Drug discovery against leishmaniasis1268389/FULLTEXT01.pdf · 2018. 12. 20. · Leishmaniasis...

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 265

Drug discovery againstleishmaniasis

Bio- and chemoinformatic guided strategies fortarget evaluation and hit identification

ELISABET VIKEVED

ISSN 1651-6192ISBN 978-91-513-0521-9urn:nbn:se:uu:diva-368499

Dissertation presented at Uppsala University to be publicly examined in A1:107a, BMC,Husargatan 3, Uppsala, Friday, 1 February 2019 at 09:15 for the degree of Doctor ofPhilosophy (Faculty of Pharmacy). The examination will be conducted in English. Facultyexaminer: Professor David Horn (The Wellcome Trust Centre for Anti-Infectives Research,School of Life Sciences, University of Dundee).

AbstractVikeved, E. 2019. Drug discovery against leishmaniasis. Bio- and chemoinformatic guidedstrategies for target evaluation and hit identification. Digital Comprehensive Summaries ofUppsala Dissertations from the Faculty of Pharmacy 265. 67 pp. Uppsala: Acta UniversitatisUpsaliensis. ISBN 978-91-513-0521-9.

Leishmaniasis is a neglected tropical disease mainly affecting poor people in developingcountries. It is caused by infections of flagellated protozoa belonging to genus Leishmania. Thefew available drugs are associated with problems such as low effectiveness, severe side effectsand resistance development. The overall aim of this thesis is to aid in drug discovery againstleishmaniasis – primarily using bio- and chemoinformtic approaches.

In the first part of the thesis potential drug targets in Leishmania parasites were identifiedand hits against these targets were thereafter suggested. In paper I bioinformatics together withexperimental work were used to evaluate lateral gene transfer (LGT) in genus Leishmania. LGTsof prokaryote origin often lack human homologs, and are therefore hypothesized to be valuabledrug targets. LGT in genus Leishmania is shown to be a dynamic process in which some acquiredgenes are conserved in the recipient genomes and others are degraded and eventually lost. SomeLGTs have also undergone pseudogenization. It is thus important to evaluate LGT productsbefore exploring them as potential drug targets.

In paper II ligand-based virtual screening and molecular docking were used to suggestpotential hits against the LGT product pteridine reductase 1 (PTR1) and the two-domain enzymedihydrofolate reductase-thymidylate synthase (DHFR-TS) both involved in folate metabolism.DHFR-TS is not encoded by an LGT but it has been hypothesised that several enzymes in thefolate pathway need to be inhibited to affect the viability of Leishmania parasites. One potentialhit compound against PTR1 and the DHFR-domain and four hit compounds against PTR1 andthe TS-domain were identified and tested on Leishmania tropica promastigotes. The suggestedPTR1/TS inhibitors had no effect in the promastigote assay, however one of them enhanced theeffect of the PTR1/DHFR inhibitor, which also had effect on its own.

In the second part of the project, focus shifted towards predictions of targets for compoundswith known anti-leishmanial activity but unknown mechanisms of actions. In paper III a ligand-based-target fishing (LBTF) method was developed. The reference compounds were metabolitesto metabolic enzymes and similarities were assessed with Euclidean distance calculations inchemical property space. The LBTF approach was used to suggest potential targets to a set ofanti-leishmanial agents retrieved from ChEMBL-database. The theory behind the LBTF methoddeveloped in paper III was also used in paper IV to predict targets of two sponge-derivedalkaloids that where shown to have anti-leishmanial activity.

Keywords: Leishmaniasis, Drug discovery, Lateral gene transfer, Comparative genomics,Virtual screening, Target fishing, Marine natural products

Elisabet Vikeved, Department of Medicinal Chemistry, Box 574, Uppsala University,SE-75123 Uppsala, Sweden.

© Elisabet Vikeved 2019

ISSN 1651-6192ISBN 978-91-513-0521-9urn:nbn:se:uu:diva-368499 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-368499)

”The voices of a million angelscould not express my gratitude

All that I am, and ever hope to be I owe it all to Thee

To God be the glory”

(Andraé Crouch)

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Vikeved, E., Backund, A., Alsmark, C. (2016) The Dynamics of Lateral Gene Transfer in Genus Leishmania - A Route for Adapta-tion and Species Diversification. PLoS Negl Trop Dis, 10(1): e0004326

II Vikeved, E., Sköld, C., Alsmark, C. Multi-targeting the folate path-way is a promising strategy against Leishmania tropica. (Manu-script)

III Vikeved, E., Alsmark, C., Sköld C. Prediction of anti-leishmanial drug targets using metabolite-based target fishing. (Manuscript)

IV Strömstedt, AA., Vikeved, E., Cardenas, P., Alsmark, C., Chen, YH. and Backlund, A. Aaptamines from Haliclona and bromopyrroles from Agelas — marine sponge alkaloids with distinct modes of ac-tion against bacteria and protozoa. (Manuscript)

Reprints were made with permission from the respective publishers.

Contents

Introduction ................................................................................................... 11Neglected tropical diseases ....................................................................... 11Leishmaniasis ........................................................................................... 11

Leishmania parasites ............................................................................ 12Current treatments and their challenges ............................................... 13

Drug discovery against leishmaniasis ....................................................... 14Target validation .................................................................................. 14Phenotypic screening ........................................................................... 19Drug repurposing ................................................................................. 20

Natural products in drug discovery .......................................................... 20

Aims .............................................................................................................. 22

Part I. Identification of valuable targets and searching for potential hits ..... 23Evaluation of LGTs as potential drug targets (paper I) ........................... 23

Methods ................................................................................................ 23Results and Discussion ........................................................................ 25Conclusions .......................................................................................... 30

Identification and evaluation of compounds targeting LGT products and other enzymes in Leishmania folate metabolism (paper II) .................... 31


Part II. Suggesting potential targets for anti-leishmanial hit compounds ..... 38Development of a ligand-based target fishing approach (Paper III) ....... 38


Evaluation of marine alkaloids against Leishmania parasites and predicting possible targets for these (paper IV) ...................................... 47


Concluding remarks ...................................................................................... 54

Populärvetenskaplig sammanfattning ........................................................... 55

Acknowledgment .......................................................................................... 58

References ..................................................................................................... 60

Abbreviations

C. albicans Candida albicans CAI Codon adaptation index CCRF-CEM Blood T lymphoblast cancer cell line CL Cutaneous leishmaniasis DHFR-TS Dihydrofolate reductase-thymidylate synthase DLD-1 Colon epithelial cancer cell line DNDi Drugs for Neglected Disease initiative dUMP Deoxyuridine monophaspahte E. coli Escherichia coli ED Euclidean distance EF-TU Elongation factor thermo unstable HIS Human enzyme inhibitory set HIS-PT Human enzyme inhibitory set (with) predicted targets HMS Human metabolite set HSP 70 Heat shock protein 70 IC50 Half-maximal inhibitory concentration K562 Bone marrow lymphoblast cancer cell line Ka/Ks Non-synonymous/synonymous substitutions L. braziliensis Leishmania braziliensis L. infantum Leishmania infantum L. major Leishmania major L. mexicana Leishmania mexicana L. tarentolae Leishmania tarentolae L. tropica Leishmania tropica LBTF Ligand-based target fishing LDS Leishmanicidal set LGT Lateral gene transfer LGTs Lateral transferred genes LIS Leishmania enzyme inhibitory set LIS-PT Leishmania enzyme inhibitory set (with) predicted targets LNCap Prostate epithelial cancer cell line LMS Leishmania metabolite set MCF7 Mammary gland epithelial cancer cell line MCL Mucocutaneous leishmaniasis MIC Minimum inhibitory concentration Molt 4 Blood T lymphoblast cancer cell line

MTT 3-(4, 5-dimethyl-2-thiazolyl)-2, 5-diphenyl-2H-tetrazolium bromide

NTD Neglected tropical diseases PC Principal component PCR Polymerase chain reactions PKDL Post kala-azar dermal leishmaniasis PTR1 Pteridine reductase 1 S. aureus Staphylococcus aureus T. brucei Trypanosoma brucei T. cruzi Trypanosoma cruzi TMP Thymidine monophosphate VL Visceral leishmaniasis VS Virtual screening WHO World health organization

11

Introduction

Neglected tropical diseases Neglected tropical diseases (NTD) is an umbrella term for twenty infectious diseases affecting more than one billion people every year. NTDs are pri-marily spread in subtropical and tropical countries and the term neglected infer that these diseases mainly affect the poorest of the poor, while they are not a major problem in the developed parts of the world. Therefore, NTDs have neither been prioritized by pharmaceutical companies nor by healthcare [1]. However, interest in these diseases are increasing and the World Health Organization (WHO) have formulated global strategies to combat NTDs [2], some of which have already been proven to be successful. For example, in the mid 1980 around 3.5 million people in 20 countries each year became infected by drinking water contaminated with the parasitic worm Dracuncu-lus medinensis causing a disease called dracunculiasis. Today dracunculiasis is on its way to become eradicated. The disease is only endemic in four countries: Chad, Ethiopia, Mali and South Sudan, and its incidence has de-creased with 99% [3].

Leishmaniasis Leishmaniasis is an NTD caused by Leishmania parasites transmitted by the bite of phlebotomine sand flies. Leishmaniasis is endemic in 98 countries and up to one million people are infected annually [4]. Leishmaniasis is di-vided into New World and Old World depending on the geographical loca-tion of the disease. New World refers to the Americas (North America, Cen-tral America, and South America) and Old World is Africa, Asia and Europe [4-6].

It is a complex disease, with different clinical manifestations depending on the strain of the parasite. Genetic characteristics and immunological sta-tus of the host can also influence the symptoms of the disease. Cutaneous leishmaniasis (CL) causes skin lesions and mucocutaneous leishmaniasis (MCL) gives ulcers in the mouth, nose and throat area. CL can self-heal and is rarely fatal, however the lesions or scars from lesions can cause disfig-urements and social stigmatizations. The ulcers caused by MCL can be widespread and cause obstructions of the airways and finally be life-

12

threatening. Visceral leishmaniasis (VL), or kala-azar, affects the inner or-gans such as spleen, liver and bone marrow and is fatal if untreated. One complication of VL is post-kala-azar dermal leishmaniasis (PKDL). The symptoms of this form of leishmaniasis, are rashes and ulcers. These lesions contain parasites, which can be spread by sand flies to other people – PKDL can thus be a cause for a prolonged high incidence of VL in a community [4, 7, 8].

Most cases of VL occur in the Indian subcontinent, East Africa, and Bra-zil. However, there are also cases of VL in other parts of the world. An VL elimination programme was initiated in 2005 and renewed in 2014 with the aim to eliminate VL in South-East Asia. Elimination means to lower the incidence to less than 1 case per 10 000 people for three consecutive years. The elimination program includes for example early diagnosis and treat-ment; and interruption of transmission through vector control [8]. The elimi-nation program is making progress, however, there is still much work to do before VL is completely eradicated [9].

Leishmania parasites Leishmania parasites are flagellated protozoa (single celled eukaryotic para-sites) harbouring the kinetoplast – a special organelle containing mitochon-drial DNA organised in a tubular network, and therefore, they belong to the kintetoplastids [10]. Until recently, genus Leishmania was divided into three subgenera; Leishmania, Viannia and Sauroleishmania. However, recently a fourth subgenus has been suggested, which is called Mundinia [11]. More than twenty species are pathogenic to humans, and different species may cause different forms of leishmaniasis [7, 12].

Leishmania parasites have a digenetic lifecycle and exist in two main forms – promastigotes and amastigotes (Figure 1). Leishmania pro-mastigotes have an elongated shape and are highly motile. They reside in the gut of the sand fly and are transferred to the mammalian host when the fe-male sand fly takes a blood meal. The promastigotes are phagocytized by phagocytic cells, such as macrophages, and are transformed into amastigotes, which will multiply until the phagocytic cell bursts. The amastigotes can thereafter infect new cells. If an infected mammalian host is bitten by a sand fly, the amastigotes can enter the sand fly and transform into promastigotes [10].

13

Figure 1. Life cycle of Leishmania parasites (CDC/Alexander J. da Silva, PhD/2002).

Current treatments and their challenges The treatment options for leishmaniasis are limited to a few drugs, which are either used alone or in combinations. Different forms of leishmaniasis are treated with different treatment strategies, and the drugs available today are [5, 13-15]:

• Antimonials (sodium stibogluconate and meglumine antimoniate); used against all forms of the disease since the 1920s, however, re-sistance development is emerging.

• Amphotericin B; primarily used against VL, but it is expensive and toxic – patients often need close monitoring and hospitaliza-tion during treatment.

• Miltefosine; the only orally available drug and it is used against VL and CL. Resistance against miltefosine has been observed.

• Paramonomycin; can be administered both parenteral and topical (for CL).

• Pentamidine; rarely used because of low effectiveness and high toxicity.

14

Drug discovery against leishmaniasis Drug discovery against leishmanisis and other NTDs, have previously not been prioritized by pharmaceutical countries, as these diseases mainly affect poor people with no or limited ability to pay for expensive treatments [1, 4]. However, this has started to change, and the Drugs for Neglected Diseases initiative (DNDi) accounts for a substantial amount of this change. DNDi was founded in 2003, and it is a non-profit organization whose main objec-tive is to develop treatments for people suffering from neglected diseases [16].

Drug discovery is a time-consuming and expensive process, which most often initiates with basic research aimed to identify potential hit compounds. These can further be optimised into lead compounds and then developed into approved drugs [17]. The two main approaches for hit identification are target-based approaches and system-based approaches. A definition of tar-get-based approaches is “hypothesis-based approaches that aim to manipu-late a biological system by pharmacologically modulating a specific compo-nent or target (an enzyme, receptor, and so on)”[18]. In system-based ap-proaches effects of compounds on biological systems such as organisms or cell-cultures are evaluated. This approach is referred to as phenotypic high-throughput screening (HTS) if a large number of compounds are tested. HTS can also be applied on target-based approaches. [18, 19].

Neither approach described above is superior to the other. An analysis of the origin of all first-in-class drugs approved by the US Food and Drug Ad-ministration from 1999 – 2013, has revealed that more drugs have been dis-covered using target-based approaches than by using phenotypic screening [18]. However, the results from another similar study have shown the oppo-site [20].

Target validation One important part of target-based approaches is target identification and validation, which is the identification of attractive protein drug targets, such as metabolic enzymes, using experimental models or bioinformatic ap-proaches [17]. In order to be classified as good drug targets in Leishmania parasites, and in other pathogens as well, enzymes should meet certain crite-ria.

One criterion is that the enzyme must be essential for the viability of the parasite [21, 22]. Analysing metabolic pathways of the parasites and identi-fying so-called choke points have identified such targets. Choke-point en-zymes either consumes unique substrates or produces unique products [23]. Targeting such an enzyme would hence result in depletion of a product or excess of a substrate, both potentially affecting the viability of the parasite [24].

15

The second criterion is that the selected target is lacking human homo-logues, as this implies that the enzyme may be inhibited without affecting the human host [21, 22]. A number of such enzymes have been identified after the complete sequencing of the genome of Leishmania major in 2005 [25]. It was at this time confirmed that L. major harbours laterally trans-ferred genes (LTGs) from prokaryote donors [26-28]. Lateral gene transfer (LGT), also known as horizontal gene transfer, is defined as the transfer of genetic material from one genome to another, specifically between divergent lineages or different species [29]. The LTGs are of prokaryote origin in all known cases and therefore they encode proteins, which are either missing in eukaryotes or by evolutionary reasons structurally different from their eu-karyotic equivalents.

A third criterion is that the enzyme is expressed in the infectious form of the parasites [21, 22]. The genes of Leishmania parasites are organized in polycistronic transcription units and gene expression is controlled with regu-latory mechanisms such as mRNA stability, gene copy number and codon usage biases [30-32]. However, knowledge about stage specific gene expres-sion in Leishmania parasites is still incomplete, and it is thus difficult to find attractive drug targets based solely on this criterion.

Hit identification Hits against suitable drug targets can be found experimentally with target-based screening or suggested with virtual screening (VS). VS is in-silico screening of large compound databases aiming to identify compounds that are most likely to have a certain biological activity. VS can be either struc-ture-based or ligand-based [33].

Structure-based VS, primarily molecular docking, simulates interactions between a compound and the binding site of an enzyme. This approach re-quires the 3D structure of the enzyme, either from a crystal structure or from homology modelling [34].

Ligand-based VS is often based on classification techniques or similarity-based approaches. The former are used to derive predictive models based on relationships between the characteristics of compounds and their biological activities [35, 36]. This is done by first dividing a data set with compounds into a training set and a test set. Certain characteristics, in the form of chem-ical descriptors, are then calculated for the compounds and the compounds are labelled according to their biological activity. The training set is thereaf-ter used to build either a linear or a non-linear model. The performance of the model is evaluated by predicting the activity of the compounds in the test set and comparing these to the labelled activities [37]. A well-performing model can subsequently be used to predict biological activities of com-pounds in other data sets.

Similarity based screening is based on “The Similarity Principle” stating that similar compounds most likely have similar biological activities [38].

16

Similarity can, however, be assessed and quantified in many different ways – it is “like beauty more or less in the eye of the beholder” [39]. One com-monly used approach to assess similarity is by comparing the 2D chemical structures of compounds. The 2D chemical structures of compounds can be described with molecular fingerprints representing the presence or absence of particular substructures in the compounds and similarity can thereafter be quantified by calculating coefficients, such as the Tanimoto coefficient. An example of similarity assessment using the Tanimoto coefficient and molec-ular fingerprints is presented in Figure 2 [40-44]

Figure 2. Similarity assessment of compound A and B derived from molecular fin-gerprints and calculation of Tanimoto coefficient.

17

Another, principally different approach for quantifying similarity is compar-ing the physicochemical properties of compounds. The physicochemical properties of compounds, for example size and aromaticity, can be calculat-ed from the chemical structure of the compounds using tools such as ChemGPS-NP. ChemGPS-NP can be explained as a model of chemical property space defined by eight principal components (PC) (Table 1) de-scribing 35 different characteristics derived from the chemical structures of a defined set of compounds. Other compounds can thereafter be positioned in this eight dimensional chemical property space based on their physicochem-ical properties [45, 46].

Table 1. Physicochemical properties associated with each principal component in ChemGPS-NP

PC Physicochemical properties 1 2 3 4 5 6 7 8

size, shape, polarizability aromaticity, conjugation-related properties lipophilicity, polarity, H-bond donor capacity flexibility, rigidity electronegativity, number of nitrogens, halogens and amides number of rings, rotational bonds, amides and OH number of double bonds, oxygen and nitrogens aromatic and aliphatic hydroxyl groups, molecular saturations, Lipinski alert index (drug like index)

[45, 46]

Compounds with similar physicochemical properties are positioned in the vicinity of each other on a ChemGPS-NP projection, which can be quanti-fied with Euclidean distance (ED) calculations [47, 48] (Figure 3).

18

Figure 3. Similarity assessment of compound A and B based on Euclidean distance in chemical property space defined by ChemGPS-NP. The principle component scores (PS) are the positions of the compounds on each principal component.

19

Phenotypic screening Phenotypic screening is the evaluation of biological effects of compounds by testing them directly on an organism or a cell culture. For the screening ap-proach to work efficiently, it is essential that the assay is pharmacological relevant, reproducible, not to expensive, and of high quality [17]. For Leish-mania screening assays it is also necessary to consider which form of the parasite to use, as well as which strain to use because the different disease manifestations are determined by the Leishmania strain infecting the host.

Promastigotes (the form of the parasite inside the sand fly vector) are easy to maintain in cell-cultures and cost-efficient to use, however the pro-mastigote are different from the amastigotes (the form of the parasite inside the host). Intracellular amastigotes can be used for biological testing of compounds, although maintaining cultures of amastigote-infected macro-phages are far more difficult and expensive than culturing free-living pro-mastigotes. Alternatives to intracellular amastigotes are axenic amastigotes. They are transformed by changing the culture conditions – media and tem-perature – of promastigotes. Axenic amastigotes display similar biochemical characteristics as intracellular amastigotes, without being dependent on mac-rophages [49, 50]. All three parasite forms are being used in screening as-says and for evaluation of single/a few number of compounds. However, numerous studies have shown differences between results from pro-mastigote-, axenic amastigote-, and intracellular amastigote screens of the same compounds [51-53]. The results from intracellular amastigote assays are considered as most reliable [53], and it has therefore been proposed that results from promastigote assays should, if possible, always be confirmed in amastigote assays [54].

The primary advantage with phenotypic screening is that it directly pro-vides information about the impact of compounds on the parasite. Phenotyp-ic screening can also suggest new previously unpredicted drug targets, if the targets of active compounds are revealed by additional experiments. The major disadvantage with phenotypic screening is though, that it rarely pro-vides information of the molecular targets of active compounds [55, 56].

Target-fishing One approach to suggest mode of actions of active compounds with un-known targets is to perform ligand-based target fishing (LBTF). LBTF ap-proaches predict target(s) of query compounds by identifying ligands with known protein target(s) that are similar to the query compound. There are thus two essential parts of all LBTF approaches: the availability of reference ligands with known protein targets, and a suitable method for assessing simi-larity [57].

20

Reference ligands are most often exogenous ligands to proteins, which can be acquired from large databases, such as the ChEMBL database. The ChEMBL database contain information about more than 15 million activities and records of 2 million compounds (2018/04/23) [58]. There are, however, problems associated with large databases of biological data, for example incomplete data (all data are not reported); there is a lack of consistency in measured endpoints of experiments and concentration units; and previously available data is recalled. It is also very important to clean and standardize downloaded data before including them in computational analyses. Cleaning can for example be removal of duplicated entries, filtering by properties such as molecular weight and extracting compounds with multiple molecular targets. Standardization can for example infer isotope removal, handling of implicit hydrogens and tautomer generalisation [59-61].

There are several suitable methods for assessing similarity between com-pounds and two major approaches for this has been described above under “Target-based approaches/Hit identification”.

Drug repurposing Another important starting point for drug discovery against leishmaniasis is drug repurposing, which is the discovering of new uses for approved drugs. One advantage of discovering already approved drugs as new anti-leishmanial agents is that the time from discovery to clinical use can be shortened because information about clinical safety and alike is already available. Repurposed drugs can also be less expensive than newly discov-ered drugs as patents most often have expired [62, 63].

Many studies have evaluated anti-leishmanial effects of “old” drugs, for example the anti-fungal agent natamycin disrupts calcium homeostasis and induce cell death in L. donovani promastigotes and intracellular amastigotes [64]; the anti-arrhythmic drug amiodarone has effect against L. infantum promastigotes and amastigotes [65]; and sertraline approved for treatment of depression has been shown to kill L. infantum and L. donovani pro-mastigotes and amastigotes [66, 67].

Natural products in drug discovery Nature has been the source for drugs and treatments for thousands of years – archaeological findings from Neanderthal tombs indicate that modern hu-mans used medicinal plants as early as 58 000 years BC [68]. Medicinal plants are still used, both in traditional medicine and as source for natural products used as drugs in western medicine. Well-known examples of clini-cally used natural products are the malaria medicine artemisinin from Arte-misia annua and the anti-cancer agent paclitaxel from Taxus brevifolia.

21

Plants are not the only source of natural products used in drug discovery. Microorganisms, fungi and animals also offer a plethora of bioactive com-pounds that have been exploited as new potential drugs or as inspirations in drug design. These sources have primarily been terrestrial, however natural products have also been isolated from marine organisms such as plankton, sponges, corals, mollusc, etc. [69-71].

Natural products, both terrestrial and marine, are often used for defence by their producing organism. Hence, they are designed to interact with pro-teins, receptors and other biomolecules in threatening organisms. These pro-found bioactive properties motivate the importance of natural products in drug discovery. Natural products are also explored in drug discovery be-cause they possess a unique and vast chemical diversity, compared to syn-thetic compounds. Comparative studies of structures and properties of syn-thetic, marine- and terrestrial natural products have revealed that synthetic compounds are most often smaller than both types of natural products. There are also differences between terrestrial and marine natural products, for ex-ample the terrestrial natural products display a lower molecular flexibility and marine natural products have more nitrogen atoms and halogens [72-74]

Both terrestrial and marine natural products have been evaluated as agents against leishmaniasis and other NTDs [70, 75, 76].

22

Aims

The overall aim of this thesis was to contribute to drug discovery against leishmaniasis, primarily using informatics-driven approaches. The first part of the project focused on identifying valuable drug targets and thereafter finding hits against these targets. In the second part of the project, focus shifted towards predictions of targets for compounds with known anti-leishmanial activity but unknown mechanisms of actions.

More specifically the aims were:

I Evaluate the potential of LGT products as drug targets by stud-ying the dynamics of LGT in genus Leishmania.

II Select promising LGT products for drug discovery using bioin-formatics and apply in-silico based approaches to suggest hits against selected LGT products.

III Investigate if Leishmania parasites are sensitive to proposed hit compounds by means of biological testing.

IV Develop a LBTF approach for predicting targets for compounds with anti-leishmanial activity.

V Evaluate the activity of selected compounds of marine natural origin on Leishmania parasites and suggest possible targets for these.

23

Part I. Identification of valuable targets and searching for potential hits

Evaluation of LGTs as potential drug targets (paper I) Since their discovery, LGTs from prokaryote donors have been suggested as potential drug targets in Leishmania parasites, as well as in other eukaryotic pathogens [26-28]. However, the dynamics of fixation in their new host ge-nome has not yet been studied. LGT between prokaryotes is transient, some genes are fixed in the host genomes but the majority is momentarily gained and subsequently rapidly lost [29, 77, 78]. Hence LGTs in prokaryotes are most often not attractive drug targets. There are, however, the occasional exceptions to the rule. In paper I, LGT in genus Leishmania was thoroughly evaluated using bioinformatics and a polymerase chain reaction (PCR)-based approach. The purpose was to relate the timing of LGT, and uptake in Leishmania genomes, and elucidate if they are conserved in their recipient genomes and hence attractive targets in drug discovery.

Methods Comparison between universal trypanosomatid and genus Leishmania specific LGTs The presence of orthologs to LGTs, previously confirmed in L. major [26-28], was determined in the genomes of Leishmania braziliensis, Leishmania infantum [79], Leishmania mexicana [80], Leishmania tarentolae [81], Trypanosoma brucei and Trypanosoma cruzi [27, 82] using homology searches. The orthologs found were, depending on their distribution in the different genomes, divided into universal trypanosomatid LGTs and genus Leishmania specific LGTs. The extent of conservation of the LGTs and their orthologs in their respective recipient genomes was estimated by calculating nucleotide and amino acid compositions, and G+C contents.

In depth analysis of genus Leishmania specific LGTs The presence of orthologs to twenty LGTs unique to genus Leishmania was determined in ten additional strains of Leishmania with unpublished ge-nomes (Table 2) employing PCR and sequence determination. Three univer-sal trypanosomatid LGTs were also included for comparison. The orthologs

24

to the LGTs were aligned and subjected to phylogenetic analyses. The orthologs were furthermore manually compared to the sequences of the LGTs in L. major with regards to insertions or deletions of nucleotides, read-ing frame disruption, and presence of premature stop codons.

When annotated orthologs to LGTs could not be detected in one, or sev-eral of the already published genomes of L. braziliensis, L. infantum, L. mex-icana and L. tarentolae, orthologs of LGT surrounding genes were identified by means of homology searches. If these orthologues were syntenic, the intergenic regions between them were aligned and manually inspected, with the purpose of finding remnants of the LGT orthologs.

The evolutionary pressure exerted on the orthologs were estimated by calculating the non-synonymous mutations over the synonymous mutations (Ka/Ks ratio), which in turn were compared to ratios calculated for a set of reference genes – Elongation Factor Thermo Unstable (EF-TU) and Heat Shock Protein 70 (HSP 70). Furthermore, the functionalities of the orthologs were predicted by calculating the Codon Adaptation Index (CAI) scores, which are measures of codon biases in the sequence.

Table 2. Strains of Leishmania parasites included in paper I.

Strain Leishmania braziliensis LB2904 MHOM/BR/75M2904a

Leishmania braziliensis L2346/05

Leishmania braziliensis complex L2237/05e Leishmania panamensis MHOM/PA/71/LS94 Leishmania panamensis L967/96 Leishmania tarentolae Parrot-TarII b Leishmania mexicana MHOM/GT/2001/U1103 c Leishmania mexicana MHOM/BZ/82/BEL21 Leishmania amazonensis HSJD-1 Leishmania major Friedlin d Leishmania major L1989/05 Leishmania tropica Leishmania infantum JPCM5 MCAN/ES/98/LLM-877 a

Leishmania infantum MHOM/00/97/3277 Leishmania donovani LV9 a [79] b [81] c [80] d [25] e The parasite belongs to the braziliensis complex, but the species of the parasite is unknown.

25

Results and Discussion Universal trypanosomatid LGTs are widely distributed and conserved in genus Leishmania As many as 90 LGTs have previously been confirmed in L. major and 59 of these had orthologs in close relatives belonging to genus Trypanosoma. Comparative genomics of trypansomatids have revealed that approximately 6200 genes are shared by L. major, T. bruci and T. cruzi [25]. Thus, less than 1% of the universal trypanosomatid genes are LGTs. However, these genes were with only a few exceptions (in total eight instances of gene loss-es) present in all published Leishmania genomes (Figure 4).

Figure 4. The presence of orthologous LGTs in five sequenced genomes of Leish-mania and the genomes of T. cruzi and T. brucei displayed on a phylogenetic tree, which is in congruence with previous systematics of genus Leishmania [83]. The dark green and light green boxes represent the number of acquired genes. The dark red boxes refer to gene loss of orthologs to universal trypanosomatid LGTs. Light red boxes refer to gene loss of orthologs to LGT unique in genus Leishmania.

The G+C contents, as well as nucleotide and amino acid identities between, the universal trypanosomatid LGT in L. braziliensis, L. major, L. infantum and L. tarentolae resembled the G+C contents and identities between the coding content of the complete genomes (Table 3 and 4). These results all indicate that the universal trypansomatid LGTs are conserved in their recipi-ent genomes.

26

Table 3. Nucleotide and amino acid identity comparisons.

Species compared

Nucleotide identity Amino acid identity

Genome Universal trypanoso-matid LGTs

Genus Leishmania unique LGTs

Genome Universal trypanoso-matid LGTs

Genus Leishmania unique LGTs

L. major – L. infantum

94.0%a 94.0% 92.1% 92.0%a 93.5% 88.8%a

L. major – L. tarentolae

84.9%b 82.9% 77.8% 81.9%b 82.5% 77.8%b

L. major – L. braziliensis

82.0%a 82.2% 77.6% 77.0%a 80.4% 74.5%a

L. infantum – L. tarentolae

85.0% 84.4% 78.0% 82.4%b 84.4% 78.0%

L. infantum – L. braziliensis

81.0%a 82.2% 78.0% 77.0%a 80.5% 74.9%a

L. tarentolae – L. braziliensis

79.2%b 76.7% 68.7% 74.8%b 76.6% 66.8%b

a [79], b [80]

Table 4. G+C content of coding content of complete genomes and in sets of LGTs.

Sequence G+C content in coding sequence

L. braziliensis (genome) 60.4%a Universal trypanosomatid LGTs in L. braziliensis 59.1% Genus Leishmania unique LGTs in L. braziliensis 57.2% L. infantum (genome) 62.5%a Universal trypanosomatid LGTs in L. infantum 61.0% Genus Leishmania unique LGTs in L. infantum 59.1% L. major (genome) 62.5%a Universal trypanosomatid LGTs in L. major 60.9% Genus Leishmania unique LGTs in L. major 59.1% L. tarentolae (genome) 58.4%a Universal trypanosomatid LGTs in L. tarentolae 58.3% Genus Leishmania unique LGTs in L. tarentolae 56.5% a [81]

27

Genus Leishmania unique LGTs have been acquired after trypanosomatid divergence Of the previously confirmed LGTs, 31 were unique to genus Leishmania, which corresponds to 3% of the total number of Leishmania specific genes (less than 1000) [25]. The LTGs unique to genus Leishmania exhibited a more heterogeneous distribution than universal trypanosomatid LGTs, with a total of 16 gene losses (Figure 4).

The G+C contents of the genus Leishmania unique LGTs in L. braziliensis, L. major, L. infantum and L. tarentolae resembled the G+C contents in the coding portions of the complete genomes of these parasites (Table 4). However, the nucleotide and amino acid identities between the genus Leishmania unique LGT in L. braziliensis, L. major, L. infantum and L. tarentolae were significantly lower than the corresponding identities be-tween the coding content of the complete genomes of these parasites (Table 3). These results indicate that the genus Leishmania unique LGTs are less conserved than the universal trypanosomatid LGTs, probably because they have resided a shorter time in their recipient genomes. This, in turn, support the conclusion that the genus Leishmania unique LGTs have been acquired after the divergence of genus Leishmania, rather than the alternative hypoth-esis which would require that they were acquired before the speciation of the trypanosomatids and then subsequently lost in the trypanosomes while being retained in genus Leishmania.

LGT in genus Leishmania is a dynamic process Eleven of the 20 LTGs unique to genus Leishmania, and two of the three universal trypanosomatid LGTs, had orthologs in all fifteen strains investi-gated in this study – ten within previously unpublished genomes and five with published genomes (Table 1). Six of the nine remaining unique LTGs in genus Leishmania had orthologs in all three subgenera, while lost in one or several strains, and final three only had orthologs in subgenus Leishma-nia. However, a careful analysis revealed remnants to one of these three LTGs in both L. tarentolae and in L. mexicana M., and another one in L. mexicana M only. The remainder of these three solely had orthologs in strains belonging to the donovani-, major- and tropica complex. These re-sults aid in correlating the uptake of genus Leishmania unique LGTs to the genus' evolutionary history. Seventeen of the LGTs appear to have been acquired before the divergence of genus Leishmania into different subgene-ra, but after the divergence of genus Leishmania from the other trypanoso-matids. Two have been acquired after the divergence of subgenera Leishma-nia and Saurolesihmania from subgenus Viannia, and one even later – after the divergence of the mexicana complex from the other complexes in subge-nus Leishmania. Hence, this analysis infers that LGT is an on-going process

28

and that it appears to contribute to species-specific diversification in genus Leishmania.

The conclusion that LGT is a dynamic process in genus Leishmania is strengthen by the Ka/Ks ratios calculated between the LGT orthologs, which showed that there is a varying degree of conservation among these genes. The Ka/Ks ratios of the reference genes EF-TU and HSP70 were close to zero, confirming that they are under purifying selection. This also applied to sixteen of the genus Leishmania specific LGT orthologs and all three uni-versal trypanosomatid LGT orthologs. The Ka/Ks ratios of the remaining LGT orthologs were greater than or close to one, indicating that these four LGT orthologs are rapidly evolving, but under low selective pressure. An additional observation supporting this conclusion of low evolutionary pres-sure is that these four LGTs also have orthologs in one or several strains that are interrupted by premature stop codons.

All of the LGT orthologs had average CAI scores that were in the same range as genes in tandem, which indicates that they are highly expressed [25, 80, 84]. However, three of the genus Leishmania unique LGT orthologs displayed extensive codon bias variation (high coefficient of variance). The-se LGT had subsets of orthologs that displayed significantly lower CAI score than the remaining LGT orthologs and were among the LGTs with orthologs that are under low selective pressure showing signs of gene inno-vation or degradation. Thus, the predicted expression level and functionality of these LGT orthologs correlates well with the dynamic nature of LGT in genus Leishmania. A phylogenetic tree displaying the fate of each of the twenty-three LGTs, investigated in detail in this study is presented in Figure 5. The four LGTs with Ka/Ks ratios ≥1 and the orthologs to the three LGTs with large variations in CAI score are highlighted in the phylogenetic tree.

29

Figure 5. Consensus tree summarizing the agreement between the phylogenetic trees for the twenty genus Leishmania unique LGTs and the three universal trypano-somatid LGTs. The strains used in the in-depth analysis of genus Leishmania specif-ic LGTs belong to subgenera Viannia, Sauroleishmania (here represented by only one species and one strain – L. tarentolae) and Leishmania. The fates of the LGT orthologs are illustrated with coloured boxes. The LGT orthologs with Ka/Ks rati-os ≥ 1 and/or with large variations in CAI score are highlighted in the phylogenetic tree with squares and circles. (Modified reprint from paper I)

30

Conclusions These conclusions can be made from paper I:

• LGTs have had an impact on the speciation and species diversifi-

cation within genus Leishmania. • LGT in genus Leishmania is a continuous dynamic process where

acquired genes are either being conserved, transformed or degrad-ed and lost.

• LGTs can be attractive as drug targets, however it is important to evaluate their stability in their recipient genome. Universal trypa-nosomatid LGTs are most often more stable and conserved than LTGs unique to genus Leishmania.

31

Identification and evaluation of compounds targeting LGT products and other enzymes in Leishmania folate metabolism (paper II) The folate pathway is essential for Leishmania parasites, just as it is for most organisms – from bacteria to plants and animals. The bifunctional enzyme dihydrofolate reductase-thymidylate synthase (DHFR-TS) is one of the key-stones in Leishmania folate metabolism. The DHFR domain reduces dihy-dro- and tetrahydrofolates and the TS domain methylates deoxyuridine mo-nophaspahte (dUMP) to thymidine monophosphate (TMP) (Figure 6) [85-87]

Figure 6. Schematic picture of reactions catalysed by DHFR-TS and PTR1. (Reprint from paper II)

32

DHFR is the target for many anti-folates used in cancer therapy and for tri-methoprim used for bacterial infections [88-91]. However anti-folate treat-ment against leishmaniasis has been shown to be ineffective, probably be-cause the enzyme pteridine reductase 1 (PTR1) function as a metabolic by-pass in folate metabolism [92, 93]. Reducing folates when DHFR is inhibit-ed is, however, not the main function of PTR1. The primary function of PTR1 is to reduce biopterins, which are important growth factors in the par-asite (Figure 6) [94]. PTR1 is a potential drug target because it is a choke-point enzyme (no other Leishmania enzymes can reduce biopterins) [24] and it is encoded by a universal trypanosomatid LGT [26, 27, 95]. There are selective PTR1 inhibitors, which have been shown to have in-vitro effect [96, 97]. However, targeting both DHFR-TS and PTR1 is the most likely successful route to find an anti-folate based treatment against leishmaniasis [93, 94, 98]. This hypothesis has been confirmed by combining PTR1 inhibi-tors with DHFR inhibitors in various in-vitro assays [99-102]. The TS do-main, on the other hand has received limited attention, despite the fact that parasites with a mutated TS domain show less viability and virulence than wild type parasites [103, 104].

In paper II potential hits against PTR1, and either of the two domains of DHFR-TS were found using similarity-based screening followed by molecu-lar docking. The anti-leishmanial activities of the potential hits were further evaluated with a L. tropica promastigote assay.

Methods Data mining Protein structures of PTR1 and DHFR-TS were retrieved from RCSB Pro-tein Data Bank (PDB) [105]. There were nine structures available for L. major PTR1 and eight different ligands were co-crystallised with these. There were no available protein structures of DHFR-TS from Leishmania, but there were eight available DHFR-TS structures from Trypanosoma cruzi – with two ligands co-crystallised with the TS domain, and six with the DHFR domain. A homology model of L. major DHFR-TS was built with SWISS-MODELL using a protein structure of TcDHFR-TS as template. A comparison of the amino acid sequences of the 3D modelled protein struc-ture of LmDHFR-TS and a 3D protein structure of TcDHFR-TS, revealed that the binding pockets of the two structures are very similar, and therefor the TcDHFR-TS structures were used in the study (Table 5).

33

Table 5. Ligands co-crystallised with proteins structures of PTR1 and DHFR-TS.

Protein Co-crystallised ligands PTR1

MTX (methotrexat ) HBI (7,8 dihydrobiopterin) TAQ (2,4,6-triaminoquinazoline) CB3 (10-propargyl-5,8-dideazafolic acid) TOP (trimethoprim) H4B (5,6,7,8-tetrahydrobiopterin) FE1 (methyl 1-(4-(((2,4-diaminopteridin-6 yl) methyl) (methyl) amino) benzoyl) piperidine-4-carboxylate)

DHFR domain 1CY (cycloguanil) 2CY (5-[3-(3-fluorophenoxy) propoxy]quinazoline-2,4-diamine) TMQ (trimetrexate) MTX (methotrexat) TS domain DU (2’-deoxyuridine-5’-monophosphate) UMP (2’-deoxyuridine-5’-monophosphate)

Selecting potential hits from and in-house dataset EDs between the co-crystallised ligands of LmPTR1 and TcDHFR-TS, and compounds in an in-house dataset (n = 91901) – comprising both synthetic and compounds of natural origin – in ChemGPS-NP chemical-property space were calculated. Compounds from the in-house dataset and ligand co-crystallised to PTR1- and DHFR-TS were considered as similar if ED ≤ 2. This cut-off ED is higher than previous suggested cut-off ED for similarity (ED ≤ 1) [47], because this similarity-based screening step was used as a filter to select compounds for further screening using molecular docking.

Compounds with similar properties as the PTR1- and DHFR-TS ligands were docked to the catalytic site of PTR1 and both domains of DHFR-TS using Glide with SP mode. The compounds with the top 50% best docking scores in each catalytic side were re-docked using Glide XP. The distribu-tion of XP docking scores among these compounds were represented in his-tograms with bin width = 1. Compounds with higher XP-docking scores than the most frequent occurring XP docking score in each data set were re-moved.

Commercially available compounds, which were predicted to inhibit both PTR1 and the TS domain, or PTR1 and the DHFR domain, were selected as potential hit compounds.

Anti-leishmanial activity of selected compounds Selected compounds were subsequently evaluated for their inhibitory effects against the viability of L. tropica promastigotes in a resazurin based assay. Resazurin is converted by metabolic active cells to resorufin – a colorimetric dye that is excited at 550 nm and emits fluorescence at 590 nm. In brief, promastigotes were seeded in a 96 well plate and the hit compounds (dis-

34

solved in either water or DMSO) were added as triplicates, alone or in com-binations, in different concentrations (160, 80, 40, 20, 10, 5 µM). Negative controls were wells with promastigotes to which only solvent was added. Wells containing only the growth media of the promastigotes and solvent were used for background corrections. Resazurin dissolved in PBS was add-ed to all wells and the plate was incubated at 24, 48 and 72 hours. The effect of resazurin on L. tropica has been evaluated previously (data not shown) and it is not toxic at the concentrations used in this study. Therefore, it could be added before incubation.

Results and Discussion Four compounds were selected against PTR1/TS and one against PTR1/DHFR The number of compounds that have similar physicochemical properties to any of the ligands bound to PTR1, and/or either of the DHFR-TS domains was 2829. After SP docking and XP docking the number decreased to 1535 and 1114, respectively. There were overlaps between compounds predicted to bind to PTR1 and the TS domain, and between compounds predicted to bind to PTR1 and the DHFR domain. However, no compounds were pre-dicted to bind to both the DHFR domain and the TS domain.

Four commercially available compounds predicted to bind to PTR1 and the TS domain, were selected as potential hit compounds. Cytarabine and decitabine are approved drugs and they were selected since the can function as potential drug repurposing candidates. 2´-deoxycytidine is an endogenous substance produced by Leishmania parasites and 5-metylcytidine is a muta-genic post-transcriptional modified nucleoside. Aminopterin, which is a methotrexate analogue, was predicted to bind to PTR1 and the DHFR do-main. Aminopterin was also selected for evaluation of anti-leishmanial ac-tivity (Figure7).

35

Figure 7. Selected potential hit compounds after similarity-based VS and molecular docking.

Inhibition of TS enhances the effect of PTR1/DHFR inhibitors All tested concentrations of aminopterin affected the viability of L. tropica promastigotes, and the inhibitory effect was seen after 24 hours and in-creased after 48 and 72 hours. These results were not unexpected because aminopterin, is similar to methotrexate – the difference is only one methyl group – and methotrexate is a crystallised ligand to one of the PTR1 protein structures. Methotrexate has also been shown to inhibit the DHFR domain of DHFR-TS from Leishmania braziliensis [106].

None of the predicted hits against PTR1 and the TS domain (cytarabine, decitabine, 2´deoxycytidine and 5-metylcytidine) had any effect against the

36

viability of L. tropica promastigotes. However, decitabine enhanced the effect of aminopterin at a significance level of 0.05. Previous studies have reported that decitabine inhibits human thymidylate synthase [107], it is therefore plausible that decitabine enhances the anti-leishmanial effect of aminopterin by inhibiting the TS domain of Leishmania DHFR-TS (Figure 11).

Figure 8. Effect of aminopterin alone or in combination with decitabine on Leish-mania parasites. (Reprint from paper II)

37

Conclusions These conclusions can be made from paper II:

• Simultaneous inhibition of DHFR-TS and the LGT encoded en-

zyme PTR1, both involved in Leishmania folate pathway, is a promising strategy against L. tropica promastigotes.

• Similarity-based screening using ChemGPS-NP in combination with molecular docking is a promising approach for finding hits against potential drug targets in Leishmania parasites.

38

Part II. Suggesting potential targets for anti-leishmanial hit compounds

Development of a ligand-based target fishing approach (Paper III) Phenotypic screening of compounds on cultured parasites is an essential part of drug discovery against leishmaniasis [49, 52, 54, 62]. One disadvantage with phenotypic screening is that it rarely provides any information of the molecular targets of active compounds. This makes it difficult to use rational drug design to optimize the interaction between the compounds and their targets.

LBTF is an in silico-based approach used to predict targets for active compounds based on similarities between the compounds and reference compounds that have known targets. Similarity can, for example, be as-sessed by comparing pharmacophore features or molecular fingerprints based on the chemical structures of the compounds. The reference com-pounds are most often experimentally validated exogenous modulators of various drug targets [57, 108-115]. However, the number of Leishmania targets with such ligands is relatively small, and such LBTF approaches can therefore not be used to predict targets for anti-leishmania compounds dis-covered through phenotypic screening.

In paper III an alternative LBTF approach was developed, in which known endogenous ligands (metabolites) to enzymes involved in Leishma-nia metabolism were used as reference compounds and similarities were assessed with ED in chemical property space defined by ChemGPS-NP [45-48]. This LBTF approach was further used to suggest targets for leishmani-cidals with currently unknown mechanisms of action.

Methods

The ligand-based target fishing workflow The LBTF approach is outlined in (Figure 9) and was as follows: the posi-tions of metabolites (reference compounds) and a query compound in physi-cochemical space were calculated with ChemGPS-NP. EDs between the query compound and all metabolites were calculated and the enzymes asso-

39

ciated with the metabolites were ordered by increasing ED between the que-ry compound and the metabolites. Enzymes associated with several metabo-lites were ranked in the list according to the metabolite with the shortest ED to the query compound. It was hypothesized that the enzymes binding the metabolites positioned close to the query compound (short ED) were possi-ble targets of the query compound. There is however no definition of “short ED” based on an established cut-off ED, for this type of approach. There-fore, the LBTF method was further developed and evaluated using datasets with known inhibitors and metabolites to human and Leishmania metabolic enzymes. Human metabolites and inhibitors were used because the number of inhibitors against Leishmania metabolic enzyme is relatively low.

Figure 9. Workflow of the ligand-based target fishing approach. (Reprint from pa-per III)

Method development Four datasets were compiled to set a proper cut-off ED for similarity in the LBTF approach and to elucidate if metabolites were suitable as reference molecules:

1. A human metabolite set (HMS) containing all substrates and products (except ions, cofactors and nucleotides) annotated for each enzyme in all human metabolic pathways (except Cyto-chrome P-450 enzymes).

2. A human enzyme inhibitor set (HIS) with compounds active against human metabolic enzymes (threshold value IC50 ≤ 10 μM).

3. A Leishmania metabolite set (LMS) containing all substrates and products (except ions, cofactors and nucleotides) annotated for each enzyme in Leishmania metabolic pathways.

4. A Leishmania enzyme inhibitor set (LIS) with compounds active against Leishmania metabolic enzymes (threshold value IC50 ≤ 10 μM).

The metabolites in LMS (n = 552) and HMS (n = 916) were retrieved from the KEGG database [116-118] and the inhibitors in LIS (n = 52) and HIS (n = 13 995) from the ChEMBL database [58, 119, 120]. The development process was split into five analyses, each over a range of cut-off EDs = 0.2 –2.4 (Figure 10). This range was set, because it covers cut-off ED = 1 sug-gested by other studies of compound similarity employing ChemGPS-NP and ED calculations [47].

40

Figure 10. Workflow of the method development process. (Reprint from paper III)

Target fishing for anti-leishmanial compounds The LBTF method was finally used to predict targets for compounds with documented anti-leishmanial activity (IC50 ≤ 10 μM), retrieved from the ChEMBL database, but with unknown mechanisms of actions. These com-pounds were compiled in a leishmanicidal data set (LDS) (n = 1875). The cut-off ED was based on the combined results from the five analyses used to develop the method and the reference compounds were metabolites from all Leishmania species available in KEGG database.

41

Results and Discussion For how many inhibitors are the true target or any target predicted? The results from the first analysis showed that the number of compounds in HIS and LIS for which targets were predicted increased as the cut-off ED increased. At cut-off ED = 2.4, true targets were predicted for 14.7% and 34.6% of the compounds in HIS and LIS respectively. The amounts of com-pounds in HIS and LIS for which any target was predicted at ED = 2.4 were 76.6% and 78.8% respectively (Figure 11).

Figure 11. Percentages of compounds in HIS and LIS for which targets were pre-dicted at different cut-off EDs.

How many targets are predicted for each inhibitor and how many of these are true targets? The results from the second analysis showed that for most inhibitors that obtained at least one predicted target (HIS-PT and LIS-PT), the numbers of predicted targets were relatively low. However, the number of inhibitors for which only 5–10 targets were predicted decreased when cut-off ED in-creased (Figure 12).

Thus, the number of compounds in HIS-PT and LIS-PT increased if the cut-off ED increased, but the number of targets predicted for each compound in these two data sets also increased. However, large cut-off ED values can result in many false positives among the predicted targets, and long EDs between compounds in chemical property space imply that they are less similar.

42

Figure 12. Percentages of compounds for which a certain number of targets are predicted at different cut-off EDs.

What is the precision and recall of the method? The balance between predicting enough targets without obtaining to many false positives was evaluated in the third analysis. Precision is the proportion of true targets among the predicted targets and recall is the proportion of true targets that the method predicts to be true. The highest average precision and average recall of the LBTF approach to predict targets for; HIS compounds, were 0.06 (ED = 0.6) and 0.26 (ED = 0.8) respectively; LIS compounds, were 0.06 (ED = 1.8) and 0.44 (ED = 2.2) respectively. These results were compared to the performance of target fishing based on random selection of 1–50 targets associated to metabolites in HMS and LMS. The average precision of randomly assigning targets for HIS com-pounds ranged from 0.0014 to 0.0019 and average recall from 0.0012 to 0.066. For LIS compounds the average precision ranged from 0–0.0064 and average recall from 0–0.19 depending on sample size (Figure 13). Hence, the LBTF approach, was shown to perform better than random assignment of targets in this dataset.

Figure 13. The average recall and average precision of the LBTF method compared to the maximum recall and maximum precision of random target assignment to the compounds in HIS and LIS.

43

What rank does the true target of each inhibitor get? The result obtained after using the LBTF method to suggest targets to a que-ry compound is list of potential targets. The fourth analysis was thus per-formed to investigate if the true targets for inhibitors in HIS-PT and LIS-PT were ranked high or low on that list. The ranks of the true targets for inhibi-tors in the HIS-PT and LIS-PT are plotted against the ED between the corre-sponding metabolite and the inhibitor in Figure 14. If an inhibitor had sever-al true targets, only the highest ranked target was considered in this analysis. The true target of HIS-PT compounds were all ranked top 10 if a metabolite of the true target were not longer than ED = 1.2 from the HIS-PT compound. The true targets of the LIS-PT compounds were consistently in top 14 if a metabolite of the true target were farther away than ED = 2.2 from the corre-sponding LIS-PT compound. So, the ranks of the true targets for compounds in both HIS-PT and LIS-PT were often high, even if many targets were pre-dicted.

Figure 14. Rank of the true target for compounds in HIS-PT and LIS-PT against the ED between the compound and the metabolite of the true target.

Are the predicted targets related to each other? In the fifth and final analysis the enzymes in the HMS and LMS were pre-sented in network graphs, in which the enzymes were represented as vertices and connected with edges if they bound the same metabolite. Vertices corre-sponding to predicted targets for each compound in HIS-PT and LIS-PT at different cut-off EDs were extracted from the graphs. Clusters of vertices – as singletons or connected with edges – were identified with the label propa-gation algorithm [121]. This is exemplified in Figure 15 in which the verti-ces representing the 22 predicted targets for CHEMBL280463 at cut-off ED = 1 are extracted from the HMS graph.

44

Figure 15. Network graph connecting human metabolic enzymes that bind the same metabolite. The extracted vertices, highlighted in the five boxes, corresponds to predicted targets for CHEMBL280463 at cut-off ED = 1. They form five clusters when they are extracted from the graph. (Reprint from paper III)

The predicted targets for more than 90% of the compounds in HIS-PT formed five or fewer clusters up to a cut-off ED = 1.6 and the predicted tar-gets for less than 25% of the compounds in LIS-PT formed more than 5 clusters up to a cut-off ED = 1.6 (Figure 16).

45

Figure 16. Percentages of predicted targets for inhibitors in HIS-PT and LIS-PT different cut-off EDs that form a certain number of clusters.

Hence, many predicted targets most often bind the same metabolite, which can decrease the number of predicted targets that are selected for the first experimental evaluation – only one predicted target in a cluster of related enzymes needs to be included in the initial round of testing.

Based on the results from the five analyses described above ED = 0.8 – 1.6 was decided as a suitable cut-off ED range for the LBTF method. This range include the cut-off ED suggested by Bounfigouli et al. [47].

Targets against anti-leishmanial hit compounds were predicted The LBTF approach – with a cut-off ED = 1.6 and the metabolites from all Leishmania species as reference compounds – was used to predict targets for 1875 compounds (referred to as leishmanicidals) in LDS. The LBTF sug-gested targets for 158 leishmanicidals, and for half of these leishmanicidals only one target was predicted. For those leishmanicidals for which several targets were predicted, the targets were most often present in the same path-ways – primarily in metabolism of cofactors, vitamins, and lipids.

Several of the predicted targets are encoded by LGTs. For example coproporphyrinogen oxidase, protoporphyrinogen oxidase, and ferrochela-tase) [26, 27], involved in heme biosynthetic pathways. Deoxyuridine tri-phosphate nucleotidhydrolase and inosine-adenosine-guanosine-preferring nucleoside hydrolase are other examples of LGT encoded enzymes that were predicted as targets for leishmanicidals. Both of these have been suggested as drug targets in previous studies [122-124].

Hence the LBTF approach can be used to identify a limited set of poten-tial and relevant targets for compounds that already have confirmed activity on parasites. However, these predictions should be experimentally evaluated since they are by no means confirmed targets for the respective leishmani-cidal.

46

Conclusions These conclusions can be made from paper III:

• An LBTF approach based on similarities between a query com-pound and endogenous ligands to metabolic enzymes, quantified by ED in the chemical space defined by ChemGPS-NP can be used to predict the true targets for compounds with known activity against metabolic enzymes in both Leishmania and human.

• A cut-off ED range of 0.8 – 1.6 is suitable for this LBTF ap-proach.

• The LBTF approach can be used to predict relevant targets for compounds that have confirmed activity on Leishmania parasites but with unknown mechanisms of action.

47

Evaluation of marine alkaloids against Leishmania parasites and predicting possible targets for these (paper IV) Marine life forms, including sponges, algae, soft corals and marine microor-ganisms are important sources of bioactive natural products [125]. Indeed, marine sponges have been called ”a drug treasure house” [126]. In paper IV the activity of four marine sponge alkaloids; aaptamine and isoaapta-mine, isolotaed from Haliclona (Gellius) cymaeformis; oroidin and nagelamide D, isolated from Agelas nemoechinata (Figure 17), were tested on Gram-negative and -positive bacterium, Candida, Leishmania and six human cancer cell lines. Furthermore, ChemGPS-NP was used to compare the physicochemical properties of the four marine sponge alkaloids with compounds (both synthetic and of natural origin) that have known activity against the above listed microbes and the cell lines. Membrane leakage assay and further exploration of the nearest neighbours in chemical property space were used to suggest possible mechanisms of action of the alkaloids.

Figure 17. A) Haliclona (Gellius) cymaeformis. B, C) Agelas nemoechinata Chem-ical structures of the sponge alkaloids. (Photo by Mr. Wei-Chieh) (Reprint from paper IV)

48

Methods Evaluation of the biological activity of the sponge alkoloids The activities of the four sponge alkaloids against four human pathogenic microbes and six cell lines were evaluated with different viability assays (Table 6).

Table 6. Microbial pathogens and cell lines with corresponding assay.

Human pathogen/Cell line Assay Escherichia coli ATCC 25922 Staphylococcus aureus ATCC 29213 Candida albicans ATCC 90028 Leishmania tropica K562 (bone marrow lymphoblast) CCRF-CEM (blood T lymphoblast) Molt 4 (blood T lymphoblast) DLD-1 (colon epithelial) LNCap (prostate epithelial) MCF7 (mammary gland epithelial)

MICa MICa MICa Resazurin MTTb MTTb MTTb MTTb MTTb MTTb

aMIC; Minimum inhibitory concentration, bMTT; 3-(4, 5-dimethyl-2-thiazolyl)-2, 5-diphenyl-2H-tetrazolium bromide

Membrane permeabilization The membrane permeabilization potential of the sponge alkaloids were stud-ied in a bacterial model membrane system consisting of large liposomes containing carboxyfluorescein [127]. Carboxyfluorescein is a fluorescent dye, which is self-quenching when it is encapsulated. Membrane disruption cause efflux of carboxyfluorescein and the de-quenching of leaked carbox-yfluorescein results in a fluorescence signal with excitation and emission wavelengths of 492 and 517 nm.

Exploration of the sponge alkaloids in chemical property space The physicochemical properties of antibacterial, cytotoxic and anti-leishmanial compounds, as well as the properties of the sponge alkaloids were calculated with ChemGPS-NP. The antibacterial, cytotoxic and anti-leishmanial compounds were retrieved from the ChEMBL database [58, 119, 120] with threshold values of MIC ≤ 32 μg/ml, IC50 ≤ 32 μg/ml, IC50 ≤ 32 μg/ml respectively.

Similarities were assessed visually by plotting the compounds in 3D plots representing the first three dimensions of chemical-property space defined by ChemGPS-NP [45, 46].

Prediction of mechanism of action in leishmania EDs in chemical property space, defined by ChemGPS-NP, between the sponge alkaloids and the antibacterial, cytotoxic and anti-leishmanial com-pounds were calculated. Compounds positioned at ED ≤ 1 (previously sug-

49

gested ED cut off for similarity [47]) were identified as nearest neighbours. Molecular targets of the near neighbours were searched for in ChEMBL database [58, 119, 120] and these were – based on the “similarity principle” [38] – suggested as potential targets for the sponge alkaloids.

Results and Discussion The sponge alkaloids have distinct activities against microbes and cell lines The sponge alkaloids had different activities in the microbes and cell lines used in this study (Table 7). The compounds classified as inactive displayed no activity at the highest concentration tested, which where; 40 µM in E. coli, S. aureus and C. albicans; 160 µM in L. tropica; and 20 µg/ml in the cell lines.

Table 7. A compilation of the biological activities of the sponge alkaloids.

Aaptamine Isoaaptamine Oroidin Nagelamide D E. coli S. aureus C. albicans L. tropica K562 CCRF-CEM Molt 4 DLD-1 LNCap MCF7

Inactive Active Inactive Active Inactive Active Active Inactive Inactive Inactive

Inactive Inactive Inactive Active Active Active Active Actve Active Active

Active Active Inactive Inactive Inactive Inactive Inactive Inactive Inactive Inactive

Active Active Inactive Inactive Inactive Inactive Inactive Inactive Inactive Inactive

Some of the above reported biological activities are in congruence with pre-vious studies. For example, the cytotoxic activities isoaaptamine against all cell lines and aaptamine on only CCRF-CEM and Molt 4, confirm results from previous studies. These studies also reported that isoaaptamine is more cytotoxic than aaptamine [128-131]. Another example is the inhibitory effect of isoaaptamine on the viability of L. tropica promastigotes (IC50 = 29 µM after 48 hours), which is somewhat in agreement with previous reported inhibitory effects (IC50 = 3 μM) of isoaaptamine on the viability of Leish-mania donovani parasites (developmental stage not reported) [132]. The antibacterial effects of oroidin and nagelamid D have also been reported before [133, 134], although these were moderate or weak compared to the antibacterial effects reported in this study.

There are also biological activities, that has not been reported before, such as the anti-leishmanial activity (although with higher IC50 than isoaap-tamine) of aaptamine and the bactericidal activity of aaptamine on S. aureus.

50

Oroidin and nagelamide D disrupt membrane Neither aaptamine nor isoaaptamine had any significant effect on membrane permeabilization (aaptamine only caused a 3% leakage at a concentration of 80 µM). Oroidin and nagelamide D, on the other hand, had disruptive effects on the liposomes used in the membrane permeabilization assay. In addition, the antibacterial activity of oroidin and nagelamide D, respectively, corre-sponded with their respective permeabilizing activity on bacterial liposomes, i.e. nagelamide D is more potent and has distinct membrane disruptive prop-erties at lower concentration than oroidin.

This suggests that a membrane permeabilizing process causes the antibac-terial effects of nagelamide D, and the less potent oroidin. Whereas, the bio-logical activities of the aaptamines have other so far unrevealed explana-tions.

The sponge alkaloids exhibit unusual properties The compounds with experimentally validated activity against E. coli (n = 9408); S. aureus (n = 17 898); Leishmania parasites (n = 4703); K562 (n = 8547); Molt 4 (n = 3943); CCRF-CEM (n = 7016); DLD-1 (n = 1376); LNCap (n = 1645) and MCF7 (n = 19 724) were positioned together with the sponge alkaloids in 3D plots of the three first dimension in chemical proper-ty space defined by ChemGPS-NP. A visual evaluation of the plots revealed that oroidin and nagelamide D were not positioned in the vicinity of most compounds with anti-E. coli and anti-S. aureus activity. Aaptamine – active against S. aureus, L. tropica, MOLT-4 and CCRF-CM, and isoaaptamine – active against L. tropica and all cell lines tested, are derivatives of each other and have similar properties. They were thus positioned near each other. However, both aaptamines were positioned in the outer border of the region most densely populated by compounds with cytotoxic, anti-leishmanial and anti-S. aureus activity (Figure 18).

Hence, the sponge alkaloids have properties that are not commonly found among compounds active against the cell lines and microbes used in this study. Further exploration of the region in chemical-property space sur-rounding the sponge alkaloids can therefore inspire new discoveries of anti-bacterial, cytotoxic and anti-leishmanial agents.

51

Figure 18. Positions of nagelamide D (green sphere), oroidin (blue sphere), aapta-mine (yellow sphere), isoaaptamine (red sphere) and A) compounds with anti-E. coli activity, B) anti-S. aureus activity, C) anti-leishmanial activity, D) cytotoxic activity (grey spheres) in the first three dimensions of chemical property space described with ChemGPS-NP (PC1 = green box, PC2 = yellow box and PC3 = red box). (Modified reprint from paper IV)

Near neighbours of the sponge alkaloids were used for target prediction Near neighbours were identified to aaptamine, isoaaptamine and oroidin by calculating EDs between the compounds with anti-microbial and cytotoxic activities and the sponge alkaloids. The numbers of near neighbours to the sponge alkaloids and suggested targets are presented in Table 8.

52

Table 8. No. near neighbours to sponge alkaloids with antibacterial, cytotoxic and anti-leishmanial activity and predicted targets.

Alkaloid Target organism No. Near neighbours (with known targets)

Target to near neighbour(s)

oroidin E. coli 6 (1) DNA gyrase1 aaptamine S. aureus 25 (1) DnaC helicase2 oroidin S. aureus 6 (0) n/a aaptamine L. tropica 6 (1) CAAX prenyl p13 isoaaptamine L. tropica 23 (1) CAAX prenyl p13 aaptamine CCRF-CEM 13 (4) Tubulin2 isoaaptamine CCRF-CEM 14 (6) Tubulin2 aaptamine Molt4 10 (4) Tubulin2 isoaaptamine Molt4 13 (6) Tubulin2 isoaaptamine DLD-1 5 (2) NO synthase4 isoaaptamine K562 23 (7) Tubulin2 isoaaptamine MCF7 70 (16) Tubulin2 1 subunit B 2 multiple predicted targets are available. 3 CAAX prenyl protease I 4 multiple predicted targets are available, but they are all isomers.

Targets for near neighbours were suggested as potential targets for the sponge alkaloids in the different microorganisms and in the cell lines, and one of the prediction could be supported by previous studies – DNA gyrase, has been reported as target for oroidin analogues in another study [135].

DNA C helicase was suggested as target for aaptamine in S. aureus, how-ever, it can neither be contradicted nor supported by previous studies, albeit according to Bowling et al aaptamine binds to DNA [136] which is the sub-strate of DNA helicase.

CAAX prenyl protease I was suggested as potential targets for aaptamine and isoaaptamine in L. tropica. CAAX prenyl protease I is a potential valua-ble drug target in Leishmania, because knock out of the gene encoding its isomer, CAAX prenyl protease II, in L. donovani significantly hamper para-site viability and infectivity [137, 138].

The predicted targets for aaptamine and isoaaptamine in CCRF-CEM, Molt4, K562 and MCF7 were all isoforms of tubulin. This suggestion is not completely confirmed by other studies, however it is in congruence with studies showing that the cytotoxic activity of some other marine alkaloids are caused by interactions with tubulin [139].

The predicted targets for the aaptamines were different in the microorgan-isms and the cell lines, because the near-neighbours (with known molecular targets) to the aaptamines among the cytotoxic- anti-leishmanial- and anti-S. aureus compounds were not the same.

53

Targets could neither be predicted to isoaaptamine in the cell line LNCap, nor to oroidin in S. aureus, because isoaaptamine were not near any of the compounds with activity against LNCap and the near neighbours of oroidin with anti-S. aureus activity had no known molecular targets.

Conclusions These conclusions can be made from paper IV:

• The four sponge alkaloids display distinct biological activities. • Oroidin and nagelamide D display antibacterial activity, and iso-

aaptamine is active against L. tropica and six different human cell lines.

• Aaptamine is also active against L. tropica but not with the same potency, and it affects the viability of two of the human cell lines. S. aureus is also sensitive to aaptamine.

• Oroidin and nagelamide D are membrane disruptive, which can explain their antibacterial effects.

• The four alkaloids have physicochemical properties different form most other compounds with known antimicrobial and cytotoxic activity.

• A few similar compounds to aaptamine, isoaaptamine and oroidin could be found and these were used to suggest potential molecular targets for these three alkaloids.

54

Concluding remarks

The work presented in this thesis is a two-front strategy against leishmania-sis. First, LGT encoded enzymes were evaluated as potential drug targets in Leishmania parasites and potential hit compounds to selected LGT products were thereafter suggested using in-silico drug discovery methods. The major finding from paper I and paper II are that:

• LGTs of prokaryotic origin can either be conserved, transformed

or degraded their recipient Leishmania genomes. Hence, it is of great value to evaluate the stability of LGTs in Leishmania ge-nomes before suggesting their products as drug targets.

• PTR1 is a stable LGT product involved in folate metabolism, and simultaneously inhibition of both PTR1 and DHFR-TS (one of the key enzymes in Leishmania folate metabolism) is a promising strategy against L. tropica.

Second, targets for compounds with anti-leishmanial activity but with un-known mechanisms of action were suggested using a LBTF approach. Ma-rine natural products with activity against Leishmania parasites were also identified and potential targets for these were predicted. The major finding from paper III and paper IV are that:

• An LBTF approach, using endogenous ligands to Leishmania metabolic enzymes as reference compounds and ED in chemical property space defined by ChemGPS-NP as similarity metric, can be used to suggest potential targets for anti-leishmanial com-pounds found by phenotypic screening.

• Isoaaptamine and aaptamine, sponge alkaloids isolated from Hali-clona (Gellius) cymaeformis, display anti-leishmanial activity, and their suggested targets is CAAX prenyl protease 1.

55

Populärvetenskaplig sammanfattning

Varje år dör mellan 20 000 och 30 000 människor i leishmaniasis – en sjuk-dom som de flesta människor i västvärlden aldrig har hört talas om. Leish-maniasis drabbar främst fattiga människor i tropiska och subtropiska länder, och den har under en lång tid varken prioriterats av läkemedelsföretag, myn-digheter eller andra hälsoorganisationer. Den har därför klassificerats som en försummad tropisk sjukdom (NTD = Neglected Tropical Disesase) av Världshälsoorganisationen (WHO).

Leishmaniasis orsakas av vissa typer av Leishmania-parasiter som sprids av sandflugor. Det finns tre huvudtyper av sjukdomen – kutan, mukokutan och visceral leishmaniasis. Symtomen av kutan leishmaniasis är sår som kan läka av sig själv, och därför behandlas inte alltid den kutana formen. En del sår kan dock vara mycket svårläkta och kan till och med bli kroniska. Patien-ter med den mukokutana formen får också svårläkta sår, men dessa uppstår i slemhinnorna runt mun och näsa, som kan bli så skadade att patienten får svårt att andas och äta. Ärren efter mukokutan leishmaniasis kan vara myck-et omfattande och kan orsaka vanställda ansikten hos personer som haft sjukdomen. Kutan och mukokutan leishmaniasis är sällan dödliga, men sviterna efter sjukdomarna kan vara stigmatiserande och människor med dessa sjukdomar kan bli förskjutna från grupper och gemenskaper i de sam-hällen som de lever. Den tredje formen, visceral leishmaniasis, kallas också för kala azar (svart/dödlig feber på hindi). Kala azar drabbar de inre organen såsom mjälte och lever, och symtomen på sjukdomen är bland annat trötthet, feber, avmagring och blodbrist. Patienter med kala azar dör om de inte behandlas. De finns ett fåtal läkemedel för att behandla de olika formerna av leishmaniasis, men de ger ofta svåra biverkningar, är dyra och fungerar inte alltid eftersom det finns resistenta Leishmania-parasiter. Det finns därför ett stort behov av forskning om Leishmania-parasiter och utveckling av nya läkemedel mot leishmaniasis. I den här avhandlingen, bestående av fyra delarbeten, har jag använt olika in-silico-metoder (datormodeller), tillsam-mans med laborativt arbete för att försöka bidra till läkemedelsutveckling mot leishmaniasis.

Leishmania-parasiter och människor är relativt lika varandra, i alla fall ge-netiskt, eftersom båda två tillhör organismgruppen eukaryoter. Detta är en av förklaringarna bakom de svåra biverkningar som läkemedel mot leishmania-sis ger. Däremot är bakterier och människor inte speciellt lika varandra och

56

det har länge varit känt att Leishmania-genomet innehåller ett antal gener som har överförts från bakterier som levt i närheten av Leishmania-parasiter. Dessa gener kallas lateralt överförda gener (LGTer =Lateral Transferred Genes). En hypotes är att de proteiner som kodas av LGTerna i Leishmania-parasiterna kan blockeras med läkemedel utan att människan som bär på parasiten påverkas. I det första delarbetet studerades LGTerna i detalj för att utröna vad som händer med generna efter överföringen från bakterier till Leishmania-parasiterna. Resultaten visade att vissa LGTer har inkorporerats (konserverats) i Leishmania-genomet och används troligtvis i deras ur-sprungsform av parasiten; andra LGTer har tagits upp och därefter modifi-erats för att passa Leishmania-parasiten; och ett flertal LGTer håller på att degraderas och försvinna från genomet. Det är därför viktigt att utvärdera de LGTer som kodar för de proteiner som är tilltänkta att blockeras med läke-medel. Endast de proteiner som kodas av konserverade LGTer och som används av Leishmania-parasiter är bra läkemedelsmål.

Ett protein som kodas av en konserverad LGT är pteridine reductase 1 (PTR1) – ett enzym som är involverat i folatmetabolismen som i sin tur är en livsnödvändig process i Leishmania-parasiter. Därför är PTR1 ett potentiellt bra läkemedelsmål och det har studerats hur Leishmania-parasiter påverkas om PTR1 blockeras av olika molekyler. Resultaten av dessa studier visade dock att blockering av PTR1 inte räcker för att hämma tillväxt av och livskraften hos Leishmania-parasiter tillräckligt mycket. Därför har det föreslagits att läkemedel riktade mot parasitens folatmetabolism bör block-era flera enzymer, exempelvis enzymet dihydrofolatreduktas-tymidylatsyntetas (DHFR-TS). DHFR-TS är ett så kallat bifunktionellt en-zym eftersom DHFR och TS sitter ihop. I de flesta andra organismer är DHFR och TS separata enzymer. I det andra delarbetet användes två olika in silico-metoder för att hitta molekyler som sannolikt blockerar PTR1 och antingen TS-delen eller DHFR-delen i DHFR-TS. Fem molekyler hittades – en (aminopterin) som predicerades blockera PTR1 och DHFR, och fyra (cy-tarabine, decitabine, 2´deoxycytidine och 5-metylcytidine) som troligtvis blockerar PTR1 och TS. Dessa molekyler testades på Leishmania-parasiter för att utvärdera om de har anti-leishmaniell effekt. Aminopterin hade anti-leishmaniell effekt, men varken cytarabine, decitabine, 2´deoxycytidine eller 5-metylcytidine hämmade tillväxten av eller dödade Leishmania-parasiterna. Däremot förstärktes den anti-leishmaniella effekten av aminopterin om den kombinerades med decitabine. Dessa resultat bekräftade hypotesen att livskraften hos Leishmania-parasiter minskar mer om flera enzymer i folat-metabolismen blockeras, än om bara ett enzym hämmas. Resultaten från delarbete II visade också att det går att hitta anti-leishmaniella molekyler genom att kombinera olika in silico-metoder.

Tillvägagångssättet i delarbete I och II kallas i läkemedelsutveck-linssammanhang för målbaserad läkemedelsutveckling (eng. target-based drug discovery). Det första steget i målbaserad läkemedelsutveckling är att

57

utse ett protein eller någon annan typ av biomolekyl som potentiellt läke-medelsmål. Detta steg baseras på tidigare forskning eller andra hypoteser som har föreslagit att modifiering av ett visst protein/biomolekyl kommer att lindra eller bota en viss sjukdom. Det andra steget är att hitta molekyler som kan binda till och påverka funktionen av det valda läkemedelsmålet.

En annan typ av läkemedelsutveckling är systembaserad läkemedelsut-veckling (eng. system-based drug discovery). Där utvärderas olika molekylers påverkan på hela organismer och om flera molekyler testas sam-tidigt på samma organism heter det fentotypisk screening. Flertalet läke-medel som används idag har hittats genom fenotypisk screening. Metoden har dock en nackdel, vilken är att resultaten från fenotypisk screening endast visar om en molekyl påverkar en organism eller inte, men de visar inte vilka läkemedelsmål som molekylerna binder till – den informationen måste hittas på annat sätt. I det tredje delarbetet utvecklades en in silico-metod som kan användas för att föreslå läkemedelsmål för aktiva molekyler som hittats med fenotypisk screening eller systembaserad läkemedelsutveckling. Metoden kallas ”ligandbaserat läkemedelsmålfiske” (eng. ligand-based target-fishing, LBTF) och bygger på den så kallade ”likhetsprincipen” som säger att molekyler som liknar varandra har samma biologiska aktivitet och med stor sannolikhet binder till samma läkemedelsmål. LBTF metoden användes därefter för att föreslå läkemedelsmål till anti-leishmaniella molekyler som hittats av andra forskare i tidigare studier.

I det fjärde delarbetet studerades fyra olika alkaloider: aaptamine, iso-aaptamine, nagelamide D och oroidin, som finns i flera olika svampdjur-sarter. Alkaloider är en heterogen grupp av kemiska föreningar som produc-eras av många växter – koffein är ett exempel på en alkaloid som bland an-nat produceras av kaffeplantan. I studien undersöktes om de fyra alkaloider-na var effektiva mot Leishmania-parasiter (även effekten mot Escherichia coli, Staphylococcus aureus, Candida albicans och sex cancercell-linjer utvärderades). Både isoaaptamine och aaptamine visade sig vara anti-leishmaniella, men oroidin och nagelamide D hade ingen effekt på tillväxten av Leishmania-parasiter. Därefter föreslogs läkemedelsmål för aaptamine och isoaaptamine i Leishmania-parasiter genom att använda en metod som bygger på samma principer som den LBTF-metod som utvecklades i delar-bete III.

Sammanfattningsvis har jag försökt angripa leishmaniasis från två olika håll. Först har jag utforskat nya läkemedelsmål och därefter letat efter lämpliga substanser att attackera dessa läkemedelsmål med. Därefter låg fokus på att hitta substanser med effekt mot leishmania-parasiter, och sen förklara hur de fungerar. Det vill säga vilka läkemedelsmål som är tänkbara för dessa anti-leishmaniella substanser. Den här avhandlingen kan således beskrivas som en tvåfrontsstrategi mot leishmaniasis.

58

Acknowledgment

During my PhD-studies, I have had the privilege to get to know and work side by side with great people. So, I want to thank everybody, that in one way or the other, have made these years a memorable time of my life.

Special thanks to my past and present supervisors: Anders Backlund, I want to thank you for giving me free hands with my computer. Many of the things, that I have done during these years, would have been hard to do without admin rights. Cecilia Alsmark and Christian Sköld, without your help, I would not have completed my PhD studies.

Curt Petterson, Anders Karlén and Birgitta Hellsing, thank you for your help and support after my sick-leave.

PhD studies are 10% success and 90% failure1. I can, however, hardly re-member any days at work when I have not laughed or at least smiled, and for this I want to especially thank: Astrid Henz Ryen – for quotes like “If it was easy, every idiot could do it”, and “I think I have ran out of …”; Karin Steffen – for replying in Spanish (when ever needed), Bailando!; Camilla Eriksson – for the secret bracelet; Erik Jacobsson – for “I see you”; Kris-tian Pirttilä – for teaching me some of the do’s and don’ts in programming (I am careful with spiders and I will turn to lists); Annelie Hansson and Alfred Haglind – for inviting me to the “book and wine club” and for being superb travel companions in Paris, Volle Vooo!. 1 common knowledge, no reference needed

I am happy that I was still around for the re-organisation of the department and I want to thank everybody from former AFK and OFK for making ILK such an enjoyable place to work. I would also like to thank all my colleagues (past and present) in the research group of Pharmacognosy. A special grati-tude to Adam Strömstedt – you are a true “catcher in the rye” (sv. “En räddare i nöden”); Hesham El-Seedi – the Yomogin-project might not have been a scientific success, but it made research fun again. Thank you!; Chris-tina Wedén – Thank you for being such a caring and supporting colleague, and for encouraging and motivating me as a teacher.

59

My mentor, Therese Ringbom, thanks for many good advices and nice dis-cussions.

Apotekarsocieteten and Smålands nation for giving me the opportunity to present parts of this thesis at internal conferences.

Jag vill också passa på att tacka vänner och familj!

Emma och Susanna. Jag har dammat av Club Cranium, så vi måste ses snart!

Jenny med familj. Jag hoppas att vi kan ses oftare nu när jag och min familj blivit värmlandsfrälsta.

Alla i aw-gruppen, speciellt Karin, Sandra och Jessica. Våra middagar har varit ett välbehövt andningshål under den senaste tiden.

Maggan. Din värme och generositet känner inga gränser och jag är så tack-sam för att du är min svärmor. Anna, Sara och Frida (med familjer). Bät-tre extrafamilj kunde jag knappast få – ni betyder mycket för mig.

Harald och Therese – min bästa konsertbror och min underbara bonanza-svägerska, Linnea och Magnus – min bästa fjällsyster och min favorit-blåbärsplockarsvåger. Ni ger mig så otroligt mycket glädje. Med er kan jag verkligen skratta tills jag kiknar. Noomie och Amelie, fastersgris saknar er. Lill-donnan, mostersgris kan knappt vänta tills hon får träffa dig.

Mamma och pappa. Jag vet att ni har bett, fortfarande ber och kommer alltid att be för mig. Utan ert bönestöd hade jag inte orkat slutföra detta. Tack!

Fanny och Vendla. Varenda stund tillsammans med er är mer värdefull än den här avhandlingen, och allt som vi gör tillsammans är roligare än att äntligen bli doktor. Ni är mitt lördagsgodis.

Johan. Tack för att du förstår mig. Tack för att du alltid låter mig sova när jag är toktrött.

60

References

1. WHO, Global report for research on infectious diseases of poverty. Geneva. WHO, 2012.

2. Savioli, L. and D. Daumerie, Global plan to combat neglected tropical diseases 2008–2015. World Health Organization, 2007.

3. Hopkins, D.R., et al., Dracunculiasis eradication: the final inch. Am J Trop Med Hyg, 2005. 73(4): p. 669-75.

4. WHO, Control of the leishmaniases. World Health Organization technical report series, 2010(949): p. xii-xiii, 1-186, back cover.

5. David, C.V. and N. Craft, Cutaneous and mucocutaneous leishmaniasis. Dermatol Ther, 2009. 22(6): p. 491-502.

6. Bates, P.A., Transmission of Leishmania metacyclic promastigotes by phlebotomine sand flies. International journal for parasitology, 2007. 37(10): p. 1097-106.

7. Pace, D., Leishmaniasis. J Infect, 2014. 69 Suppl 1: p. S10-8. 8. WHO, Kala-Azar elimination programme. Report of a WHO consultation of

partners, 2015. 9. Olliaro, P.L., et al., Investments in Research and Surveillance Are Needed to

Go Beyond Elimination and Stop Transmission of Leishmania in the Indian Subcontinent. PLoS Negl Trop Dis, 2017. 11(1): p. e0005190.

10. Sunter, J. and K. Gull, Shape, form, function and Leishmania pathogenicity: from textbook descriptions to biological understanding. Open Biol, 2017. 7(9).

11. Espinosa, O.A., et al., An appraisal of the taxonomy and nomenclature of trypanosomatids presently classified as Leishmania and Endotrypanum. Parasitology, 2018. 145(4): p. 430-442.

12. McGwire, B.S. and A.R. Satoskar, Leishmaniasis: clinical syndromes and treatment. QJM, 2014. 107(1): p. 7-14.

13. Copeland, N.K. and N.E. Aronson, Leishmaniasis: treatment updates and clinical practice guidelines review. Curr Opin Infect Dis, 2015. 28(5): p. 426-37.

14. Uliana, S.R.B., C.T. Trinconi, and A.C. Coelho, Chemotherapy of leishmaniasis: present challenges. Parasitology, 2018. 145(4): p. 464-480.

15. Alves, F., et al., Recent Development of Visceral Leishmaniasis Treatments: Successes, Pitfalls, and Perspectives. Clin Microbiol Rev, 2018. 31(4).

16. Chatelain, E. and J.R. Ioset, Drug discovery and development for neglected diseases: the DNDi model. Drug Des Devel Ther, 2011. 5: p. 175-81.

17. Hughes, J.P., et al., Principles of early drug discovery. Br J Pharmacol, 2011. 162(6): p. 1239-49.

18. Eder, J., R. Sedrani, and C. Wiesmann, The discovery of first-in-class drugs: origins and evolution. Nat Rev Drug Discov, 2014. 13(8): p. 577-87.

19. An, W.F. and N. Tolliday, Cell-based assays for high-throughput screening. Mol Biotechnol, 2010. 45(2): p. 180-6.

61

20. Swinney, D.C. and J. Anthony, How were new medicines discovered? Nat Rev Drug Discov, 2011. 10(7): p. 507-19.

21. Smirlis, D. and M.B. Soares, Selection of molecular targets for drug development against trypanosomatids. Subcell Biochem, 2014. 74: p. 43-76.

22. Vasaikar, S., et al., Complementary Approaches to Existing Target Based Drug Discovery for Identifying Novel Drug Targets. Biomedicines, 2016. 4(4).

23. Rahman, S.A. and D. Schomburg, Observing local and global properties of metabolic pathways: 'load points' and 'choke points' in the metabolic networks. Bioinformatics, 2006. 22(14): p. 1767-74.

24. Doyle, M.A., et al., LeishCyc: a biochemical pathways database for Leishmania major. BMC systems biology, 2009. 3: p. 57.

25. Ivens, A.C., et al., The genome of the kinetoplastid parasite, Leishmania major. Science, 2005. 309(5733): p. 436-42.

26. Alsmark, C., et al., Patterns of prokaryotic lateral gene transfers affecting parasitic microbial eukaryotes. Genome biology, 2013. 14(2): p. R19.

27. Berriman, M., et al., The genome of the African trypanosome Trypanosoma brucei. Science, 2005. 309(5733): p. 416-22.

28. Opperdoes, F.R. and P.A. Michels, Horizontal gene transfer in trypanosomatids. Trends in parasitology, 2007. 23(10): p. 470-6.

29. Keeling, P.J. and J.D. Palmer, Horizontal gene transfer in eukaryotic evolution. Nature reviews. Genetics, 2008. 9(8): p. 605-18.

30. Chauhan, N., A.S. Vidyarthi, and R. Poddar, Comparative multivariate analysis of codon and amino acid usage in three Leishmania genomes. Genomics, proteomics & bioinformatics, 2011. 9(6): p. 218-28.

31. Leifso, K., et al., Genomic and proteomic expression analysis of Leishmania promastigote and amastigote life stages: the Leishmania genome is constitutively expressed. Mol Biochem Parasitol, 2007. 152(1): p. 35-46.

32. Worthey, E.A., et al., Leishmania major chromosome 3 contains two long convergent polycistronic gene clusters separated by a tRNA gene. Nucleic acids research, 2003. 31(14): p. 4201-10.

33. Glaab, E., Building a virtual ligand screening pipeline using free software: a survey. Brief Bioinform, 2016. 17(2): p. 352-66.

34. Kitchen, D.B., et al., Docking and scoring in virtual screening for drug discovery: methods and applications. Nat Rev Drug Discov, 2004. 3(11): p. 935-49.

35. Bajorath, J., Integration of virtual and high-throughput screening. Nat Rev Drug Discov, 2002. 1(11): p. 882-94.

36. Lavecchia, A., Machine-learning approaches in drug discovery: methods and applications. Drug Discov Today, 2015. 20(3): p. 318-31.

37. Lo, Y.C., et al., Machine learning in chemoinformatics and drug discovery. Drug Discov Today, 2018. 23(8): p. 1538-1546.

38. Johnson, M. and G.M. Maggiora, Concepts and applications of molecular similarity. 1990, New York: John Wiley & Sons.

39. Maggiora, G., et al., Molecular similarity in medicinal chemistry. J Med Chem, 2014. 57(8): p. 3186-204.

40. Bero, S.A., et al., Similarity Measure for Molecular Structure: A Brief Review. Journal of Physics: Conference Series, 2017. 892(1): p. 012015.

41. Duan, J., et al., Analysis and comparison of 2D fingerprints: insights into database screening performance using eight fingerprint methods. J Mol Graph Model, 2010. 29(2): p. 157-70.

62

42. Mitchell, J.B., Machine learning methods in chemoinformatics. Wiley Interdiscip Rev Comput Mol Sci, 2014. 4(5): p. 468-481.

43. Rogers, D. and M. Hahn, Extended-connectivity fingerprints. J Chem Inf Model, 2010. 50(5): p. 742-54.

44. Rogers, D.J. and T.T. Tanimoto, A Computer Program for Classifying Plants. Science, 1960. 132(3434): p. 1115-8.

45. Larsson, J., et al., ChemGPS-NP: tuned for navigation in biologically relevant chemical space. J Nat Prod, 2007. 70(5): p. 789-94.

46. Rosen, J., et al., ChemGPS-NP(Web): chemical space navigation online. Journal of computer-aided molecular design, 2009. 23(4): p. 253-9.

47. Buonfiglio, R., et al., Investigating Pharmacological Similarity by Charting Chemical Space. J Chem Inf Model, 2015. 55(11): p. 2375-90.

48. Godden, J.W. and J. Bajorath, A distance function for retrieval of active molecules from complex chemical space representations. J Chem Inf Model, 2006. 46(3): p. 1094-7.

49. Zulfiqar, B., T.B. Shelper, and V.M. Avery, Leishmaniasis drug discovery: recent progress and challenges in assay development. Drug Discov Today, 2017. 22(10): p. 1516-1531.

50. Fumarola, L., R. Spinelli, and O. Brandonisio, In vitro assays for evaluation of drug activity against Leishmania spp. Res Microbiol, 2004. 155(4): p. 224-30.

51. De Muylder, G., et al., A screen against Leishmania intracellular amastigotes: comparison to a promastigote screen and identification of a host cell-specific hit. PLoS Negl Trop Dis, 2011. 5(7): p. e1253.

52. Ortiz, D., et al., Discovery of novel, orally bioavailable, antileishmanial compounds using phenotypic screening. PLoS Negl Trop Dis, 2017. 11(12): p. e0006157.

53. De Rycker, M., et al., Comparison of a high-throughput high-content intracellular Leishmania donovani assay with an axenic amastigote assay. Antimicrob Agents Chemother, 2013. 57(7): p. 2913-22.

54. Don, R. and J.R. Ioset, Screening strategies to identify new chemical diversity for drug development to treat kinetoplastid infections. Parasitology, 2014. 141(1): p. 140-6.

55. Reguera, R.M., et al., Target-based vs. phenotypic screenings in Leishmania drug discovery: A marriage of convenience or a dialogue of the deaf? Int J Parasitol Drugs Drug Resist, 2014. 4(3): p. 355-7.

56. Zulfiqar, B., et al., Screening a Natural Product-Based Library against Kinetoplastid Parasites. Molecules, 2017. 22(10).

57. Cereto-Massague, A., et al., Tools for in silico target fishing. Methods, 2015. 71: p. 98-103.

58. Bento, A.P., et al., The ChEMBL bioactivity database: an update. Nucleic Acids Res, 2014. 42(Database issue): p. D1083-90.

59. Richter, L. and G.F. Ecker, Medicinal chemistry in the era of big data. Drug Discov Today Technol, 2015. 14: p. 37-41.

60. Sun, J., et al., ExCAPE-DB: an integrated large scale dataset facilitating Big Data analysis in chemogenomics. J Cheminform, 2017. 9: p. 17.

61. Tetko, I.V., et al., BIGCHEM: Challenges and Opportunities for Big Data Analysis in Chemistry. Mol Inform, 2016. 35(11-12): p. 615-621.

62. Alcantara, L.M., et al., Challenges in drug discovery targeting TriTryp diseases with an emphasis on leishmaniasis. Int J Parasitol Drugs Drug Resist, 2018. 8(3): p. 430-439.

63

63. Charlton, R.L., et al., Repurposing as a strategy for the discovery of new anti-leishmanials: the-state-of-the-art. Parasitology, 2018. 145(2): p. 219-236.

64. Awasthi, B.P. and K. Mitra, In vitro leishmanicidal effects of the anti-fungal drug natamycin are mediated through disruption of calcium homeostasis and mitochondrial dysfunction. Apoptosis, 2018. 23(7-8): p. 420-435.

65. Pinto, E.G. and A.G. Tempone, Activity of the antiarrhythmic drug amiodarone against Leishmania (L.) infantum: an in vitro and in vivo approach. J Venom Anim Toxins Incl Trop Dis, 2018. 24: p. 29.

66. Lima, M.L., et al., Drug repurposing for Leishmania. Molecular basis of the leishmanicidal activity of the antidepressant sertraline. Antimicrob Agents Chemother, 2018.

67. Palit, P. and N. Ali, Oral therapy with sertraline, a selective serotonin reuptake inhibitor, shows activity against Leishmania donovani. J Antimicrob Chemother, 2008. 61(5): p. 1120-4.

68. Lietava, J., Medicinal plants in a Middle Paleolithic grave Shanidar IV? J Ethnopharmacol, 1992. 35(3): p. 263-6.

69. Katz, L. and R.H. Baltz, Natural product discovery: past, present, and future. J Ind Microbiol Biotechnol, 2016. 43(2-3): p. 155-76.

70. Tempone, A.G., C. Martins de Oliveira, and R.G. Berlinck, Current approaches to discover marine antileishmanial natural products. Planta Med, 2011. 77(6): p. 572-85.

71. Thomford, N.E., et al., Natural Products for Drug Discovery in the 21st Century: Innovations for Novel Drug Discovery. Int J Mol Sci, 2018. 19(6).

72. Hong, J., Role of natural product diversity in chemical biology. Curr Opin Chem Biol, 2011. 15(3): p. 350-4.

73. Muigg, P., et al., In silico comparison of marine, terrestrial and synthetic compounds using ChemGPS-NP for navigating chemical space. Phytochemistry Reviews, 2013. 12(3): p. 449-457.

74. Shang, J., et al., Cheminformatic Insight into the Differences between Terrestrial and Marine Originated Natural Products. J Chem Inf Model, 2018. 58(6): p. 1182-1193.

75. Cheuka, P.M., et al., The Role of Natural Products in Drug Discovery and Development against Neglected Tropical Diseases. Molecules, 2016. 22(1).

76. Simoben, C.V., et al., Compounds from African Medicinal Plants with Activities Against Selected Parasitic Diseases: Schistosomiasis, Trypanosomiasis and Leishmaniasis. Natural Products and Bioprospecting, 2018. 8(3): p. 151-169.

77. Hao, W. and G.B. Golding, Patterns of bacterial gene movement. Molecular biology and evolution, 2004. 21(7): p. 1294-307.

78. Marri, P.R., W. Hao, and G.B. Golding, The role of laterally transferred genes in adaptive evolution. BMC evolutionary biology, 2007. 7 Suppl 1: p. S8.

79. Peacock, C.S., et al., Comparative genomic analysis of three Leishmania species that cause diverse human disease. Nature genetics, 2007. 39(7): p. 839-47.

80. Rogers, M.B., et al., Chromosome and gene copy number variation allow major structural change between species and strains of Leishmania. Genome research, 2011.

81. Raymond, F., et al., Genome sequencing of the lizard parasite Leishmania tarentolae reveals loss of genes associated to the intracellular stage of human pathogenic species. Nucleic acids research, 2012. 40(3): p. 1131-47.

64

82. El-Sayed, N.M., et al., The genome sequence of Trypanosoma cruzi, etiologic agent of Chagas disease. Science, 2005. 309(5733): p. 409-15.

83. Fraga, J., et al., Phylogeny of Leishmania species based on the heat-shock protein 70 gene. Infection, genetics and evolution : journal of molecular epidemiology and evolutionary genetics in infectious diseases, 2010. 10(2): p. 238-45.

84. Horn, D., Codon usage suggests that translational selection has a major impact on protein expression in trypanosomatids. BMC genomics, 2008. 9: p. 2.

85. Anderson, K.S., Understanding the molecular mechanism of substrate channeling and domain communication in protozoal bifunctional TS-DHFR. Protein Eng Des Sel, 2017. 30(3): p. 253-261.

86. Knighton, D.R., et al., Structure of and kinetic channelling in bifunctional dihydrofolate reductase-thymidylate synthase. Nat Struct Biol, 1994. 1(3): p. 186-94.

87. Vickers, T.J. and S.M. Beverley, Folate metabolic pathways in Leishmania. Essays Biochem, 2011. 51: p. 63-80.

88. Bertino, J.R., The Mechanism of Action of the Folate Antagonists in Man. Cancer Res, 1963. 23: p. 1286-306.

89. Campbell, J.M., et al., Methotrexate-induced toxicity pharmacogenetics: an umbrella review of systematic reviews and meta-analyses. Cancer Chemother Pharmacol, 2016.

90. Coughter, J.P., J.L. Johnston, and G.L. Archer, Characterization of a staphylococcal trimethoprim resistance gene and its product. Antimicrob Agents Chemother, 1987. 31(7): p. 1027-32.

91. Werkheiser, W.C., The relation of folic acid reductase to aminopterin toxicity. J Pharmacol Exp Ther, 1962. 137: p. 167-72.

92. El Fadili, A., et al., Inactivation of the Leishmania tarentolae pterin transporter (BT1) and reductase (PTR1) genes leads to viable parasites with changes in folate metabolism and hypersensitivity to the antifolate methotrexate. J Biol Chem, 2004. 279(18): p. 18575-82.

93. Nare, B., L.W. Hardy, and S.M. Beverley, The roles of pteridine reductase 1 and dihydrofolate reductase-thymidylate synthase in pteridine metabolism in the protozoan parasite Leishmania major. J Biol Chem, 1997. 272(21): p. 13883-91.

94. Ouellette, M., et al., Pterin transport and metabolism in Leishmania and related trypanosomatid parasites. Int J Parasitol, 2002. 32(4): p. 385-98.

95. Vikeved, E., A. Backlund, and C. Alsmark, The Dynamics of Lateral Gene Transfer in Genus Leishmania - A Route for Adaptation and Species Diversification. PLoS Negl Trop Dis, 2016. 10(1): p. e0004326.

96. Cavazzuti, A., et al., Discovery of potent pteridine reductase inhibitors to guide antiparasite drug development. Proc Natl Acad Sci U S A, 2008. 105(5): p. 1448-53.

97. McLuskey, K., et al., Inhibition of Leishmania major pteridine reductase by 2,4,6-triaminoquinazoline: structure of the NADPH ternary complex. Acta Crystallogr D Biol Crystallogr, 2004. 60(Pt 10): p. 1780-5.

98. Opperdoes, F.R. and G.H. Coombs, Metabolism of Leishmania: proven and predicted. Trends Parasitol, 2007. 23(4): p. 149-58.

99. Corona, P., et al., Structure-based selectivity optimization of piperidine-pteridine derivatives as potent Leishmania pteridine reductase inhibitors. J Med Chem, 2012. 55(19): p. 8318-29.

65

100. Ferrari, S., et al., Virtual screening identification of nonfolate compounds, including a CNS drug, as antiparasitic agents inhibiting pteridine reductase. J Med Chem, 2011. 54(1): p. 211-21.

101. Hardy, L.W., et al., Biochemical and Genetic Tests for Inhibitors of Leishmania Pteridine Pathways. Exp Parasitol, 1997. 87(3): p. 158-70.

102. Tulloch, L.B., et al., Structure-based design of pteridine reductase inhibitors targeting African sleeping sickness and the leishmaniases. J Med Chem, 2010. 53(1): p. 221-9.

103. Cruz, A. and S.M. Beverley, Gene replacement in parasitic protozoa. Nature, 1990. 348(6297): p. 171-3.

104. Titus, R.G., et al., Development of a safe live Leishmania vaccine line by gene replacement. Proc Natl Acad Sci U S A, 1995. 92(22): p. 10267-71.

105. Berman, H.M., et al., The Protein Data Bank. Nucleic Acids Res, 2000. 28(1): p. 235-42.

106. Osorio, E., et al., Biochemical characterization of the bifunctional enzyme dihydrofolate reductase-thymidylate synthase from Leishmania (Viannia) and its evaluation as a drug target. Biomedica, 2013. 33(3): p. 393-401.

107. Almqvist, H., et al., CETSA screening identifies known and novel thymidylate synthase inhibitors and slow intracellular activation of 5-fluorouracil. Nat Commun, 2016. 7: p. 11040.

108. Peragovics, A., et al., Virtual affinity fingerprints for target fishing: a new application of Drug Profile Matching. J Chem Inf Model, 2013. 53(1): p. 103-13.

109. Peon, A., C.C. Dang, and P.J. Ballester, How Reliable Are Ligand-Centric Methods for Target Fishing? Front Chem, 2016. 4: p. 15.

110. Liu, X., et al., In Silico target fishing: addressing a "Big Data" problem by ligand-based similarity rankings with data fusion. J Cheminform, 2014. 6: p. 33.

111. Koutsoukas, A., et al., From in silico target prediction to multi-target drug design: current databases, methods and applications. J Proteomics, 2011. 74(12): p. 2554-74.

112. Keiser, M.J., et al., Relating protein pharmacology by ligand chemistry. Nat Biotechnol, 2007. 25(2): p. 197-206.

113. Huang, T., et al., MOST: most-similar ligand based approach to target prediction. BMC Bioinformatics, 2017. 18(1): p. 165.

114. Gfeller, D., et al., SwissTargetPrediction: a web server for target prediction of bioactive small molecules. Nucleic Acids Res, 2014. 42(Web Server issue): p. W32-8.

115. AbdulHameed, M.D., et al., Exploring polypharmacology using a ROCS-based target fishing approach. J Chem Inf Model, 2012. 52(2): p. 492-505.

116. Kanehisa, M., et al., KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res, 2017. 45(D1): p. D353-D361.

117. Kanehisa, M. and S. Goto, KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res, 2000. 28(1): p. 27-30.

118. Kanehisa, M., et al., KEGG as a reference resource for gene and protein annotation. Nucleic Acids Res, 2016. 44(D1): p. D457-62.

119. Davies, M., et al., MyChEMBL: A Virtual Platform for Distributing Cheminformatics Tools and Open Data. Challenges, 2014. 5(2): p. 334.

120. Jupp, S., et al., The EBI RDF platform: linked open data for the life sciences. Bioinformatics, 2014. 30(9): p. 1338-9.

66

121. Raghavan, U.N., R. Albert, and S. Kumara, Near linear time algorithm to detect community structures in large-scale networks. Phys Rev E Stat Nonlin Soft Matter Phys, 2007. 76(3 Pt 2): p. 036106.

122. Carter, N.S., et al., Purine and pyrimidine metabolism in Leishmania. Adv Exp Med Biol, 2008. 625: p. 141-54.

123. Hemsworth, G.R., et al., The crystal structure of the Leishmania major deoxyuridine triphosphate nucleotidohydrolase in complex with nucleotide analogues, dUMP, and deoxyuridine. J Biol Chem, 2011. 286(18): p. 16470-81.

124. Shi, W., V.L. Schramm, and S.C. Almo, Nucleoside hydrolase from Leishmania major. Cloning, expression, catalytic properties, transition state inhibitors, and the 2.5-a crystal structure. J Biol Chem, 1999. 274(30): p. 21114-20.

125. Gerwick, W.H. and B.S. Moore, Lessons from the past and charting the future of marine natural products drug discovery and chemical biology. Chem Biol, 2012. 19(1): p. 85-98.

126. Saxena, A. and S. Mishra, Marine sponge derived natural products as inhibitors of mycothiol-S-conjugate amidase. Bioinformation, 2017. 13(8): p. 256-260.

127. Strömstedt, A.A., et al., Selective membrane disruption by the cyclotide kalata B7: complex ions and essential functional groups in the phosphatidylethanolamine binding pocket. Biochim Biophys Acta, 2016. 1858(6): p. 1317-27.

128. Dyshlovoy, S.A., et al., Aaptamines from the marine sponge Aaptos sp. display anticancer activities in human cancer cell lines and modulate AP-1-, NF-kappaB-, and p53-dependent transcriptional activity in mouse JB6 Cl41 cells. Biomed Res Int, 2014. 2014: p. 469309.

129. Pettit, G.R., et al., Antineoplastic agents 491. Synthetic conversion of aaptamine to isoaaptamine, 9-demethylaaptamine, and 4-methylaaptamine. J Org Chem, 2004. 69(7): p. 2251-6.

130. Pettit, G.R., et al., Antineoplastic agents. 380. Isolation and X-ray crystal structure determination of isoaaptamine from the Republic of Singapore Hymeniacidon sp. and conversion to the phosphate prodrug hystatin 1. J Nat Prod, 2004. 67(3): p. 506-9.

131. Shen, Y.C., et al., Structures and cytotoxicity relationship of isoaaptamine and aaptamine derivatives. J Nat Prod, 1999. 62(9): p. 1264-7.

132. Gul, W., et al., Modification at the C9 position of the marine natural product isoaaptamine and the impact on HIV-1, mycobacterial, and tumor cell activity. Bioorg Med Chem, 2006. 14(24): p. 8495-505.

133. Endo, T., et al., Nagelamides A-H, new dimeric bromopyrrole alkaloids from marine sponge Agelas species. J Nat Prod, 2004. 67(8): p. 1262-7.

134. Zidar, N., et al., Antimicrobial activity of the marine alkaloids, clathrodin and oroidin, and their synthetic analogues. Mar Drugs, 2014. 12(2): p. 940-63.

135. Tomasic, T., et al., Discovery of 4,5,6,7-Tetrahydrobenzo[1,2-d]thiazoles as Novel DNA Gyrase Inhibitors Targeting the ATP-Binding Site. J Med Chem, 2015. 58(14): p. 5501-21.

136. Bowling, J.J., et al., Antiviral and anticancer optimization studies of the DNA-binding marine natural product aaptamine. Chem Biol Drug Des, 2008. 71(3): p. 205-15.

67

137. Bhardwaj, R., et al., Evaluation of CAAX prenyl protease II of Leishmania donovani as potential drug target: Infectivity and growth of the parasite is significantly lowered after the gene knockout. Eur J Pharm Sci, 2017. 102: p. 156-160.

138. Di Giorgio, C., et al., In vitro activity of the beta-carboline alkaloids harmane, harmine, and harmaline toward parasites of the species Leishmania infantum. Exp Parasitol, 2004. 106(3-4): p. 67-74.

139. Frolova, L.V., et al., Exploring natural product chemistry and biology with multicomponent reactions. 5. Discovery of a novel tubulin-targeting scaffold derived from the rigidin family of marine alkaloids. J Med Chem, 2013. 56(17): p. 6886-900.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 265

Editor: The Dean of the Faculty of Pharmacy

A doctoral dissertation from the Faculty of Pharmacy, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofPharmacy. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Pharmacy”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-368499

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2019

Date post:	18-Jan-2021
Category:	Documents
Upload:	others
View:	3 times
Download:	0 times

Drug discovery against leishmaniasis1268389/FULLTEXT01.pdf · 2018. 12. 20. · Leishmaniasis...

Documents