability was measured separately, allowing the calculation of selectivity indices.
Of the compounds tested 170 showed some activity against the parasite, the inhibitor specificity of which is shown in Fig. 16. Inhibitors of all classes showed activity, but interestingly inhibitors of HDMs, including LSD1 were overrepresented among the positive compounds (26%) compared to the numbers tested (18% of all compounds).
Of the compounds with activity against the parasite, only 10 meet the Katsuno criteria, three compounds have EC50 values of <5 µM, which is comparable to that of the currently used drug benznidazole (3.8 µM), and SIs >10.
4.1.3.2.d. Phenotypic screening against P. falciparum
Compounds were tested for anti-malarial activity against two strains of P. falciparum intraerythrocytic stages maintained in culture, the chloroquine-sensitive 3D7 strain and the resistant Dd2 strain. Further testing on some ‘hits’ was also done using the resistant W2 strain. Two testing methods were used. For the 3D7 and W2 strains parasitemia was detected by flow cytometry after Sybr Green staining and for the Dd2 strain, parasitemia was measured after 3H-hypoxanthine incorporation.
A large number of compounds proved positive according to the Katsuno criteria initially used and a more stringent cut-off with an EC50 value of <100 nM was subsequently used. Nevertheless, 47 compounds were positive at this level against the 3D7 strain and 28 against the Dd2 strain. The best compounds had EC50 values of < 10 nM. Moreover, results obtained using different strains were broadly coherent. For example, the most active compound had an EC50 of 4.2 nM in strain 3D7 and 6.6 nM in strain W2. From all these results, 28 compounds were classified as having the most potent activity and the 8 best compounds were placed in the pipeline for in vivo testing.
4.1.1.2. Bioguided optimization of HME inhibitors
The optimization of HME inhibitors for their capacity both to inhibit their target enzymes and for their toxic effects on the parasites has been initiated for several compound series:
- The structure based optimization of TcDAC2 hydroxamate inhibitors using the homology model. Docking and ranking led to a subset of ~30 compounds that were tested against the enzyme in vitro.
- A homology model generated for L. braziliensis DAC2 was docked with hydroxamates of the TH series [17] and structural optimization led to the selection of one compound which proved active in phenotypic screening against the parasite.
- A homology model for P. falciparum DACs 2 and 3 was again docked with TH series hydroxamates, leading to the testing of one compound that gave EC50 values of 50 and 70 nM against the 3D7 and Dd2 strains respectively.
- More generally, the identification of hits by phenotypic screening has led to the synthesis of analogues for further testing for a variety of compound series. These compounds are among those selected for in vivo testing.
The best example of bioguided optimization of a compound series is provided by the TH series of hydroxamate-based SmHDAC8 inhibitors [17]. Compounds identified as isoform-selective inhibitors by in silico screening of a large virtual compound library and validated by enzyme inhibition assays, co-crystallization with the enzyme and tests on schistosomes in culture [6], were used as starting points for the synthesis of compound series with the aim of generating more selective inhibitors, but which were also more potent against the parasite. The starting point for this series was the inhibitor JM1038 (Fig. 17)
The TH series of benzamido-hydroxamates [17] showed increased affinity for the SmHDAC8 catalytic pocket, due to an SmHDAC8-specific clamp, involving specific interactions with catalytic pocket histidine and lysine residues and the flipped-out configuration of phenylalanine 151, which is sterically impeded in the human orthologue (Fig. 18).
Both increasing selectivity and increasing in vitro potency against the parasite has been achieved in this series and in a parallel series (TB) using a different inhibitor as the starting point. The EC50 curves of the two most potent compounds are shown in Fig. 19.
4.1.1.3. In vivo studies on HME inhibitors
During the project 27 compounds were selected, synthesized and submitted for formulation and PK studies. The in vitro evaluation prior to PK studies consisted in the determination of (i) identity, purity, chemical stability and solubility of the compounds, (ii) an LC-MS/MS method to be used for bioanalysis of the compound in the plasma samples and (iii) a formulation suitable for the dosing solution. Of these compounds 23 were found to be suitable for testing and possible to formulate for dosing. These compounds were given to mice both i.v. and orally and sequential plasma and urine samples measured in order to determine plasma compound concentrations, estimate half-life and bioavailability. Using these results compounds have been selected for in vivo experimentation in mouse models of infection for S. mansoni, P. berghei and T. cruzi.
4.1.3.4.a. In vivo testing in S. mansoni-infected mice
Experiments have been carried out with two TH-series SmHDAC8 inhibitors. A single dose of compound (50 mg/kg) administered orally either at day 5 of the infection, targeting the infective larvae (schistosomula) migrating in the blood stream, or at day 35, targeting paired adult worms at the start of egg laying in the mesenteric veins, was ineffective in reducing either the worm burden or eggs recovered from the liver and intestines. Four successive daily oral doses (50 mg/kg) of the compound TH92, either on days 5-8 after infection or on days 30-33, did lead to a partial reduction in both worm and egg burdens in treated animals compared to controls dosed with the vehicle alone. However, these results have proved to be inconsistent. Further experiments will be done with the most potent SmHDAC8 inhibitors, TH143 and TB87 as well as another compound, MC3974, which is also very potent against the parasite in vitro.
4.1.3.4.b. In vivo testing in P. berghei-infected mice
Eight compounds were initially selected for in vivo experiments, but one proved unsuitable. For five compounds P. berghei-infected Balb/c mice (106 parasites) received 8 (twice daily) doses of 50 mg/kg orally. For three compounds, no protection against infection was observed. One compound proved toxic although some protection against infection was also observed. Mice given compound MC3681 all survived and no parasites were present in blood smears 4 days following infection.
Two further compounds, in view of their PK properties, were given as one oral and one i.v. dose daily for four days. However, no protection was observed using these compounds despite their in vitro potency.
The compound MC3681 is a highly promising compound and will be tested further. IP protection is also envisaged for this compound.
4.1.3.4.c. In vivo testing in T. cruzi-infected mice
The most potent compound in the phenotypic compound screen at the time, IN0441411, was tested in 10 mice infected with the T. cruzi DM28c strain modified to express luciferase that were infected with 105 trypomastigotes i.p. Three days after infection, the level of parasitemia was determined by the fluorescence generated after injection of luciferin. The mice were divided into a test and a control group. Compound IN0441411 was injected i.p. at 25 mg/kg every 12 hours from days 4-8 post-infection. Luminescence was tested on days 7 and 9 post-infection. The compound had a growth inhibitory effect on the parasite. At day 7 levels were similar to those at day 3 in the test group, whereas levels had doubled in the control group. At day 9, corresponding to the peak of infection, three of five mice in the test group had low levels of luminescence intensity (Fig. 20). This experiment will be repeated with additional controls, but IN0441411 is a lead candidate for development.
Conclusions
The A-ParaDDisE project has generated a wealth of data showing that many epigenetic enzymes of parasites are bona fide therapeutic targets. Crucially, species selectivity of inhibitors of these enzymes can be achieved through bioguided optimization and these inhibitors can have potent effects on the parasite both in vitro and in vivo. Three compounds have been so far shown to be active in vivo in mouse models of infection and can be considered to be lead candidates for further development. Moreover, more compounds are in the pipeline, including inhibitors that are more potent than those yet tested.
Work on several aspects of the project will continue with the aims of producing high-impact publications, patent applications and stimulating future funding. This work has reached a pivotal phase where continued investment is necessary to bring the results already obtained to fruition.
References
1. Hotez P. et al. (2016) PLoS Negl. Trop. Dis. 10(3):e0003895
2. Mwinzi PN et al. (2015) Front. Public Health, 3:48.
3. Roquis D et al. (2015) PLoS Negl Trop Dis, 9(8):e0003853.
4. Pierce RJ et al. (2012) Curr Pharm Des, 18:3567-78.
5. Andrews KT et al. (2012) Curr Pharm Des, 18:3467-79.
6. Marek M et al. (2013) PLoS Pathog, 9(9):e1003645.
7. Scholte L et al. (2017) Submitted.
8. Melesina J et al. (2015) J Mol Graph Model, 62:342-61.
9. Peng D et al. (2014) MBio, 6(1):e02097.
10. Picchi GF et al. (2016) J Proteome Res, 16:1167-79.
11. Weltregny D et al. (2004) Am J Pathol, 165:553-64.
12. Deardorff MA et al. (2012) Nature, 489:313-7.
13. Anderson L et al. (2015) PLoS Negl Trop Dis, 9(12):e0004334.
14. Anderson L et al. (2017) PLoS Negl Trop Dis, in press.
15. Chao MW et al. (2016) Oncotarget, 7:1796-807.
16. Fiebig M et al. (2015) PLoS Pathog, 11(10):e1005186.
17. Heimburg T et al. (2016) J Med Chem, 59:2423-35.
18. Josling GA et al. (2015) Cell Host Microbe, 17, 741-51.
19. Zhang X et al. (2016) Nat Chem Biol, 12:614-20.
20. Katsuno K. et al. (2015) Nat. Rev. Drug Discov. 14:751-758.
21. Panic G et al. (2015) Parasit Vectors, 8:624.
Potential Impact:
The aim of the A-ParaDDisE project was to develop new drug candidates for use in treating four neglected parasitic diseases, schistosomiasis, leishmaniasis, Chagas disease and malaria, by targeting enzymes (histone modifying enzymes) and other proteins involved in epigenetic processes that regulate gene transcription in these organisms. The project is built on a strategy developed in a previous, successful EC-funded project, SEtTReND, which focused on schistosomiasis, expanding the number of teams participating to create an innovative platform for drug discovery. In all 17 teams participated in the project and included those involved in the chemical synthesis of epigenetic inhibitors combined with others that carried out phenotypic screening on the parasites, molecular modeling and virtual screening, recombinant protein production and structural analysis, high-throughput screening, target validation and characterization, pharmacokinetics and in vivo testing. In this way inhibitors of histone modifying enzymes (HME) from focused inhibitor libraries or identified by target-based screening methods have been characterized as novel compounds with potent activity against the parasite in vitro and in vivo. The working hypothesis that, despite the evolutionary conservation of epigenetic processes, inhibitors could be selected or developed that show species selectivity has been validated. A number of compounds that have activity in vivo (mouse infection models) have already been identified and can be considered as lead compounds, justifying further development through compound optimization and pre-clinical testing. Further compounds that have potent effects on the parasites in culture are in the development pipeline. One challenge facing the teams involved in the project is to attract collaborations and funding that will allow the continued development of these drug candidates through pre-clinical and clinical testing.
Key assets of our approach were set out in the impact statement of the initial application and have been confirmed during the project. These are:
- Proven concept. The extension of the approach developed in the SEtTReND project, which focused on schistosomiasis, to diseases caused by evolutionarily distant protozoan parasites with very different etiologies, has been fully justified. The combination of both target and structure based HME inhibitor identification with the phenotypic screening of focused inhibitor libraries made up of compounds initially developed against other pathologies such as cancer, has allowed the identification of compounds active against all the parasites studied.
- Different parasites, the same enzyme families. Genome data-mining showed that all these parasites indeed have HMEs in all the families studied in the project: histone acetyltransferases, histone deacetylases, histone methyltransferases and histone demethylases as well as bromodomain proteins and DNA methyltransferases. Moreover, inhibitors of the main HME types are toxic to the parasites in culture and gene knockout studies carried out in T. cruzi have shown that some individual HMEs are essential for parasite survival, as was shown for some S. mansoni HMEs by transcript knockdown.
- Inhibitors active against related species. Homology modeling of HMEs, particularly HDACs, from the species under study show that some do show cross-species similarity. This is the case, for example, for HDAC8 from a variety of parasitic flatworms including both trematodes and cestodes. However, there are significant differences between the metazoan flatworms and the protozoa, and between different protozoan parasites. It is not probable that an inhibitor developed against an HDAC from T. cruzi would be effective against Leishmania parasites for example.
4.1.4.1. Strategic Impact
4.1.4.1.a. Innovation and Scientific Impact
The scientific impact of the project, both within the domain of parasitology and more widely in the areas of drug development and epigenetics, has been considerable. The combination of the structural studies with compound screening and the exploration of the roles of HMEs has generated considerable interest in the scientific community, through high-impact publications, presentations in international conferences and other dissemination activities. Again, the principal areas of projected impact set out in the project proposal are revisited below to define the impact generated during the project.
- Identification and characterization of parasite HMEs. The identification and characterization of the enzymes involved in histone acetylation, deacetylation, methylation and demethylation, as well as bromodomain proteins, has generated a wealth of data that has been used in the project to identify potential drug targets via phylogenetic studies and molecular modeling. A stringent phylogenetic study of the class I and II zinc-dependent HDACs and the class III HDACs, the sirtuins, has led to the clarification of the sub-class relationships of the parasite enzymes to their human counterparts, the measure of evolutionary distances and the correction of some misconceptions in published research from other groups. This work has also improved annotation and has been added to sequence databases.
- Transcriptomic analyses. The analysis of the effects on parasite transcriptomes of treatment with HME inhibitors, and in the case of T. cruzi, after the invalidation of a gene encoding an HDAC, has already provided crucial insights into metabolic and signaling pathways influenced by HMEs in the parasites. One practical consequence has been the identification of the transcriptional regulation of the expression of a component of a histone methyltransferase complex by histone deacetylases following the analysis of the transcriptome of schistosomes treated with an HDAC inhibitor. This suggested the combination of the HDAC inhibitor with the HMT inhibitor, which led to synergistic effects on the parasite: lower doses of each inhibitor, in combination, lead to parasite killing. Moreover, detailed analysis of the genes up- and downregulated by inhibitor treatment has given indications of the molecular mechanism of action on the parasite and will be utilized to assess eventual off-target effects.
- Comparative structural studies. The study of the X-ray structure of the catalytic domain of SmHDAC8 has yielded important insights into inhibitor binding and the mechanisms leading to isoform and species inhibitor selectivity. This has allowed the development of increasingly selective inhibitors through bioguided optimization and has aroused a great deal of interest from groups outside the A-ParaDDisE consortium, leading to collaborative studies. The work carried out is exemplary and will doubtless serve as a paradigm for the development of selective inhibitors against other targets. The successful crystallization of an HDAC from T. cruzi was possible due to a novel combination of bioengineering methodologies. The possibility of IP protection of this procedure is under consideration. In parallel a comparative study of parasite HME catalytic domains, particularly for the HDACs, has been done using generated homology models. These models have allowed the identification of variable residues within the catalytic pocket and identify divergent enzymes as potential therapeutic targets.
- Novel HME inhibitor structures. During the project more than 500 compounds were tested against the parasites in phenotypic assays, many of which were generated specifically for the project after enzyme activity and structure-based inhibitor optimization studies. Some of these series, such as the hydroxamate-based inhibitor series generated as selective inhibitors of SmHDAC8, have been published and have been taken up by other research groups. For others, the possibility of IP protection is under consideration, according to the conditions laid down in the Consortium Agreement concerning the exploitation of foreground. A linear strategy has allowed the selection of a significant number of compounds to go forward for in vivo studies, based on their activity on parasites in culture and according to established criteria. Three compounds that are well tolerated and are active in vivo are considered as leads for further development and more than 15 compounds are in the pipeline.
- Formulation, pharmacokinetics and in vivo studies. The procedure established for establishing conditions for the formulation of inhibitors for administration in vivo, and the study of the pharmacokinetic properties has allowed the rapid and efficient compound selection.
The scientific impact of the project has been ensured by the effective dissemination of the results via publications, communications and presentations during international conferences (see section 4.2.2 below).
4.1.4.1.b. Societal Impact
The principal societal impact of the work to develop new drugs against neglected parasitic diseases are the populations for which these diseases are leading causes of chronic disability and poverty, as well as being a leading cause of mortality, particularly in childhood. The diseases affect the countries the least able to combat them. The aim of delivering improved treatments will have a dual impact on both public health and living standards in countries in the tropics. Moreover, as well as being a legitimate humane objective for the European community, improved health and economic prospects would, in the long term, lessen the need for aid and ameliorate economic exchanges.
New drugs are required to replace existing treatments where these lead to significant side-effects or where drug resistance is appearing, which is the case for all the diseases studied in the project. In the absence of effective vaccines, and there is currently little prospect of their development, drug treatment remains a crucial arm in control strategies. Moreover, as well as a dearth of effective, well tolerated treatments, the development of drug resistance means that it will be necessary to continuously develop new drugs. For this reason it is important to explore new drug targets and the epigenetic enzymes are particularly promising in this respect.
The health issues posed by the neglected parasitic diseases are of common concern to the endemic countries and the EC. In addition, it is important to consider that these diseases are spreading northwards, possibly due to climate change and increased immigration. Schistosomiasis transmissible to humans has appeared in Corsica and visceral leishmaniasis is endemic in dogs in southern Europe. While it would be wrong to overestimate the impact of these diseases in Europe, this should serve as a reminder that they should not be considered only as problems of developing countries.
In order for the work of the A-ParaDDisE project to have the hoped for societal impact, it will be necessary for the hits and leads discovered to follow the drug development pipeline through lead development, pre-clinical testing and clinical trials. For this financial support will be necessary. Ways in which this can be achieved through a European dimension are discussed in section 4.1.4.
4.1.4.2. Training and Technology Transfer
Training activities in the project have concerned the exchange of students and post-doctoral fellows between beneficiary laboratories and the presentation of their work at international conferences, as well as the organization of a project Workshop. A non-exhaustive list of specific training activities and technology transfers follows:
- Presentations at conferences by project students and post-doctoral fellows were numerous and are listed in the dissemination table. In particular, the FRIAS Epigenetics symposium in Freiburg (April 2016), organized by beneficiary 6, was attended by post-doctoral fellows and students from European and Brazilian beneficiaries.
- Direct training activities during the project concerned PhD students and post-doctoral fellows from beneficiaries 1, 4, 5, 7 and 13 who underwent training in beneficiary laboratories (1, 4, 5, and 6) for enzyme inhibition assays, RNASeq experiments and related technologies.
- Exchanges and discussions between beneficiaries 1, 4, and 5 and 7a concerning HME inhibitor treatment of schistosomes and RNASeq experiments. Similar exchanges were undertaken between beneficiaries 5 and 7b concerning T. cruzi RNASeq experiments.
- Exchanges and discussions concerning phenotypic screening methods applied to schistosomes between beneficiaries 1 and 7a.
- Development and standardization of enzyme assay methods between beneficiaries 2, 6 and 9 concerning TcDAC2.
One important societal impact is also the training of the next generation of scientists in the field of anti-parasitic drug discovery in epigenetics that will be active in the field of research and teaching in the future. Young scientists at different career levels were trained to be specialists in their requisite fields (molecular biology, parasitology, screening etc.) but learned early on to work in a highly interdisciplinary approach and to keep all these different aspects in perspective. One example is a post-doctoral fellow who worked as core support personnel in the project at the beneficiary 6 laboratory. With support from the consortium he received a scholarship from the German Science Foundation to work at the University of Oxford also in the field of drug discovery regarding parasite epigenetics. He will become a junior group leader on his return to Germany and hence will further develop and spread our approach.
A workshop organized by the Consortium and in particular beneficiary 13, was held in Rome in December 2016 on “Epigenetic drug discovery for Neglected Parasitic Diseases”. Held over two days it was attended by 100 participants and presentations were given by 6 invited speakers and 17 speakers who were members of the A-ParaDDisE consortium. All aspects of epigenetic drug discovery were covered from chemical synthesis of inhibitors to structural studies, compound screening, bioguided optimization and the target product profile.
4.1.4.3. European Dimension
A project of the size and ambition of A-ParaDDisE necessitated the participation of numerous partners. These included 10 from Europe, 5 from Brazil and 2 from Australia. The European beneficiaries represented 5 EC member states (France, Germany, Sweden, Italy and the UK). The project required the collaboration of beneficiaries working on all aspects of drug discovery, parasite biology and molecular biology. The structure of the Consortium also ensured access to the parasites endemic to Brazil and maintained in the beneficiary laboratories. The financial support of the EC was essential to the construction of the project: no other comparable funding mechanisms exist for collaborative projects of this scale between European and endemic countries. Moreover, although there is intense competition in the search for drugs against parasitic diseases, apart from other European consortia, no comparable projects exist, for example between the USA and partner countries. This means that consortia like A-ParaDDisE maintain Europe at the forefront of research in the domain.
4.1.4.4. Synergies with other EC-supported consortia and future funding
As well as the A-ParaDDisE consortium, three other consortia, Kindred, NMTrypI and PDE4NPD, were also funded under the same FP7 call. There is some overlap concerning the parasites studied: NMTrypI and Kindred both study kinetoplastid parasites and the PDE4NPD consortium studies helminth infections including schistosomiasis. However, although there are similarities in the approaches used for drug discovery, the chosen targets are different. With the aim of avoiding redundancy of targets and/or compounds studied the four projects share a specific “Synergy” work package and have communicated regularly during the projects. Moreover, a common symposium was organized by the NMTrypI consortium in Modena (June 2016) to allow comparisons of approaches and progress. At this meeting it was evident that all the projects had produced large numbers of ‘hits’ and “leads’ warranting further development and piloting through preclinical and clinical testing. Discussions have since focused on developing a common approach to obtaining future support for the work to be carried out. Two of the actions taken have been to write an article in press in Trends in Parasitology and a letter to the European Commission, both setting out a strategy by which the EC can support future work and remain at the forefront of drug discovery for neglected parasitic diseases. The assessment of the situation and a strategy to remedy it can be summarized as follows:
- The FP-programme has now been completed and replaced by H2020. Under this programme there is provision for vaccine development and support for clinical trials (but only from Phase 2 onwards) in Africa for NTDs. Although this support is significant and important, there is currently no provision for ensuring progress in the work on drug discovery.
- The next logical step of the investment strategy of the EU would be to create a call within H2020, to enable continued interaction between academia and SMEs to create a platform which is capable of continuing the work started by the consortia funded in this final FP7 call.
- This platform would specifically take on the task of bridging the gap between lead optimization and early-stage clinical trials (Phases 1 -2). Lead compounds developed by the four consortia, but also from other sources (e.g. teams participating in the current COST 1307 initiative) would feed the platform regardless of the parasitic disease targeted i.e. moving beyond the current FP7 call and overcoming the problem of fragmentation amongst groups in the field
- As identified in the current FP7 call, not all groups involved in the drug discovery process have the same goals therefore a larger view of this platform would be to divide it into at least two complementary groups with different specificities. One focused on early stage screening, FBDD and innovative ways to design more effective drug development and a second, intent on developing lead compounds towards the clinic from in vivo animal studies, ADMET and GLP/GMP and registration to phase 1. We think that this type of platform structure is highly desired, as the number of candidates populating the different stages of the drug development pipeline needs to be dense to statistically ensure enough drug candidates enter phase 1.
- We see that this as an opportunity to build on the existing consortia, a rich source of knowledge and expertise to create one or more new consortia, in order to drive the various different innovative approaches with one consolidated aim the intent to produce drug candidates appropriate for use in the clinic.
- In a larger vision, once the infrastructure was in place and proven to be productive, this could be a dynamic European/International platform bringing together different partners from academic /industrial/foundations, attracting co-funding from large NGOs, Foundations and Industry. Even partial funding from the EU will facilitate attracting support from these other sources. The Innovative Medicines Initiative (IMI) could become involved in financing partnerships between the pharmaceutical industry and academic groups for drug development for neglected parasitic diseases. Several large pharmaceutical companies (GSK, Merck Serono, Novartis..) are already engaged and contribute to the field.
4.1.4.5. European competitiveness
The A-ParaDDisE consortium included two industrial beneficiaries, the SMEs Kancera and Adlego, as well as a company engaged in intellectual property management and dissemination, Inserm Transfert. Objectives of these beneficiaries included:
- To increase awareness of their expertise within and outside Europe to provide new contacts leading to business opportunities.
- To influence the direction of the project and help focus on the primary objective: drug development.
- In the case of Kancera, to exploit novel drug structures derived from their background for other research projects, notably in cancer.
- To participate in the industrial exploitation of ‘hits’ and ‘leads’ developed during the project. The involvement of large pharmaceutical companies in this aspect, as outlined in the preceding section represents a major objective for progress in neglected parasitic disease drug development and for improving European competitiveness in the domain.
A further contribution to the integration of European research initiatives and competitiveness made by A-ParaDDisE concerns the project submitted to and accepted, after a highly competitive evaluation, by the European Lead Factory (ELF), an organization within the IMI. This project concerns the high-throughput screening of the T. cruzi HDAC TcDAC2, which is an enzyme essential to the growth and survival of the parasite. The recombinant enzyme will be screened against the large compound library of the ELF, which will supply a qualified hit list of a maximum of the 50 best inhibitors. These will be tested by consortium members implicated in the project (beneficiaries 2, 6, 7b and 9) and compounds active on the enzyme and the parasite will be developed. This latter stage of the work will be subject to discussion with the provider(s) of the compound(s) selected and may lead to collaborative projects with pharmaceutical industry partners.
List of Websites:
http://a-paraddise.cebio.org/
The aim of the A-ParaDDisE project was to feed the drug development pipeline for Neglected Tropical Diseases (NTDs) and in particular leishmaniasis, Chagas disease, malaria and schistosomiasis. These diseases collectively infect more than one billion people, mainly in developing countries, cause hundreds of thousands of deaths annually and represent a major economic burden. Treatments for all these diseases face problems due to developing resistance, severe side-effects and/or prolonged dosing schedules. The development of new drugs is a priority and we extended and expanded a strategy applied to schistosomiasis drug development in a previous EC-funded project, SEtTReND. This involved targeting enzymes and other proteins involved in the epigenetic processes that control gene transcription and in particular the modification of histones via acetylation/deacetylation and methylation/demethylation: histone modifying enzymes (HMEs). These enzymes are acetyl- (HAT) or methyltransferases (HMT), deacetylases (HDAC) or demethylases (HDM) and bromodomain-containing proteins that are readers of acetylated histones and recruit multi-protein complexes involved in transcriptional regulation. These enzymes are increasingly targeted in the therapy of a variety of pathologies including cancer, which allows us to benefit from the expertise and inhibitors already developed as starting points for developing anti-parasitic drugs.
We therefore aimed to identify HME inhibitors that will be lead compounds for drug development. We employed two complementary strategies: phenotypic screening using focused HME inhibitor libraries and a structure-based approach involving validated target enzymes and high-throughput or in silico screening. Bioguided optimization of the hits identified fed into lead development, basic ADMET and in vivo testing to generate lead compounds. The main achievements of the A-ParaDDisE project are as follows:
All the HMEs encoded in the genomes of the parasites studied have been identified. Phylogenetic analyses and molecular modeling allowed the choice of HMEs showing significant differences, particularly within the catalytic pocket, from their human counterparts, likely to facilitate the development of selective inhibitors. A total of 37 HMEs from Schistosoma mansoni, Leishmania braziliensis, Plasmodium falciparum and Trypanosoma cruzi were cloned.
- Transcriptomic studies (RNASeq) allowed the identification of gene expression signatures in parasites either treated with HME inhibitors or invalidated for a target HME gene/transcript.
- Gene knockout studies for 8 T. cruzi HMEs has shown that three (two HDACs and one HMT) are essential for parasite growth and survival. Knockout of one other HDAC does not affect viability but leads to a metacyclogenesis defect.
- Recombinant proteins have been obtained for a large number of HMEs. Notably two T. cruzi HMEs, TcHDAC2 (essential for parasite survival) and TcHAT1 have been produced as active enzymes following extensive protein engineering. Diffracting crystals of the former enzyme have been produced and structural analysis is ongoing. The S. mansoni arginine methyltransferase SmPRMT3 has also been produced as an active enzyme and crystallized.
- High throughput screening (HTS) of SmPRMT3 has been carried out. HTS using TcDAC2 has been submitted to the European Lead Factory and screening will start shortly.
- Bioguided inhibitor optimization was carried out for the essential S. mansoni HDAC8. A combination of enzyme inhibition assays, co-crystallization with the enzyme and phenotypic screening yielded compounds with increased selectivity and potency.
- Phenotypic screening was carried out on all the parasites studied in the project using focused libraries of HME inhibitors supplied by the chemist groups in the project. More than 500 compounds were tested and a large number (> 100) of hits identified.
- Pharmacokinetic and toxicity studies were carried out on the most promising hits and in vivo testing was carried out in mouse infection models for Plasmodium, T. cruzi and S. mansoni infections. Three compounds with significant in vivo activity have so far been identified and others are continuing to be tested.
The A-ParaDDisE project has fulfilled its principal objectives and the results obtained validate the strategy of targeting parasite epigenetic process for drug development.
Project Context and Objectives:
The disease burden imposed by parasitic infections worldwide is huge, and disproportionately affects poor countries in tropical and subtropical regions. Their impact is not only due to the mortality that they provoke, but also, more insidiously, to disease morbidity, which affects overall well-being of the populations concerned and is an underlying cause of persisting poverty [1]. The aim of our project was to develop novel drug candidates against four major parasitic diseases, schistosomiasis, leishmaniasis, Chagas disease and malaria, by targeting the enzymes responsible for epigenetic processes in the parasites. These diseases are among those defined as “Neglected Tropical Diseases” (NTDs) in the London declaration of 2012 “Uniting to Combat Neglected Tropical Diseases” (http://unitingtocombatntds.org/resource/london-declaration) which put the control or elimination of NTDs in the vanguard of global efforts to “chart a course toward health and sustainability among the world’s poorest communities to a stronger, healthier future”. A key element of the objectives set out concerns the development of new drug treatments in order to combat developing resistance against some treatments and to replace other treatments that are unsatisfactory due to significant side-effects, prolonged treatment regimens or are not effective at all disease stages. The drug development process is challenging and the attrition rate for new compounds entering the pipeline is high, adding to the cost of new drugs that successfully negotiate the difficult and expensive road from the initial identification of a “hit” during the screening process via lead optimization, pre-clinical testing and clinical trials. In order to produce more drugs, a large number of “hits” are required to feed the pipeline and this has been the fundamental aim of the A-ParaDDisE project.
The need for new drugs to treat NTDs
Our project focuses on parasitic diseases that all present problems for their control in endemic countries. There are no effective vaccines currently available and although some alternative control strategies exist, for example the use of insecticide-impregnated bednets in the case of mosquito-borne infections, drug treatments remain in the front line of disease control. There is only one drug available for the treatment of schistosomiasis, praziquantel, which is used for mass chemotherapy in sub-Saharan Africa, and besides increasing evidence for treatment failure, concerns have been raised about the real efficacy of this drug [2]. The treatment of choice for systemic leishmaniasis, liposomal Amphotericin B, has severe side-effects and is costly. Miltefosine has been used as an oral leishmaniasis treatment, but in India resistance to the drug is now a significant problem. The treatment of Chagas disease using benznidazole or nifurtimox requires long treatments, causes side-effects and it has recently been shown that benznidazole has no effect on the chronic, most clinically important, form of the disease. Finally, the increasing level of resistance to Artemesinin and Artemesinin-based combination therapies in south-east Asia, as well as reports of treatment failure in patients infected in Africa, has rendered urgent the introduction of alternative therapies.
Targeting epigenetic processes in parasites
The term “epigenetics” encompasses several types of heritable changes in gene expression that are linked to structural changes in chromatin, but without any modification of the DNA sequence. They include methylation of the DNA itself, variant histones, chromatin remodeling factors, non-coding RNAs and reversible, post-translational modifications of histone “tails”. The latter are numerous and include acetylation, methylation, phosphorylation, SUMOylation, ubiquitinylation, which, together with other “histone marks” constitute a complex “histone code”. This has a central role in regulating chromatin architecture, the binding of protein cofactors, the regulation of gene expression and its dysregulation in pathology. The enzymes that effect these modifications, “writers” and “erasers” of histone marks, and, increasingly, “reader” proteins that interact with them, are privileged targets for chemotherapeutic interventions. In particular, the writers, readers and erasers associated with histone acetylation and methylation have been under active investigation as potential drug targets for some time, particularly in cancer. Five FDA-approved inhibitors of histone deacetylases (HDAC) are now used in cancer treatment and a large number of other inhibitors of histone modifying enzymes (HME) or of bromodomain proteins, which interact with acetylated lysine residues, are in clinical trials.
Parasites are eukaryotic organisms and, as such, share common histone marks and epigenetic mechanisms with humans [3]. Moreover, HMEs have been shown to have vital roles in the growth and survival of parasites [4] and are therefore valid therapeutic targets. In the malaria parasite, Plasmodium falciparum, HDACs, histone acetyltransferases (HAT) and methyltransferases (HMT) have roles in the control of gene expression and cell cycle progression, and particularly in the control of variable surface gene expression involved in immune evasion by the parasite. In Trypanosoma brucei, a protozoan parasite of the Kinetoplastid family, which includes the agents of leishmaniasis (Leishmania sp.) and Chagas disease (Trypanosoma cruzi), both HDACs and HATs have been shown to be essential for growth and survival using gene knockout. Moreover, during a previous EC-funded project (SEtTReND) transcript knockdown by RNA interference was used to identify three HMEs, HDAC8, the lysine demethylase KDM1 and the arginine methyltransferase PRMT3 as valid targets in Schistosoma mansoni, the flatworm parasite causing intestinal schistosomiasis.
Inhibitors of HMEs, particularly HDAC inhibitors, have been shown to be effective against various parasites maintained in culture. For example, Trichostatin A (TSA) and the approved cancer drug Vorinostat are effective against P. falciparum and S. mansoni, but less so against Leishmania sp. [reviewed in 4]. However, these are pan-HDAC inhibitors that are active against all the isoforms of HDACs and inhibit human HDACs very effectively. The key to using HME inhibitors as a basis for new drugs against parasites is to develop compounds that are selective for parasites: i.e. that selectively inhibit a particular HME isotype and which also show species selectivity. We consider these criteria essential in order to develop candidate drugs that show minimal side-effects when used in humans. Moreover, the feasibility of this approach has been demonstrated by the development of HDAC inhibitors selective for P. falciparum that were effective in the low nM range against the parasite in culture and with selectivity indices >600 when tested on a macrophage cell line [5]. Similarly, selective inhibitors of class III histone deacetylases, the sirtuins, have been developed that are selectively active against L. infantum.
During the SEtTReND project, which focused on S. mansoni, we developed a target- and structure-based strategy to develop selective inhibitors as drug precursors. The most striking example of this strategy is the development of selective inhibitors of S. mansoni HDAC8 (SmHDAC8). This involved the crystallization of the recombinant protein produced in Escherichia coli, the solution of its 3D structure by X-ray crystallography, in silico screening of the catalytic pocket structure by docking compounds in a virtual library and the testing of selected compounds for their capacity to inhibit the recombinant SmHDAC8 in an enzymatic assay and for their ability to kill parasite larvae in a phenotypic assay [6]. This strategy was complemented by high-throughput screening of a compound library using recombinant SmHDAC8 and this latter strategy was also successfully used to identify inhibitors of S. mansoni sirtuin 2 (SmSirt2). Target and structure-based screening was also carried out utilizing in silico homology modeling of P. falciparum HDAC1 [5] to develop selective inhibitors that are promising drug precursors.
Objectives and strategy of the A-ParaDDisE project
The A-ParaDDisE project drug discovery strategy expanded the successful approach developed for the flatworm S. mansoni during the SEtTReND project to three protozoan parasites, P. falciparum, Leishmania sp. and T. cruzi. It combined structure-based approaches on selected, validated HME targets, with phenotypic screening using focused libraries of HME inhibitors. The “hits” identified by these screening steps then underwent a bioguided optimization process in order to improve their selectivity and activity against the parasite. The best compounds identified during this process were then subjected to toxicological studies and basic pharmacokinetic analysis before testing in vivo in mouse infection models. This strategy is summarized in Fig. 1. (see report attached).
This central strategy was supported by the specific principal objectives:
- The identification of HMEs from all four parasite species by data mining of the respective genome sequences. This was followed up by molecular and functional characterization of selected targets, target validation using gene knockout and transcript knockdown.
- Transcriptomic studies were carried out on parasites treated with HME inhibitors or knocked out for selected HMEs in order to determine the pathways affected by the treatment and to validate inhibitor selectivity.
- Phenotypic screening of all four parasites using focused HME inhibitor libraries.
- Production of recombinant parasite HME proteins for selected therapeutic targets in order to allow high-throughput screening (following enzyme assay development) crystallization and structure determination using X-ray crystallography.
- Optimization of inhibitor structures by chemical synthesis based on molecular modeling and structural studies. Bioguided feedback was used to direct optimization of inhibitor potency, selectivity and biological activity.
- Pharmacological and toxicological studies (in vitro and in vivo) were carried out on the best hit compounds derived from the screening/optimization process. Inhibitors showing high selectivity indices, satisfactory PK characteristics and absence of toxicity in mice were then tested in vivo in mouse infection models.
The A-ParaDDisE project has used an innovative and comprehensive strategy for drug discovery pooling resources and expertise garnered against different parasites by experts in their domains. (Project members listed in report attached).
The project partners brought different areas of expertise to the project including bioinformatics and target identification (2), biochemical characterization and transcriptomics (1, 4, 5, 7, 11, 15, 16), parasite phenotypic screening (1, 7, 11, 15 and 16), chemical synthesis (3, 6, 9, 12, 13 and 14), protein engineering and structural analysis (2), molecular modeling and in silico screening (3, 12), enzyme assay development (6, 9) high-throughput screening (9), pharmacokinetics, toxicology and in vivo testing (1, 7, 9, 10, 15) and project management (8). All of these aspects contributed to the overall success of the project.
Project Results:
(SEE FIGURES AND TABLES IN REPORT ATTACHED)
The drug discovery strategy of the A-ParaDDisE project involved both the direct screening of compound libraries composed of HME inhibitors (phenotypic screening), and the screening of individual HMEs, either in silico by docking compound structures into homology models of parasite HMEs, or by high-throughput screening of compound libraries. Both strategies have yielded compounds that are active against parasites maintained in culture that were fed into the drug development pipeline. However, this process involved a variety of steps necessary to the overall success of the strategy. This summary describes these steps, and the principal results obtained.
4.1.3.1. Structure-based screening strategy.
Two working hypotheses underpinned the possibility of targeting individual HMEs. First, individual HMEs would be valid therapeutic targets. Since these enzymes are members of protein families, the different isoforms may have overlapping functions, this requires verification. The second hypothesis is that such HMEs would have structural differences, particularly in the catalytic pocket, that would allow the development of inhibitors that are both isoform and species selective. The choice of individual HMEs to be screened as therapeutic targets was informed by several complementary approaches. These are summarized in Fig. 2
Bioinformatic genomic data mining was used to identify all the possible targets in the genomes of the studied organisms and molecular modeling of catalytic domains identified those that showed significant differences compared to the human counterpart. Gene or transcript knockdown (depending on the parasite studied) allowed the identification of HMEs that are essential to parasite development and survival, particularly in life-cycle stages that occur within the human host. Gene expression signatures, corresponding to the transcriptomes of parasites either treated with HME inhibitors or by HME gene knockout, characterize the biological functions of the parasite that are altered and define the molecular mechanism of action (MMA). Finally, phenotypic screening with HME inhibitors defined classes of HMEs that are promising targets.
4.1.3.1.a. Identification of the targets: mining parasite genomes.
In order to identify the possible HME targets in the parasites 6 species of kinetoplastids, which include 4 species of Leishmania, the Chagas disease agent, Trypanosoma cruzi and the agent of sleeping sickness, Trypanosoma brucei as a reference organism, the malaria parasite Plasmodium falciparum and 3 species of schistosome (Schistosoma mansoni, S. haematobium and S. japonicum) were included in the analysis. In the case of the schistosomes, improvements in genome assembly and annotation allowed an update of the analysis previously carried out during the SEtTReND project. In each case the predicted proteome database for each parasite was analyzed using InterProscan 5, which combines different protein signature methods into a single resource and allows functional analysis of proteins via family classification, domain and site prediction. The mass of data recovered in this way necessitated the construction of the A-ParaDDisE SQL (structured query language) database to allow data storage and recovery. This then became the main source for data-mining with a single table for each species proteome. To identify the HME proteins, significant matches with Hidden Markov Models (HMM) built for the catalytic domains of histone acetyltransferases (HAT) histone deacetylases (HDAC), histone methyltransferases (HMT) or histone demethylases (HDM) were searched using SQL queries. The final results for this search for the protozoan parasites are shown in Fig. 3.
This analysis shows that numbers of HDACs are similar for the protozoans (4 in kinetoplastids and 3 in P. falciparum, and this is also true for the GNAT class HATs with 10-13 in kinetoplastids and 9 in P. falciparum. However, the latter parasite has many fewer HMTs: 7 compared to more than 20 in the kinetoplastids and only one histone demethylase. This suggests that these enzymes may be promising therapeutic targets since their functional redundancy is likely to be reduced.
In addition to this analysis, the proteins containing bromodomains, which interact with acetylated lysine residues, were also identified in all the parasites studied using the same methodology. These turned out to be numerous (21-26 in Leishmania sp., 22 in T. cruzi, 29 in P. falciparum and 22 in S. mansoni) complicating their exploitation as therapeutic targets.
In addition to this enumeration, a phylogenetic analysis was carried out for the Zn2+-dependent class I and II HDACs and the NAD+-dependent class III HDACs, known as sirtuins, in order to identify the orthology relationships of parasite enzymes to their human counterparts and to determine which of them were the most divergent and therefore most likely to allow the development of selective inhibitors. Overall, as illustrated by Fig. 4, which shows the class IIa HDACs, schistosome enzymes were the most conserved. This is expected due to the relative evolutionary closeness of the metazoan platyhelminth parasite to humans, compared to the unicellular protozoans. Nevertheless, whilst certain schistosome HDACs, like the class I enzymes SmHDAC1/2 and SmHDAC3 are highly conserved, SmHDAC8 and all the class II enzymes are more distant. In contrast, all the kinetoplastid class II HDACs and some of the class I enzymes, are on long branches that diverge early in the phylogenetic tree and have no clear orthologous relationships with the human isoforms. This suggests that they may be promising targets.
The analysis of the class III HDACs, which bear no structural or phylogenetic relationship to the class I and II enzymes, and which have an NAD+-dependent catalytic mechanism, showed similar results in terms of the relative phylogenetic distances separating protozoan sirtuins from human isoforms. It also highlighted the absence of class IV sirtuins (Sirt6 and 7) from the kinetoplastids and of class II sirtuins (Sirt4) from schistosomes.
These analyses are the subject of a publication currently in revision [7]. Further to this work an in-depth phylogenetic analysis of Leishmania HMEs has also been carried out.
4.1.3.1.b. Construction of homology models of parasite HDACs
The first X-ray structure of a parasite HDAC, SmHDAC8, was obtained during the SEtTReND project. However, information was lacking on the structures of the HDACs from other parasites. For this reason, homology models of HDACs from Leishmania, T. cruzi and P. falciparum were prepared and compared to HDACs from S. mansoni, other flatworms and humans [8]. In addition, docking and molecular dynamics studies were carried out using known HDAC inhibitors. The general structure of class I HDACs is shown in Fig. 5.
Docking studies showed that, overall, the residues coordinating the Zn2+ ion are perfectly conserved in all the HDACs studied, and residues coordinating the warheads of inhibitors are also conserved. Differences are observed in the lower part of the substrate-binding tunnel and at the entrance of the so-called foot pocket (Fig. 6), which is not always present, and particularly in the walls of the binding pocket tunnel and the entrance of the side-pocket. In this way, a catalogue of differences has been compiled that could be exploited in the design of selective inhibitors.
Docking studies with HDAC inhibitors with reported activity against the parasites supported the mechanism of action of these compounds as acting selectively on HDACs. For example, HDAC inhibitors such as SAHA, Trichostatin A and the selective HDAC6 inhibitor Tubastatin A, as well as other compounds, could be docked to homology models of L. braziliensis and L. major HDACs [8]. However, this was not always the case. The HDAC8 selective inhibitor PCI-34051 could be docked into the homology models of Leishmania HDACs 2 and 4 (Fig. 7), but not HDACs 1 and 3. A homology model of the T. cruzi HDAC2 (TcDAC2) was used for an in silico docking screen to identify potential inhibitors (see below).
Interestingly, all inhibitors developed against SmHDAC8 could be docked into platyhelminth HDAC8 homology models and the docking poses were similar due to the high level of similarity of the binding pockets. This suggests that the inhibitors developed against the schistosome enzyme might also inhibit the orthologues in other schistosome species, as well as those in Clonorchis sinensis and in the Echinococcus, Taenia and Hymenolepis genera.
4.1.3.1.c. Target validation and characterization
In order to validate individual HMEs as therapeutic targets, we have carried out gene or transcript knockdown studies, characterized in detail the functional activity of certain HMEs and investigated the effects on gene transcription in parasites either treated with HME inhibitors or knocked down for an HME in order to elucidate molecular mechanisms of action.
i) Target validation
During the previous SEtTReND project, transcript invalidation of a large number S. mansoni HMEs using RNA interference (RNAi) showed that three were essential to the survival and/or development of the parasite within its mammalian host. These were SmHDAC8, the histone demethylase SmLSD1 and the arginine methyltransferase SmPRMT3. Studies on these three HMEs have continued during the A-ParaDDisE project.
Of the 50 HMEs encoded in the T. cruzi genome (Fig. 3), 8 were selected for gene knockout studies. These include the four HDACs, two HATs the methyltransferase DOT1A and a JmjC family histone demethylase. Gene knockdown was carried out principally using a classical genetic recombination strategy. Briefly, an initial knockout in the first allele was done using a construct containing flanking regions of the targeted gene, intergenic regions of two genes and a selectable marker for hygromycin resistance. Confirmed heterozygous knockout parasites were then transfected with a second construct carrying a different selectable antibiotic resistance marker. Homozygous knockouts were selected using both antibiotics.
In this way it was demonstrated that two HDACs, TcDACs 1 and 2, as well as the histone methyltransferase TcDOT1A, were essential for parasite survival and development: it was not possible to obtain viable homozygous knockout parasites. In the case of a further HDAC, TcDAC4, homozygous knockout parasites were viable. However, they did show morphological defects during metacyclogenesis (Fig. 8), indicating that TcDAC4 does fulfill a distinct specific role in T. cruzi. The viability of the TcDAC4-/- parasites allowed the investigation of their transcriptome using RNASeq (see below).
Following the publication of a methodology allowing the application of CRISPR/Cas9 gene invalidation to T. cruzi [9] this is also being applied to obtain knockouts of HME genes. This work will be completed after the end of the A-ParaDDisE project.
Whilst similar studies were not done on Leishmania parasites, a drug target deconvolution study has been undertaken for a compound that was a hit in the screening assays. This compound is tested at increasing concentrations on L. donovani promastigotes transfected with a homologous cosmid library from virulent hamster-derived parasites. Resistant parasites have been isolated and selected cosmids sequenced. They are now under analysis for the identification of the potential target.
ii) HME functional activity
The role of histone modifications in kinetoplastid parasites is not known, particularly since these organisms carry out genome-wide transcription producing polycistronic mRNAs, which are then divided by trans-splicing. The control of protein expression seems to occur mainly at the post-transcriptional level. However, histone marks do exist and as part of the A-ParaDDisE project a mass-spectrometry-based proteomic study was carried out to identify them in T. cruzi [10]. A total of 13 different chemically-distinct modifications were characterized, including some novel and unusual marks such as lysine acylations, Ser/Thr acetylation and N-terminal methylation. Some of the marks correlate in nature and position to those in other eukaryotic organisms, suggesting the existence of similar regulatory mechanisms. Others are unique to trypanosomatids and the enzymes responsible may be drug targets.
Functional characterization of T. cruzi HMEs has involved their subcellular localization using fusions with fluorescent proteins and tags, or immunolocalization. These analyses showed that most are nuclear as expected. However, a cytosolic localization for one HAT was detected, suggesting that it acts on targets other than histones. The analysis of the phenotypes of the T. cruzi HME mutants also gives insights into their functions. In the case of TcDAC2, the alterations include cell division defects, and DNA accumulation.
The functional study of S. mansoni HMEs was focused on SmHDAC8 and SmLSD1, both of which had been shown to be essential to the parasite by RNAi treatment. The use of two LSD1 inhibitors showed that these induced tegumental and muscle abnormalities as well as the presence of apoptotic and necrotic vesicles and granules. Adult schistosomes treated with these inhibitors also showed extensive alterations to their reproductive systems. Death of the parasites after treatment was due to the induction of apoptosis as confirmed by the expression of caspases 3 and 7 and the TUNEL staining of DNA double-strand breaks. That these alterations were due to the inhibition of LSD1 was shown by the resulting hypermethylation of histone 3 at lysine 4 (H3K4), the specific target of this enzyme. This study was completed by RNASeq analysis of inhibitor-treated parasites.
The precise biological roles of HDAC8 remain enigmatic. Relatively few partner proteins and substrates have been identified. We therefore undertook a study to identify partner proteins of SmHDAC8 using a yeast two-hybrid (Y2H) strategy. Using full-length SmHDAC8 as bait a large number of potential partners were identified corresponding to biological processes ranging from cytoskeleton organization to oxidative stress. The interaction of human HDAC8 with actin in the cytoskeleton of smooth muscle cells has already been described [11] and we are therefore investigating more closely the interaction of SmHDAC8 with cytoskeletal regulatory proteins identified in the Y2H experiment. A further function for HDAC8 that has been described concerns the deacetylation of the cohesin complex component SMC3 [12]. The cohesin complex is involved in mitosis, gene expression and DNA repair, forming a “lasso” around DNA strands, and the deacetylation of SMC3 is essential to the release of DNA strands by the complex. SmHDAC8 also deacetylates SMC3 and interacts with another complex component, suggesting that it is an integral part of the complex.
iii) Gene expression signatures
Transcriptional analysis (RNASeq) has been used in order to determine gene expression signatures in three parasites, S. mansoni, T. cruzi and L. braziliensis, under different conditions. The principal objective is to elucidate the processes and signaling pathways affected by treatment of the parasites with HME inhibitors and to determine their molecular mechanisms of action. RNASeq has also been performed following the invalidation of a T. cruzi HDAC with the aim of determining its function.
At the start of the project an RNASeq analysis pipeline was developed in order to analyze the RNASeq data obtained from S. mansoni worms treated with HME inhibitors. The pipeline applies the de novo assembly of transcripts with Trinity tools, without aligning the reads to the genome. Using unpublished S. mansoni male, female and egg RNASeq data previously acquired, gene expression signatures were obtained for these life-cycle stages. These results, highlighting 238 novel protein coding genes and extending the sequences of hundreds of predicted genes already annotated in the S. mansoni genome, were published [13]. Further to this, a local genome browser database was implemented at the University of Sao Paolo (http://schistosoma.usp.br/folders/geneNetwork/) incorporating the genome sequence, annotated genes, transcripts and ChIPseq histone marks and made public.
RNASeq analysis methodology was further improved and automated in order to introduce the use of different quantification tools. It is considered necessary to run different tools on the same dataset and consider only genes listed by all the tools (the main intersection in a Venn diagram) in the final analysis. After exhaustive tests of different programs, the final pipeline is composed of the following steps (some tools used are in parentheses): data download (ftp), data checking (md5sum), data decompression (gunzip and cat), metadata preparation, statistics about the reads (FastQC), adapter and quality trimming (Trimmomatic), read mapping (Tophat2), visualization of the mapped reads (UCSC Genome Browser), counting indexing, counting quantification (Cuffquant, Sailfish, and Kallisto), differential expression analysis (Cuffdiff, EdgeR, and Sleuth), comparison between sets of DEGs, gene enrichment analysis (Cytoscape with Bingo, Revigo, and Ontologizer) and finding normalizer genes. It is important to emphasize that for counting quantification and differential expression analysis steps three sets of tools were used: Metagenes (consolidated transcripts from publicly available RNASeq reads which represent a more complete transcriptome than the Sanger Centre Smp database), the new genome browser and hierarchically clustered heatmaps of differentially expressed genes.
These tools were applied to S. mansoni transcriptomes analyzed after treatment with the pan-HDAC inhibitor, Trichostatin A (TSA), the histone methyltransferase inhibitor GSK343, the SmHDAC8 selective inhibitor TH65 and the LSD1 inhibitor TB7/36.
In schistosomes treated with TSA, the subject of an article currently in press [14], histone acetylation at H3K9 and H3K14, indicating transcriptional activation, was increased while H3K27me3 levels (transcriptional repression) were unaffected. The effect of TSA on gene expression in schistosomula was tested at three different time points and found a marked genome-wide change in the transcription profile. ChIPSeq public data was combined with the RNASeq data and showed that differentially expressed genes with the H3K4me3 (activating) mark at their promoter regions generally showed transcription activation on TSA treatment, whereas those without this mark were mostly down-regulated. The affected processes include an upregulation of DNA replication genes, a down-regulation of 20 out of 22 genes involved in the accumulation of reactive oxygen species and changes in the expression levels of dozens of genes encoding proteins with histone reader motifs. These included SmEED, an orthologue of a member of the PRC2 transcriptional regulatory complex. This complex contains the histone methyltransferase EZH2 and we therefore tested the highly selective EZH2 inhibitor GSK343 for its activity on the parasite. This inhibitor showed a synergistic effect in combination with TSA: a sublethal dose of TSA (1 µM) combined with a sublethal dose of GSK343 (20 µM) led to significant schistosomulum mortality after 3 days in culture. Moreover, when tested alone, GSK343 caused substantial damage to the adult male and female worm teguments and eggs released by treated female worms showed eggshell fissures and were not viable.
Transcriptomes of adult worms and schistosomula treated with the sub-lethal dose of 20 µM GSK343 showed a sex-specific response. Three analysis tools were used and only genes identified by all three as showing expression variations were analyzed. Whilst 607 female and 580 male transcripts were downregulated, and 412 and 532 were upregulated in females and males respectively (Fig. 9), only 54 downregulated genes and 86 upregulated genes were in common. In male worms genes related to membrane constituents, cell homeostasis and drug transport were downregulated. In females, genes involved in egg formation were downregulated, coherent with the effects of the drug on eggs. Interestingly, EED was also downregulated in female worms, as was the case after TSA treatment, as was a key gene involved in stem cell differentiation. These data underline the potential of targeting EZH2 in schistosomes.
tified by all three methods (cutdiff, Sleuth and EdgeR) were retained in the analysis.
In order to help determine why SmHDAC8 is essential for schistosome development and survival in the mammalian host, the selective inhibitor TH65 (see below) was used to treat adult worms and their transcriptomes were then analyzed. At a dose of 20 µM, this inhibitor caused both tegumental damage, including blebbing and loss of tubercles, and effects on the reproductive organs: notably release into the medium of vitelline cells. Due to these severe effects, three doses of TH65 were tested: 5, 10 and 20 µM. Table 1 shows the dose-dependent increase in the numbers of differentially expressed genes.
SmHDAC8 seems to regulate common pathways in both sexes since about 50% of the differentially expressed genes are common between females and males. In female worms genes related to egg development and transmembrane ion transport are the most striking, corroborating the phenotypic effects of TH65 treatment on egg morphology and viability. In males, genes encoding tegument surface proteins are downregulated, again in line with the phenotypic effects of drug treatment. On the other hand, upregulated genes in both sexes included those related to nucleic acid binding, ribosome biogenesis and RNA processing, flagging SmHDAC8 as an important actor in transcriptional and post-transcriptional control of expression. Interestingly, genes encoding proteins identified as partners or substrates of HDAC8 were not affected.
RNASeq analysis of adult worms treated with the LSD1 inhibitor TB7/36 showed that 2709 genes were differentially expressed in males and 938 in females. In addition, 270 genes were differentially expressed in schistosomula treated with the inhibitor. Downregulated genes in male worms included those involved in drug pumps and cytoskeleton formation, whereas in females transmembrane transporter genes were downregulated. Notch signaling pathway genes were upregulated in female worms and schistosomula. Notch pathways have multiple roles including in neuronal function and development, but also in the regulation of cell fate decisions in various human tissues and the role of HDAC8 in maintaining Notch1 stability was recently reported [15].
At this stage of the analysis of the results obtained with different HME inhibitors in schistosomes it is too early to draw hard and fast conclusions. However, some common effects of inhibitors of different enzymes can be identified:
- All the inhibitors studied have striking effects on the morphology of female reproductive tissues and the eggs. This may be partly due to the fact that, in adult worms, these tissues are the most metabolically active and show the most intense cell proliferation. However, these effects are correlated with the downregulation of genes involved in eggshell formation and transmembrane transporters.
- Effects on signaling pathways and genes involved in cell differentiation and cell fate determination are affected. This suggests that HME inhibitors affect parasite development pathways. The effects of TH65 on gene transcription can be correlated with the effect of SmHDAC8 knockdown on schistosome development and survival in mice.
In addition to transcriptomic experiments undertaken on schistosomes, studies were also carried out on HDAC4-/- T. cruzi and on L. braziliensis parasites treated with an HME inhibitor.
RNASeq results indicate that several genes are differentially expressed when TcHDAC4 is not present. More than 50 different genes are up-regulated in T. cruzi HDAC4 null mutant epimastigotes and epimastigotes under nutritional stress, suggesting that, in these stages, histone deacetylation by TcHDAC4 is important to maintain the expression of some genes at a low level. Later on in the metacyclogenesis process (attached differentiating epimatigote cells), lacking of TcHDAC4 causes both up- and down-regulation of almost 2000 genes. Preliminary analysis also indicates that a number of surface protein genes are differentially expressed in T. cruzi HDAC4 null mutant cells. As these surface proteins play a direct role in the T. cruzi infection process, our data indicate that HDAC4 should also be considered as target for T. cruzi drug design to block expression of surface proteins. These results also show that an individual HDAC controls the expression of a specific gene subset, although whether this control is exerted at the transcriptional or post-transcriptional level, via effects on mRNA stability for example, remains to be determined.
We have adopted the Dual-RNA-Seq approach [16] in order to evaluate the effect of two selected HME inhibitors on the gene expression of both Leishmania braziliensis (M2904/GFP strain) and murine macrophages (RAW 264.7). Macrophages were cultivated for 24 hours before infection with Leishmania parasites. After 24h of infection, treatments were performed for 48 hours with either the HDAC inhibitor TH85, the HMT inhibitor BSF2 or with vehicle (DMSO) for the control. After treatment, Leishmania-infected macrophages were collected, total RNA from the cells was extracted, quantified and checked for quality and integrity. The experiments were performed with 4 biological replicates each for a total of 12 samples. The extracted RNAs were of high quality and were used for the generation of Illumina cDNA libraries. The sequencing was very balanced, such that approximately 60 million reads were obtained for each sample. Preliminary mapping analysis showed that around 1.5% of the reads from each sample map onto the genome of Leishmania (approx. 1 million reads for each sample). Differential gene expression analyses are being performed now and will help to elucidate the role of histone modifying enzymes on gene expression programs in Leishmania.
4.1.3.1.d. Production of recombinant HMEs and structural determination
Recombinant HMEs have been produced with two main objectives: to supply enzymatically active proteins to allow high-throughput inhibitor screening and to produce protein crystals for X-ray crystallography and the solution of the 3D structures of HMEs alone or in complexes with inhibitors. This in turn would permit the in silico screening of virtual compound libraries. Moreover, the study of crystal structures of inhibitors bound to the active site of SmHDAC8 has been instrumental in the development of selective inhibitors of this enzyme.
i) Production of recombinant HMEs
During the project the production of a variety of HME recombinant proteins from all the parasite species studied has been attempted. These include 7 HDACs, one sirtuin (class III HDAC), one HAT, two PRMTs and two HDMs. Some of these proteins proved difficult to produce in quantity or in an intact form, for instance the S. mansoni LSD1 protein. The production of bromodomain-containing proteins was put on hold in the absence of their validation of targets. In addition, the possibility of reconstituting complexes between target HMEs and regulatory subunits has been investigated. The main motivations for this were: (i) the stabilization of HMEs through complex formation and (ii) the reconstitution of a biological context for the exploration of HME activity and inhibition. However, a major hurdle has been the correct identification of the partner subunits, for which clear orthologues of human counterparts may not exist, particularly in the case of protozoan parasites. Stabilization of HMEs has also been explored through protein engineering: using shortened constructs, point mutations and the removal of insertions. Many of the parasite HMEs contain insertions within their catalytic domains.
The principal HME proteins produced were:
Arginine methyltransferases (PRMTs). Two S. mansoni PRMTs have been studied, SmPRMT1 and SmPRMT3. The latter was validated as a therapeutic target using RNAi transcript knockdown and was produced in a catalytically active form. This protein was then used for high-throughput inhibitor screening. The production of diffracting crystals proved difficult, however, and current efforts are directed at the coexpression of SmPRMT3 with a putative protein partner, the ribosomal protein RPS2.
Lysine demethylases. Following difficulties in obtaining SmLSD1 in an intact form, another schistosome histone demethylase, SmUtx was studied. This enzyme is the only schistosome demethylase capable of demethylating H3K27me3, a key repressive histone modification, and this enzyme may represent a uniquely sensitive target in the parasite. This has been successful (Fig. 10), but it has not so far been possible to measure the enzyme activity of the recombinant protein, perhaps due to the incompatibility of the human substrate used.
Histone acetyltransferases. The T. cruzi histone acetyltransferase TcHAT1 has been produced in a stable form following protein engineering to remove some insertions that form unstructured loops. The resulting recombinant protein is enzymatically active and has been used for tests of potential inhibitors.
Histone deacetylases. Production of the enzymatically active S. mansoni sirtuin SmSirt2 has continued with the aim of exploring the selective inhibition of this enzyme. Concerning the zinc-dependent class I and II HDACs the main strategy has been to expedite the search for anti-parasitic drugs using FDA-approved scaffolds for designing selective drug leads. With this aim the production of large quantities of several HDACs has been carried out for work both within and outside the A-ParaDDisE consortium.
A major effort has gone into producing L. braziliensis HDAC3. The full-length protein expressed on its own poses problems of degradation and aggregation. The removal of a large insertion within the catalytic domain solved some of the problems of degradation and increased yields. However, its propensity to aggregate was not diminished. Putative protein partners (SANT-domain containing proteins) were identified and cloned, but their coexpression with LbDAC3 did not solve the problem and work on this protein was stopped.
Work on the selective inhibition of SmHDAC8 has continued and intensified during the A-ParaDDisE project. Very large quantities of SmHDAC8 have been produced and have been used for biophysical and crystallization studies, as well as for inhibitor testing (determination of IC50 values). The enzyme has also been distributed to groups outside the consortium in collaborations on the selective inhibition of SmHDAC8. Mutational analysis has also been carried out in order to determine the eventual importance of catalytic site loops. It should be noted that the work on the selective inhibition of SmHDAC8 has required the comparison with human HDAC8 (hHDAC8). Large amounts of the latter have also been produced for work within and outside the consortium.
A potential therapeutic target of major interest identified during the A-ParaDDisE project is HDAC2 from T. cruzi (TcDAC2). TcDAC2-/- parasites are not viable, showing that this enzyme is essential to survival. It was initially produced in as a full-length protein in an enzymatically active form (Fig. 11). This protein was used for initial studies of its activity and inhibition but its crystallization proved difficult to reproduce. A systematic study combining protein engineering with biophysical studies (see below) have yielded a protein with markedly improved stability that has been used successfully for reproducible crystallization.
ii) Biophysical studies
The simple use of SDS-PAGE analysis of recombinant proteins to assess degradation and yields is insufficient as a predictor of their enzymatic activity, inhibition and stability with a view to their crystallization. Various biophysical methods were therefore brought to bear and, although these methods sometimes generated different results, the integration of the analyses enabled choices of constructs and inhibitors for further structural characterization:
- Dynamic Light Scattering (DLS). This has been used to assess the aggregation state of HMEs and their possible stabilization through the use of small molecules such as inhibitors.
- Thermal Shift Assay (TSA). TSA provides similar, but more precise information than DLS and has been use to assess the stability of different HME constructs produced. It has also been used extensively to determine the (de)stabilization of HMEs by mutations and inhibitors.
- Isothermal Titration Calorimetry (ITC). This has been used to determine thermodynamic parameters of the binding of inhibitors to HMEs (Fig. 12). The Kd values obtained by ITC could be compared to IC50 values measured in order to show up similarities or dissimilarities. ITC was employed both for mutated and non-mutated proteins, providing a fine measure of the effects of the mutations on inhibitor binding. This technique is complex to set up, however, meaning that only a subset of HMEs, mutants and inhibitors was considered.
iii) Structural studies of HMEs and their inhibitors
The structural characterization of HMEs and HME/inhibitor complexes has been a major objective of the A-ParaDDisE project. This aspect has depended on the success of protein production and the support of biophysical studies.
SmPRMT3. This enzyme has been produced as a full-length protein, but the best results were obtained using a catalytic core construct. Although crystals were obtained using both recombinant proteins, they failed to diffract adequately for structural determinations (Table 2).
The co-crystallization of SmPRMT3 with its protein partner SmRps2 is under study with the aim of improving crystal quality.
TcHAT1. This protein has a tendency to aggregate, but this was reduced by adding the enzyme cofactor acetyl-CoA. However, crystals could not be obtained, whatever the construct used. Attempts will be renewed when a substrate of the enzyme has been developed to enable co-crystallization experiments.
SmSirt2. This enzyme has also proved to be refractory to crystallization. The development of selective inhibitors should allow co-crystallization studies to go ahead in the near future.
SmHDAC8. The crystal structure of this enzyme was solved during the previous SEtTReND project [6]. During the A-ParaDDisE project a large number of inhibitors of SmHDAC8 have been generated within the consortium (see below), particularly using a strategy of bioguided optimization. The best compounds, as assessed by their low nM IC50 values and/or their effect on the parasite in culture, were used to solve the structures of numerous SmHDAC8/inhibitor complexes.
In particular, the solved crystal structure of SmHDAC8 bound to an inhibitor has revealed the mode of binding of these inhibitors to the enzyme via a schistosome-specific clamp [17] (Fig. 13). This structure also explains the strongly increased selectivity of these inhibitors for SmHDAC8 and their loss of activity against most human HDAC isoforms, apart from hHDAC8. Nevertheless, selectivity against the latter has also been achieved (up to about 8-fold) and much of the structural analysis has aimed at achieving a better understanding of the determinants for these observations.
The work on smHDAC8 selective inhibition has also been carried out by using known human HDAC8-selective inhibitors. A large amount of biochemical and biophysical data has been collected on these inhibitors as well as structural data (manuscript in preparation). Interestingly, these data further refine our view on HDAC8 selective inhibition and open new avenues towards the selective inhibition of other HDACs as well as the comprehension of their catalytic mechanisms and target recognition.
In addition, many other compounds have been received from within and outside the consortium. Structural data was also collected on many complexes of these inhibitors with SmHDAC8, providing a wealth of further structural data (manuscripts in preparation).
TcDAC2. Large quantities of enzymatically active full-length TcDAC2 were prepared, but this protein failed to crystallize reproducibly. Major protein engineering involving the partial removal of insertions, backed up by biophysical methods to determine the stability of the resulting constructs in the presence of inhibitors allowed a significant increase in protein stability. This allowed the growth of large, well-diffracting crystals. Structure determination is underway.
The process and methodologies leading to this latter result are highly original and are the subject of a patent application under preparation. No further details can be given here.
4.1.3.1.e. In silico and high-throughput screening of HMEs
i) Virtual screening of HME inhibitors.
In silico docking studies were done using generated homology models:
- Hydroxamic acid-containing SmHDAC8 inhibitors previously reported [6, 17] were docked into platyhelminth HDAC8 homology models. The docking protocol established for SmHDAC8 was used. The catalytic pockets of the HDAC8 enzymes from other schistosome species, from other trematodes and from cestodes were highly similar and all the inhibitors could be docked into the homology models with similar docking poses. The hydroxamic acid residue was always coordinated by the zinc ion and a conserved tyrosine and two histidine residues. The derived data suggest that SmHDAC8 inhibitors may also inhibit the enzymes of these phylogenetically related species and that a drug targeting SmHDAC8 would also be effective in these other species.
- The same compound libraries were docked into HDAC homology models of TcDAC2, L. braziliensis LbDAC2 and the three P. falciparum HDACs, PfDACs 1, 3 and 3 with the object of selecting compounds for phenotypic screening. Virtual screening was also carried out on a bromodomain from a recently-described essential P. falciparum protein [18]. Again, hits were subjected to phenotypic screening.
- The homology model derived for TcDAC2 was used to explain the binding modes of inhibitors tested in an enzyme assay. The results suggested starting points for chemical optimization of inhibitors. These screens also suggested that human HDAC inhibitors currently in clinical trials for cancer treatment might represent starting points for lead optimization.
- A homology model of SmSirt2 was also generated for docking studies. In order to take into account the conformational plasticity of this protein, several available conformations/structures of human Sirt2 were used in model generation. Several compound libraries including the GSK-PKIS kinase inhibitor data set, the GSK Malaria Box, KinetoBox and hits from HTS screening during the SEtTReND project, were docked and interaction/binding modes evaluated. The derived models suggest that active compounds interact with a large cavity known as the extended C-site and selectivity pocket.
ii) Enzyme assay development and enzyme inhibition-based screening
Enzyme assays have been developed for a variety of HMEs produced as recombinant proteins. These have been utilized to screen inhibitors identified from virtual compound screening and high-throughput screening and in the following iterative process where optimized inhibitors have been developed by means of synthetic chemistry.
- In vitro assays for SmHDAC8, TcDAC2 and SmSirt2 were developed and large numbers of inhibitors screened. In the cases of TcHAT1 and SmUtx assay development is ongoing.
- It has recently been demonstrated that some sirtuins are fatty acylases in addition to their deacetylase activity [19]. For SmSirt2 an assay was developed showing that the enzyme deacylates substrates with long-chain fatty acids linked to a fluorogenic substrate. This assay will be used to determine SmSirt2 substrate preference: demyristoylation over deacetylation for example.
- Screening campaigns using the FDA approved drug library and the GSK-PKIS kinase inhibitor set allowed the identification of some hits for SmSirt2 in the low µM range. Two hits from the kinase inhibitor set are now used in optimization campaigns.
Assay development adapted to High-Throughput Screening (HTS) was also undertaken:
- For SmPRMT3 screening an assay for primary screening was developed based on the Cisbio EPIgeneous Methyltransferase assay, the enzyme activity being measured by the conversion of the S-adenosylmethionine (SAM) methyl donor to SAH in a homogeneous time-resolved FRET assay (HTRF; Fig. 14). Dose-response curves were determined using both the Cisbio EPIgeneous Methyltransferase assay and another commercial assay: the Perkin-Elmer LANCE assay. Both assays had to be adapted to use in 384-well HTS plates.
- For TcDAC2 HTS an assay was developed and formatted to 384-well plates using the recombinant TcDAC2 protein. The assay measures fluorescence intensity using a tripeptidic trifluoro-lysine substrate with a coumarin fluorophore (Ac-Leu-Gly-(TFA)Lys-AMC). Further optimization resulted in an assay protocol with excellent statistics to meet the requirements for HTS of the European Lead Factory (see below) who will carry out the HTS. This assay was also used in order to evaluate HDAC8 inhibitors against TcDAC2 and to test different TcDAC2 protein constructs.
iii) High-throughput screening of HMEs
HTS was performed against SmPRMT3 by screening a diversity set of 79696 distinct compounds at 20 µM concentration. The HTS went technically well (Z’-factor 0.73) and yielded 114 primary hits, although many were very weak. Follow-up assays of the primary hits including quality control and determination of dose-responses showed that many of them interfered with the assay read-out and were therefore false-positives. Three compounds showed dose-responses in both available SmPRMT3 assays. One was pentamidine, an old drug previously used against leishmaniasis but which has severe side-effects. Small chemistry programs around the other two hits were initiated and 27 analogues were tested, but none were active. The SmPRMT3 chemistry program was therefore stopped. In addition 66 potential PRMT inhibitors generated within the consortium were tested, but none were active.
TcDAC2 is a priority therapeutic target for the A-ParaDDisE consortium: it is essential for the survival of T. cruzi, the enzyme has been produced in quantity and in an enzymatically active form, an HTS-compatible assay has been developed and diffracting crystals have been produced, indicating that the detailed catalytic site structure will be available in the near future. Since SmHDAC8 had already been screened against the compound diversity set available within the consortium, it was decided to submit a project to the European Lead Factory (ELF) in order to screen TcDAC2 against the extensive ELF library derived from a variety of academic and Pharma industry sources. This project was accepted and is currently underway. The ELF will supply a qualified hit list of a maximum of 50 of the best inhibitors for testing against the parasite, docking studies, compound optimization and lead development.
HTS campaigns against SmHDAC8 and SmSirt2 performed in the SEtTReND project provided hits which during the A-PARADDISE project have been optimized in an iterative process involving synthesis of >500 compounds, biochemical assays, phenotypic assays, in-vivo PK studies and an in-vivo efficacy study.
4.1.1.1. Phenotypic screening of focused inhibitor libraries
Phenotypic screening using focused inhibitor libraries, supplied by the “chemist” partners in the A-ParaDDisE project, formed a major part of the overall drug development approach. More than 500 compounds were tested against S. mansoni (in two screening centers), P. falciparum (two screening centers), Leishmania sp. (two centers) and T. cruzi. In general, compounds were initially screened at 10 and 20 µM against cultured parasites and compounds giving >50% growth inhibition or mortality were further tested to determine EC50 values and, in the case of S. mansoni, against different life-cycle stages. The compounds tested included inhibitors of HDACs, sirtuins, HATs, HMTs, HDMs and also DNA methyltransferase (DNMT) inhibitors and bromodomain inhibitors. Inhibitors were designated as ‘hits’ according to the ‘Katsuno criteria’ [20]. For kinetoplastid parasites, for instance, this implies an EC50 value <10 µM and a selectivity index (for toxicity on cell lines) >10. For P. falciparum these criteria are more stringent (EC50 < 1µM, SI>100). For schistosomes criteria were not defined, but we have decided to use the same criteria as for the kinetoplastids.
4.1.3.2.a. Phenotypic screening against S. mansoni
Phenotypic screening against schistosomes is well known to present problems, with different methods sometimes giving radically different results [21]. For this reason, S. mansoni phenotypic screening was carried out in two centers using different methods. Primary screening against the schistosomula larvae using a fluorescence-based method measuring mitochondrial function via the reduction of resazurin (Alamar blue) to resafurin (fluorescent) and a secondary visual screen measuring the maintenance of adult worm pairing and egg laying was carried out in one center. In another center, primary screening on schistosomula was carried out with a propidium iodide exclusion assay and secondary screens measuring lactate production by schistosomula and adult worm motility were carried out in a second center.
In the first center, 36 compounds were selected for supplementary screening on adult worms, of which 11 have so far been found to induce complete separation of worm pairs at 20 µM. Of the compounds for which EC50 values were determined, four were <10 µM and are considered the best candidates for in vivo testing.
Comparison of the results of the tests on schistosomula and adult worms in the second center showed that, of a total of 132 compounds testing positive in one or other of the tests carried out, only 33 were positive in all tests (Fig. 15). Of these, 8 compounds had selectivity indices >20 and were considered the best candidates for in vivo testing. Only two of these were in common with the compounds selected in the first center.
4.1.3.2.b. Phenotypic screening against Leishmania
Due to the variability between different Leishmania species, phenotypic assays were carried out on L. amazonensis and L. donovani in one center and on L. braziliensis in a second center. Tests were carried out on both free-living promastigotes and on intramacrophagic amastigotes.
In the first center, tests on L. amazonensis and L. donovani promastigotes were done using an Alamar blue viability assay and on L. amazonensis amastigotes using a high-content cell-based assay allowing simultaneous evaluation of effects on parasite load and on macrophage viability. Tests on promastigotes yielded 47 compounds active at 10 µM on promastigotes of both species. In contrast, only 22 compounds were active on intracellular amastigotes. Of these, 8 compounds were active both on promastigotes at 10 µM and on amastigotes at 1 µM. Five compounds were putative DNMT inhibitors and one other was an HDM inhibitor. Analogues of active compounds were synthesized and supplied for testing, yielding further active compounds. The best compounds showed activity in the low µM range.
Tests on L. braziliensis promastigotes and intramacrophagic amastigotes yielded fewer compounds that were active against the parasite and simultaneously showed low toxicity to the macrophages. Only three compounds, both HDAC inhibitors, were selected as candidates for in vivo studies. Two were also selected for studies on their effects on the parasite transcriptome.
4.1.3.2.c. Phenotypic screening against T. cruzi
Compounds were tested using the Tulahuen strain of T. cruzi engineered to express E. coli β–galactosidase. Infective trypomastigotes were used to infect primary cultures on L929 fibroblasts in microtiter plates. Parasite viability after 4 days of drug treatment was measured by quantifying the β–galactosidase activity in the culture lysate. L929 cell viability was measured separately, allowing the calculation of selectivity indices.
Of the compounds tested 170 showed some activity against the parasite, the inhibitor specificity of which is shown in Fig. 16. Inhibitors of all classes showed activity, but interestingly inhibitors of HDMs, including LSD1 were overrepresented among the positive compounds (26%) compared to the numbers tested (18% of all compounds).
Of the compounds with activity against the parasite, only 10 meet the Katsuno criteria, three compounds have EC50 values of <5 µM, which is comparable to that of the currently used drug benznidazole (3.8 µM), and SIs >10.
4.1.3.2.d. Phenotypic screening against P. falciparum
Compounds were tested for anti-malarial activity against two strains of P. falciparum intraerythrocytic stages maintained in culture, the chloroquine-sensitive 3D7 strain and the resistant Dd2 strain. Further testing on some ‘hits’ was also done using the resistant W2 strain. Two testing methods were used. For the 3D7 and W2 strains parasitemia was detected by flow cytometry after Sybr Green staining and for the Dd2 strain, parasitemia was measured after 3H-hypoxanthine incorporation.
A large number of compounds proved positive according to the Katsuno criteria initially used and a more stringent cut-off with an EC50 value of <100 nM was subsequently used. Nevertheless, 47 compounds were positive at this level against the 3D7 strain and 28 against the Dd2 strain. The best compounds had EC50 values of < 10 nM. Moreover, results obtained using different strains were broadly coherent. For example, the most active compound had an EC50 of 4.2 nM in strain 3D7 and 6.6 nM in strain W2. From all these results, 28 compounds were classified as having the most potent activity and the 8 best compounds were placed in the pipeline for in vivo testing.
4.1.1.2. Bioguided optimization of HME inhibitors
The optimization of HME inhibitors for their capacity both to inhibit their target enzymes and for their toxic effects on the parasites has been initiated for several compound series:
- The structure based optimization of TcDAC2 hydroxamate inhibitors using the homology model. Docking and ranking led to a subset of ~30 compounds that were tested against the enzyme in vitro.
- A homology model generated for L. braziliensis DAC2 was docked with hydroxamates of the TH series [17] and structural optimization led to the selection of one compound which proved active in phenotypic screening against the parasite.
- A homology model for P. falciparum DACs 2 and 3 was again docked with TH series hydroxamates, leading to the testing of one compound that gave EC50 values of 50 and 70 nM against the 3D7 and Dd2 strains respectively.
- More generally, the identification of hits by phenotypic screening has led to the synthesis of analogues for further testing for a variety of compound series. These compounds are among those selected for in vivo testing.
The best example of bioguided optimization of a compound series is provided by the TH series of hydroxamate-based SmHDAC8 inhibitors [17]. Compounds identified as isoform-selective inhibitors by in silico screening of a large virtual compound library and validated by enzyme inhibition assays, co-crystallization with the enzyme and tests on schistosomes in culture [6], were used as starting points for the synthesis of compound series with the aim of generating more selective inhibitors, but which were also more potent against the parasite. The starting point for this series was the inhibitor JM1038 (Fig. 17)
The TH series of benzamido-hydroxamates [17] showed increased affinity for the SmHDAC8 catalytic pocket, due to an SmHDAC8-specific clamp, involving specific interactions with catalytic pocket histidine and lysine residues and the flipped-out configuration of phenylalanine 151, which is sterically impeded in the human orthologue (Fig. 18).
Both increasing selectivity and increasing in vitro potency against the parasite has been achieved in this series and in a parallel series (TB) using a different inhibitor as the starting point. The EC50 curves of the two most potent compounds are shown in Fig. 19.
4.1.1.3. In vivo studies on HME inhibitors
During the project 27 compounds were selected, synthesized and submitted for formulation and PK studies. The in vitro evaluation prior to PK studies consisted in the determination of (i) identity, purity, chemical stability and solubility of the compounds, (ii) an LC-MS/MS method to be used for bioanalysis of the compound in the plasma samples and (iii) a formulation suitable for the dosing solution. Of these compounds 23 were found to be suitable for testing and possible to formulate for dosing. These compounds were given to mice both i.v. and orally and sequential plasma and urine samples measured in order to determine plasma compound concentrations, estimate half-life and bioavailability. Using these results compounds have been selected for in vivo experimentation in mouse models of infection for S. mansoni, P. berghei and T. cruzi.
4.1.3.4.a. In vivo testing in S. mansoni-infected mice
Experiments have been carried out with two TH-series SmHDAC8 inhibitors. A single dose of compound (50 mg/kg) administered orally either at day 5 of the infection, targeting the infective larvae (schistosomula) migrating in the blood stream, or at day 35, targeting paired adult worms at the start of egg laying in the mesenteric veins, was ineffective in reducing either the worm burden or eggs recovered from the liver and intestines. Four successive daily oral doses (50 mg/kg) of the compound TH92, either on days 5-8 after infection or on days 30-33, did lead to a partial reduction in both worm and egg burdens in treated animals compared to controls dosed with the vehicle alone. However, these results have proved to be inconsistent. Further experiments will be done with the most potent SmHDAC8 inhibitors, TH143 and TB87 as well as another compound, MC3974, which is also very potent against the parasite in vitro.
4.1.3.4.b. In vivo testing in P. berghei-infected mice
Eight compounds were initially selected for in vivo experiments, but one proved unsuitable. For five compounds P. berghei-infected Balb/c mice (106 parasites) received 8 (twice daily) doses of 50 mg/kg orally. For three compounds, no protection against infection was observed. One compound proved toxic although some protection against infection was also observed. Mice given compound MC3681 all survived and no parasites were present in blood smears 4 days following infection.
Two further compounds, in view of their PK properties, were given as one oral and one i.v. dose daily for four days. However, no protection was observed using these compounds despite their in vitro potency.
The compound MC3681 is a highly promising compound and will be tested further. IP protection is also envisaged for this compound.
4.1.3.4.c. In vivo testing in T. cruzi-infected mice
The most potent compound in the phenotypic compound screen at the time, IN0441411, was tested in 10 mice infected with the T. cruzi DM28c strain modified to express luciferase that were infected with 105 trypomastigotes i.p. Three days after infection, the level of parasitemia was determined by the fluorescence generated after injection of luciferin. The mice were divided into a test and a control group. Compound IN0441411 was injected i.p. at 25 mg/kg every 12 hours from days 4-8 post-infection. Luminescence was tested on days 7 and 9 post-infection. The compound had a growth inhibitory effect on the parasite. At day 7 levels were similar to those at day 3 in the test group, whereas levels had doubled in the control group. At day 9, corresponding to the peak of infection, three of five mice in the test group had low levels of luminescence intensity (Fig. 20). This experiment will be repeated with additional controls, but IN0441411 is a lead candidate for development.
Conclusions
The A-ParaDDisE project has generated a wealth of data showing that many epigenetic enzymes of parasites are bona fide therapeutic targets. Crucially, species selectivity of inhibitors of these enzymes can be achieved through bioguided optimization and these inhibitors can have potent effects on the parasite both in vitro and in vivo. Three compounds have been so far shown to be active in vivo in mouse models of infection and can be considered to be lead candidates for further development. Moreover, more compounds are in the pipeline, including inhibitors that are more potent than those yet tested.
Work on several aspects of the project will continue with the aims of producing high-impact publications, patent applications and stimulating future funding. This work has reached a pivotal phase where continued investment is necessary to bring the results already obtained to fruition.
References
1. Hotez P. et al. (2016) PLoS Negl. Trop. Dis. 10(3):e0003895
2. Mwinzi PN et al. (2015) Front. Public Health, 3:48.
3. Roquis D et al. (2015) PLoS Negl Trop Dis, 9(8):e0003853.
4. Pierce RJ et al. (2012) Curr Pharm Des, 18:3567-78.
5. Andrews KT et al. (2012) Curr Pharm Des, 18:3467-79.
6. Marek M et al. (2013) PLoS Pathog, 9(9):e1003645.
7. Scholte L et al. (2017) Submitted.
8. Melesina J et al. (2015) J Mol Graph Model, 62:342-61.
9. Peng D et al. (2014) MBio, 6(1):e02097.
10. Picchi GF et al. (2016) J Proteome Res, 16:1167-79.
11. Weltregny D et al. (2004) Am J Pathol, 165:553-64.
12. Deardorff MA et al. (2012) Nature, 489:313-7.
13. Anderson L et al. (2015) PLoS Negl Trop Dis, 9(12):e0004334.
14. Anderson L et al. (2017) PLoS Negl Trop Dis, in press.
15. Chao MW et al. (2016) Oncotarget, 7:1796-807.
16. Fiebig M et al. (2015) PLoS Pathog, 11(10):e1005186.
17. Heimburg T et al. (2016) J Med Chem, 59:2423-35.
18. Josling GA et al. (2015) Cell Host Microbe, 17, 741-51.
19. Zhang X et al. (2016) Nat Chem Biol, 12:614-20.
20. Katsuno K. et al. (2015) Nat. Rev. Drug Discov. 14:751-758.
21. Panic G et al. (2015) Parasit Vectors, 8:624.
Potential Impact:
The aim of the A-ParaDDisE project was to develop new drug candidates for use in treating four neglected parasitic diseases, schistosomiasis, leishmaniasis, Chagas disease and malaria, by targeting enzymes (histone modifying enzymes) and other proteins involved in epigenetic processes that regulate gene transcription in these organisms. The project is built on a strategy developed in a previous, successful EC-funded project, SEtTReND, which focused on schistosomiasis, expanding the number of teams participating to create an innovative platform for drug discovery. In all 17 teams participated in the project and included those involved in the chemical synthesis of epigenetic inhibitors combined with others that carried out phenotypic screening on the parasites, molecular modeling and virtual screening, recombinant protein production and structural analysis, high-throughput screening, target validation and characterization, pharmacokinetics and in vivo testing. In this way inhibitors of histone modifying enzymes (HME) from focused inhibitor libraries or identified by target-based screening methods have been characterized as novel compounds with potent activity against the parasite in vitro and in vivo. The working hypothesis that, despite the evolutionary conservation of epigenetic processes, inhibitors could be selected or developed that show species selectivity has been validated. A number of compounds that have activity in vivo (mouse infection models) have already been identified and can be considered as lead compounds, justifying further development through compound optimization and pre-clinical testing. Further compounds that have potent effects on the parasites in culture are in the development pipeline. One challenge facing the teams involved in the project is to attract collaborations and funding that will allow the continued development of these drug candidates through pre-clinical and clinical testing.
Key assets of our approach were set out in the impact statement of the initial application and have been confirmed during the project. These are:
- Proven concept. The extension of the approach developed in the SEtTReND project, which focused on schistosomiasis, to diseases caused by evolutionarily distant protozoan parasites with very different etiologies, has been fully justified. The combination of both target and structure based HME inhibitor identification with the phenotypic screening of focused inhibitor libraries made up of compounds initially developed against other pathologies such as cancer, has allowed the identification of compounds active against all the parasites studied.
- Different parasites, the same enzyme families. Genome data-mining showed that all these parasites indeed have HMEs in all the families studied in the project: histone acetyltransferases, histone deacetylases, histone methyltransferases and histone demethylases as well as bromodomain proteins and DNA methyltransferases. Moreover, inhibitors of the main HME types are toxic to the parasites in culture and gene knockout studies carried out in T. cruzi have shown that some individual HMEs are essential for parasite survival, as was shown for some S. mansoni HMEs by transcript knockdown.
- Inhibitors active against related species. Homology modeling of HMEs, particularly HDACs, from the species under study show that some do show cross-species similarity. This is the case, for example, for HDAC8 from a variety of parasitic flatworms including both trematodes and cestodes. However, there are significant differences between the metazoan flatworms and the protozoa, and between different protozoan parasites. It is not probable that an inhibitor developed against an HDAC from T. cruzi would be effective against Leishmania parasites for example.
4.1.4.1. Strategic Impact
4.1.4.1.a. Innovation and Scientific Impact
The scientific impact of the project, both within the domain of parasitology and more widely in the areas of drug development and epigenetics, has been considerable. The combination of the structural studies with compound screening and the exploration of the roles of HMEs has generated considerable interest in the scientific community, through high-impact publications, presentations in international conferences and other dissemination activities. Again, the principal areas of projected impact set out in the project proposal are revisited below to define the impact generated during the project.
- Identification and characterization of parasite HMEs. The identification and characterization of the enzymes involved in histone acetylation, deacetylation, methylation and demethylation, as well as bromodomain proteins, has generated a wealth of data that has been used in the project to identify potential drug targets via phylogenetic studies and molecular modeling. A stringent phylogenetic study of the class I and II zinc-dependent HDACs and the class III HDACs, the sirtuins, has led to the clarification of the sub-class relationships of the parasite enzymes to their human counterparts, the measure of evolutionary distances and the correction of some misconceptions in published research from other groups. This work has also improved annotation and has been added to sequence databases.
- Transcriptomic analyses. The analysis of the effects on parasite transcriptomes of treatment with HME inhibitors, and in the case of T. cruzi, after the invalidation of a gene encoding an HDAC, has already provided crucial insights into metabolic and signaling pathways influenced by HMEs in the parasites. One practical consequence has been the identification of the transcriptional regulation of the expression of a component of a histone methyltransferase complex by histone deacetylases following the analysis of the transcriptome of schistosomes treated with an HDAC inhibitor. This suggested the combination of the HDAC inhibitor with the HMT inhibitor, which led to synergistic effects on the parasite: lower doses of each inhibitor, in combination, lead to parasite killing. Moreover, detailed analysis of the genes up- and downregulated by inhibitor treatment has given indications of the molecular mechanism of action on the parasite and will be utilized to assess eventual off-target effects.
- Comparative structural studies. The study of the X-ray structure of the catalytic domain of SmHDAC8 has yielded important insights into inhibitor binding and the mechanisms leading to isoform and species inhibitor selectivity. This has allowed the development of increasingly selective inhibitors through bioguided optimization and has aroused a great deal of interest from groups outside the A-ParaDDisE consortium, leading to collaborative studies. The work carried out is exemplary and will doubtless serve as a paradigm for the development of selective inhibitors against other targets. The successful crystallization of an HDAC from T. cruzi was possible due to a novel combination of bioengineering methodologies. The possibility of IP protection of this procedure is under consideration. In parallel a comparative study of parasite HME catalytic domains, particularly for the HDACs, has been done using generated homology models. These models have allowed the identification of variable residues within the catalytic pocket and identify divergent enzymes as potential therapeutic targets.
- Novel HME inhibitor structures. During the project more than 500 compounds were tested against the parasites in phenotypic assays, many of which were generated specifically for the project after enzyme activity and structure-based inhibitor optimization studies. Some of these series, such as the hydroxamate-based inhibitor series generated as selective inhibitors of SmHDAC8, have been published and have been taken up by other research groups. For others, the possibility of IP protection is under consideration, according to the conditions laid down in the Consortium Agreement concerning the exploitation of foreground. A linear strategy has allowed the selection of a significant number of compounds to go forward for in vivo studies, based on their activity on parasites in culture and according to established criteria. Three compounds that are well tolerated and are active in vivo are considered as leads for further development and more than 15 compounds are in the pipeline.
- Formulation, pharmacokinetics and in vivo studies. The procedure established for establishing conditions for the formulation of inhibitors for administration in vivo, and the study of the pharmacokinetic properties has allowed the rapid and efficient compound selection.
The scientific impact of the project has been ensured by the effective dissemination of the results via publications, communications and presentations during international conferences (see section 4.2.2 below).
4.1.4.1.b. Societal Impact
The principal societal impact of the work to develop new drugs against neglected parasitic diseases are the populations for which these diseases are leading causes of chronic disability and poverty, as well as being a leading cause of mortality, particularly in childhood. The diseases affect the countries the least able to combat them. The aim of delivering improved treatments will have a dual impact on both public health and living standards in countries in the tropics. Moreover, as well as being a legitimate humane objective for the European community, improved health and economic prospects would, in the long term, lessen the need for aid and ameliorate economic exchanges.
New drugs are required to replace existing treatments where these lead to significant side-effects or where drug resistance is appearing, which is the case for all the diseases studied in the project. In the absence of effective vaccines, and there is currently little prospect of their development, drug treatment remains a crucial arm in control strategies. Moreover, as well as a dearth of effective, well tolerated treatments, the development of drug resistance means that it will be necessary to continuously develop new drugs. For this reason it is important to explore new drug targets and the epigenetic enzymes are particularly promising in this respect.
The health issues posed by the neglected parasitic diseases are of common concern to the endemic countries and the EC. In addition, it is important to consider that these diseases are spreading northwards, possibly due to climate change and increased immigration. Schistosomiasis transmissible to humans has appeared in Corsica and visceral leishmaniasis is endemic in dogs in southern Europe. While it would be wrong to overestimate the impact of these diseases in Europe, this should serve as a reminder that they should not be considered only as problems of developing countries.
In order for the work of the A-ParaDDisE project to have the hoped for societal impact, it will be necessary for the hits and leads discovered to follow the drug development pipeline through lead development, pre-clinical testing and clinical trials. For this financial support will be necessary. Ways in which this can be achieved through a European dimension are discussed in section 4.1.4.
4.1.4.2. Training and Technology Transfer
Training activities in the project have concerned the exchange of students and post-doctoral fellows between beneficiary laboratories and the presentation of their work at international conferences, as well as the organization of a project Workshop. A non-exhaustive list of specific training activities and technology transfers follows:
- Presentations at conferences by project students and post-doctoral fellows were numerous and are listed in the dissemination table. In particular, the FRIAS Epigenetics symposium in Freiburg (April 2016), organized by beneficiary 6, was attended by post-doctoral fellows and students from European and Brazilian beneficiaries.
- Direct training activities during the project concerned PhD students and post-doctoral fellows from beneficiaries 1, 4, 5, 7 and 13 who underwent training in beneficiary laboratories (1, 4, 5, and 6) for enzyme inhibition assays, RNASeq experiments and related technologies.
- Exchanges and discussions between beneficiaries 1, 4, and 5 and 7a concerning HME inhibitor treatment of schistosomes and RNASeq experiments. Similar exchanges were undertaken between beneficiaries 5 and 7b concerning T. cruzi RNASeq experiments.
- Exchanges and discussions concerning phenotypic screening methods applied to schistosomes between beneficiaries 1 and 7a.
- Development and standardization of enzyme assay methods between beneficiaries 2, 6 and 9 concerning TcDAC2.
One important societal impact is also the training of the next generation of scientists in the field of anti-parasitic drug discovery in epigenetics that will be active in the field of research and teaching in the future. Young scientists at different career levels were trained to be specialists in their requisite fields (molecular biology, parasitology, screening etc.) but learned early on to work in a highly interdisciplinary approach and to keep all these different aspects in perspective. One example is a post-doctoral fellow who worked as core support personnel in the project at the beneficiary 6 laboratory. With support from the consortium he received a scholarship from the German Science Foundation to work at the University of Oxford also in the field of drug discovery regarding parasite epigenetics. He will become a junior group leader on his return to Germany and hence will further develop and spread our approach.
A workshop organized by the Consortium and in particular beneficiary 13, was held in Rome in December 2016 on “Epigenetic drug discovery for Neglected Parasitic Diseases”. Held over two days it was attended by 100 participants and presentations were given by 6 invited speakers and 17 speakers who were members of the A-ParaDDisE consortium. All aspects of epigenetic drug discovery were covered from chemical synthesis of inhibitors to structural studies, compound screening, bioguided optimization and the target product profile.
4.1.4.3. European Dimension
A project of the size and ambition of A-ParaDDisE necessitated the participation of numerous partners. These included 10 from Europe, 5 from Brazil and 2 from Australia. The European beneficiaries represented 5 EC member states (France, Germany, Sweden, Italy and the UK). The project required the collaboration of beneficiaries working on all aspects of drug discovery, parasite biology and molecular biology. The structure of the Consortium also ensured access to the parasites endemic to Brazil and maintained in the beneficiary laboratories. The financial support of the EC was essential to the construction of the project: no other comparable funding mechanisms exist for collaborative projects of this scale between European and endemic countries. Moreover, although there is intense competition in the search for drugs against parasitic diseases, apart from other European consortia, no comparable projects exist, for example between the USA and partner countries. This means that consortia like A-ParaDDisE maintain Europe at the forefront of research in the domain.
4.1.4.4. Synergies with other EC-supported consortia and future funding
As well as the A-ParaDDisE consortium, three other consortia, Kindred, NMTrypI and PDE4NPD, were also funded under the same FP7 call. There is some overlap concerning the parasites studied: NMTrypI and Kindred both study kinetoplastid parasites and the PDE4NPD consortium studies helminth infections including schistosomiasis. However, although there are similarities in the approaches used for drug discovery, the chosen targets are different. With the aim of avoiding redundancy of targets and/or compounds studied the four projects share a specific “Synergy” work package and have communicated regularly during the projects. Moreover, a common symposium was organized by the NMTrypI consortium in Modena (June 2016) to allow comparisons of approaches and progress. At this meeting it was evident that all the projects had produced large numbers of ‘hits’ and “leads’ warranting further development and piloting through preclinical and clinical testing. Discussions have since focused on developing a common approach to obtaining future support for the work to be carried out. Two of the actions taken have been to write an article in press in Trends in Parasitology and a letter to the European Commission, both setting out a strategy by which the EC can support future work and remain at the forefront of drug discovery for neglected parasitic diseases. The assessment of the situation and a strategy to remedy it can be summarized as follows:
- The FP-programme has now been completed and replaced by H2020. Under this programme there is provision for vaccine development and support for clinical trials (but only from Phase 2 onwards) in Africa for NTDs. Although this support is significant and important, there is currently no provision for ensuring progress in the work on drug discovery.
- The next logical step of the investment strategy of the EU would be to create a call within H2020, to enable continued interaction between academia and SMEs to create a platform which is capable of continuing the work started by the consortia funded in this final FP7 call.
- This platform would specifically take on the task of bridging the gap between lead optimization and early-stage clinical trials (Phases 1 -2). Lead compounds developed by the four consortia, but also from other sources (e.g. teams participating in the current COST 1307 initiative) would feed the platform regardless of the parasitic disease targeted i.e. moving beyond the current FP7 call and overcoming the problem of fragmentation amongst groups in the field
- As identified in the current FP7 call, not all groups involved in the drug discovery process have the same goals therefore a larger view of this platform would be to divide it into at least two complementary groups with different specificities. One focused on early stage screening, FBDD and innovative ways to design more effective drug development and a second, intent on developing lead compounds towards the clinic from in vivo animal studies, ADMET and GLP/GMP and registration to phase 1. We think that this type of platform structure is highly desired, as the number of candidates populating the different stages of the drug development pipeline needs to be dense to statistically ensure enough drug candidates enter phase 1.
- We see that this as an opportunity to build on the existing consortia, a rich source of knowledge and expertise to create one or more new consortia, in order to drive the various different innovative approaches with one consolidated aim the intent to produce drug candidates appropriate for use in the clinic.
- In a larger vision, once the infrastructure was in place and proven to be productive, this could be a dynamic European/International platform bringing together different partners from academic /industrial/foundations, attracting co-funding from large NGOs, Foundations and Industry. Even partial funding from the EU will facilitate attracting support from these other sources. The Innovative Medicines Initiative (IMI) could become involved in financing partnerships between the pharmaceutical industry and academic groups for drug development for neglected parasitic diseases. Several large pharmaceutical companies (GSK, Merck Serono, Novartis..) are already engaged and contribute to the field.
4.1.4.5. European competitiveness
The A-ParaDDisE consortium included two industrial beneficiaries, the SMEs Kancera and Adlego, as well as a company engaged in intellectual property management and dissemination, Inserm Transfert. Objectives of these beneficiaries included:
- To increase awareness of their expertise within and outside Europe to provide new contacts leading to business opportunities.
- To influence the direction of the project and help focus on the primary objective: drug development.
- In the case of Kancera, to exploit novel drug structures derived from their background for other research projects, notably in cancer.
- To participate in the industrial exploitation of ‘hits’ and ‘leads’ developed during the project. The involvement of large pharmaceutical companies in this aspect, as outlined in the preceding section represents a major objective for progress in neglected parasitic disease drug development and for improving European competitiveness in the domain.
A further contribution to the integration of European research initiatives and competitiveness made by A-ParaDDisE concerns the project submitted to and accepted, after a highly competitive evaluation, by the European Lead Factory (ELF), an organization within the IMI. This project concerns the high-throughput screening of the T. cruzi HDAC TcDAC2, which is an enzyme essential to the growth and survival of the parasite. The recombinant enzyme will be screened against the large compound library of the ELF, which will supply a qualified hit list of a maximum of the 50 best inhibitors. These will be tested by consortium members implicated in the project (beneficiaries 2, 6, 7b and 9) and compounds active on the enzyme and the parasite will be developed. This latter stage of the work will be subject to discussion with the provider(s) of the compound(s) selected and may lead to collaborative projects with pharmaceutical industry partners.
List of Websites:
http://a-paraddise.cebio.org/