L’allogreffe de cellules hématopoïétiques est une stratégie thérapeutique importante dans les hémopathies malignes. La maladie du greffon contre l’hôte (GvH) en est une complication majeure menaçant le pronostic vital. Elle est due à la reconnaissance des antigènes du receveur par les lymphocytes T du donneur et à l’activation de ceux-ci, à l’origine de dommages tissulaires. L’altération de la barrière intestinale joue un rôle critique dans la GvH. La famille des protéines de choc thermique (HSP)90 comporte cinq membres dont trois cytosoliques dénommés HSP90, et un localisé dans le réticulum endoplasmique (RE), Gp96, qui peut être sécrété en cas de stress. Nous montrons dans nos travaux de thèse que la 17AAG, un inhibiteur des HSP90, réduit la mortalité liée à la GvH dans un modèle murin. Cet effet est associé à une augmentation de la réponse au stress du RE dans les cellules épithéliales intestinales comme en atteste l’augmentation de l’épissage du facteur de transcription XBP-1, corrélée à une diminution du dommage tissulaire intestinal. Ces résultats permettent d’envisager une place pour la 17AAG ou d’autres inhibiteurs de HSP90 dans la prévention de la GvH chez l’homme. D’autre part, nous montrons que Gp96 est sécrétée dans le sérum de patients développant une GvH aiguë sévère avec atteinte intestinale. Nous suggérons de valider la pertinence de Gp96 comme biomarqueur de GvH intestinale dans une étude de plus grande ampleur. Enfin, nous trouvons que Gp96 s’associe avec le composant 3 du complément, une protéine impliquée dans l’immunité innée et adaptative, et inhibe certaines de ses fonctions. Les conséquences fonctionnelles de cette association sont discutées. / Allogeneic hematopoietic cell transplantation is a treatment for certain disorders including hematologic malignancies. Graft-versus-host-disease (GvHD) is a major, life-threatening complication. It is due to the recognition of recipient antigens by donor T cells, which activate and damage tissues. Intestinal barrier alteration plays a critical role in GvHD. Heat shock proteins (HSP)90 include five members, three cytosolic members named HSP90, and one member localized in endoplasmic reticulum (ER) called Gp96 and able to gain extracellular level in case of stress. We show in our thesis that 17AAG, a HSP90 inhibitor, reduces GvHD mortality in a mouse model. This effect is associated with an increase in ER stress pathway in intestinal epithelial cells as figured by transcription factor XBP-1 splicing, correlated to a decrease in intestinal tissue damage. These results suggest that 17AAG could be considered in GvHD prevention in human. Moreover, we show that Gp96 is secreted in serum of patients developing an acute GvHD with intestinal involvement. We propose to validate the relevance of Gp96 as an intestinal GvHD biomarker is a larger study. Finally, we find that Gp96 associate with complement component 3, a protein involved in innate and adaptive immunity, and inhibit some of its functions. Functional consequences of this association are discussed.

Molecular chaperones are a group of proteins which maintain cellular homeostasis by assisting de novo protein folding and their refolding to native state after destabilization due to external stress. They are also known as heat shock proteins as they were first discovered as a response to heat stress. It is now well established that the function of this group of proteins is not only restricted to protein homeostasis but also extends to diverse cellular processes such signal transduction, development and differentiation. Heat shock protein 90 (Hsp90) is one of the most abundant molecular chaperones that is highly conserved from prokaryotes to eukaryotes. Hsp90 is an essential chaperone and is required for the viability of all eukaryotes examined so far including yeast, Drosophila and Caenorhabditis elegans. Hsp90 has emerged as an important regulator of cellular activities by virtue of its ability to interact with a diverse set of client proteins many of which include transcription factors, protein kinases and signaling molecules. Through interaction with these proteins it is involved in regulating cellular processes including growth, cell cycle, endocrine functions, apoptosis, differentiation and development. Further in Drosophila and plants, Hsp90 is thought to function as a capacitor for morphological evolution and phenotypic variation. Recently, it has also been implicated in the emergence of drug resistance in Candida albicans. Furthermore, the importance of Hsp90 in disease states, particularly in cancer, is strongly evident, where chaperoning of mutated and oncogenic proteins is critical for continuous proliferation of cells. This has led to the development of Hsp90 inhibitors as an anti-cancer drug. Geldanamycin (GA), a benzoquinone ansamycin was the first molecule shown to inhibit Hsp90 activity by binding to its ATP binding domain. A derivative of GA, 17-allylamino-17-demethoxygeldanamycin (17AAG), has shown promise in clinical studies and has entered Phase III clinical trials. Hsp90 has been shown to be important for growth and development of many protozoan parasites. Inhibition of Hsp90 function in Leishmania, Emiera, Toxoplasma, Trypanosoma as well as Plasmodium causes a block in their developmental cycle. Previous studies from our laboratory have shown that inhibition of Hsp90 function prevents growth of malaria parasite in human erythrocytes in vitro. P. falciparum Hsp90 (PfHsp90) has also been shown to regulate parasite growth during the febrile episodes that are characteristic of malaria. While most of the studies highlighting the importance of PfHsp90 have relied on its pharmacological inhibition, its biochemical characterization and quantitative measurement of its interaction with GA in isolated system has not been explored. It was also not understood whether the in vitro model of Hsp90 inhibition could translate into inhibition of the parasite growth in an animal model of malaria. Since Hsp90 is a split ATPase requiring proper co-ordination between the residues on its N-terminal and middle domains, it would be desirable to biochemically characterize full length PfHsp90 to gain insights into its potential as an anti-malarial target. The present study was initiated with an objective of understanding the biochemical properties of Hsp90 from P. falciparum in terms of ATP binding, ATP hydrolysis and its GA binding ability. We have also examined the potential of PfHsp90 to serve as a chemotherapeutic target using its clinically well-established inhibitor, 17AAG, in a preclinical mice model. Apart from using in vitro and in vivo models of malaria, we have also explored the efficacy of 17AAG in the P. falciparum samples collected from malaria patients. Additionally, we have examined the relevance of chaperones, in particular PfHsp90 in the samples collected from malaria patients. Finally, we have attempted to understand the unexplored biology of another malaria parasite P. vivax by a high throughput proteomics approach. Biochemical characterization of PfHsp90 and its comparison with host Hsp90 Hsp90 belongs to GHKL (gyrase, Hsp90, histidine kinase, MutL) protein family having a characteristic novel ATP-binding Bergerat fold. The ATP binding pocket of GHKL family differs from the conventional nucleotide binding fold in the formation of a cone shaped pocket made up of four anti-parallel β-sheets and three α helices as opposed to parallel βsheets surrounded by α-helices in the latter. The most distinctive feature of Bergerat fold is the presence of ATP lid. Further, even within the GHKL family members the composition and the conformation of this ATP-lid differs, leading to different solvent exposure of the bound ATP. All Hsp90s from different organisms, characterized so far, have been shown to posses ATP binding and hydrolysis activity but so far PfHsp90 ATPase activity has not been characterized. Using intrinsic tryptophan fluorescence measurements, we found PfHsp90 to bind ATP with about 30% higher affinity than human Hsp90 (hHsp90). We further, 32 determined the ATPase activity of PfHsp90 by monitoring the direct conversion of (γ-P) 32-2 ATP to Pi. PfHsp90 bound and hydrolyzed ATP with a Km of 611 µM and kcat of 9.9 x 10 -1m . Interestingly, PfHsp90 showed six times higher ATPase activity as compared to its human homologue and more intriguingly the ATPase activity exhibited by PfHsp90 was highest among all the Hsp90s studied so far. Previous studies from our laboratory have provided sufficient evidence for inhibitory action of GA on Plasmodium growth inside the infected erythrocytes. GA is known to exert its inhibitory effect by binding to the ATP binding domain of Hsp90 thus inhibiting its chaperone activity. Earlier reports have shown that despite a high similarity between the ATP/GA binding region in Hsp90 from different organisms, there is a difference in their ability to bind GA. For example, in spite of all the hallmarks of ATP-binding pocket of Hsp90 family C. elegans Hsp90 does not bind GA. We have employed fluorescence spectroscopy to examine whether PfHsp90 can bind to GA. In parallel, we have also determined the binding affinity of human Hsp90 (hHsp90) to GA. We observed small but reproducible differences in the binding affinity of GA to Hsp90s from human host and P. falciparum with latter having fourfold higher affinity. A sequence analysis of the GA binding domain of Hsp90s from P. falciparum and human host showed a homologous substitution of K112 of hHsp90 to R98 in PfHsp90. In order to examine the effect of this substitution, if any, on the observed difference in GA binding abilities, we mutated R98 to K in PfHsp90. However, we did not find any difference in the binding ability of R98K PfHsp90 to GA, suggesting that this homologous substitution has minimal or no effect on drug protein interaction in vitro. However, in view of this phylogenetically conserved substitution, we cannot rule out its role in vivo. The chaperone function of Hsp90 is dependent on its ATPase activity which is susceptible to GA mediated inhibition. We next examined the extent of inhibition of GA on the ATPase activity of Hsp90s from P. falciparum and human host. Interestingly, we found the PfHsp90-ATPase activity to be three times more sensitive than hHsp90-ATPase activity to GA mediated inhibition suggesting that the malaria parasite, P. falciparum is likely to be more sensitive to GA when compared to human host. This result is in accordance with a recent study, which has shown that yeast expressing PfHsp90 in lieu of native yeast Hsp90 was more sensitive to GA than yeast expressing either yeast Hsp90 or human Hsp90. Acetylation of Plasmodium falciparum Hsp90 Post-translational modification of Hsp90 such as acetylation has been shown to affect its binding with GA. We first examined whether, PfHsp90 can be acetylated. With the use of various purified Histone acetyl transferases (HATs) of human origin, we have shown PfHsp90 to undergo acetylation in vitro. We found that among different HATs (pCAF, Gcn5 and p300) used, only p300 was able to acetylate PfHsp90 suggesting a role for it in PfHsp90 in vivo acetylation as well. We next examined the in vivo acetylation status of PfHsp90 from parasite lysate. To enrich the acetylated fraction of PfHsp90, we have used Histone deacetylase (HDAC) inhibitor, trichostatin A (TSA). Immunoprecipitation of PfHsp90 followed by immunoblotting with an acetyl-lysine antibody confirmed that PfHsp90 undergoes acetylation in vivo. In order to identify the lysine residues which underwent acetylation we subjected the acetylation enriched fraction of PfHsp90 to in-gel trypsin digestion followed by mass spectrometry. Analysis of trypsin digested PfHsp90 from the parasites identified three sites of acetylation, one of which overlapped with PfHsp90 cochaperone (Aha1 and p23) binding residue, suggesting that acetylation could play a potential role in modulating PfHsp90 multi-chaperone complex assembly. Indeed, treatment of P. falciparum cultures with a HDAC-inhibitor resulted in partial dissociation of PfHsp90 complex as observed from size-exclusion chromatography. Adding to this observation, we also found that co-treatment of TSA and GA showed a synergistic and additive effect in inhibiting parasite growth in vitro. The above results suggest the possibility of using Hsp90 inhibitor in combination with HDAC inhibitor to arrest Plasmodium growth and development. Clinically tested GA-analogue 17AAG inhibits Plasmodium growth in vitro and in vivo The specificity of GA inside the cell has been a matter of debate since the discovery of its medicinal importance. In the past, Hsp90 has been implicated as a target of GA by carrying out immunoblotting of GA pull-down fraction with an anti-Hsp90 antibody. Crystal structure of GA with yeast Hsp90 has shown it to bind within the well conserved ATP-binding pocket of Hsp90. However, the specificity of GA inside the cell is still a conjecture. We have performed GA pull down assays from the parasite lysate followed by Coomassie Blue staining, which gave a single band corresponding to 86 kDa PfHsp90. The identity of PfHsp90 was further confirmed by immunoblotting with antibody specific to PfHsp90. This result indicates that inside the cells, inhibitory effect of GA is mediated by and large through its interaction with Hsp90. However, we cannot rule out the presence of other minor, less significant, interactors of GA. Earlier work from our laboratory has shown that GA inhibits Plasmodium growth inside the infected erythrocytes. However, issues related to GA toxicity have excluded its development as a therapeutic. Nevertheless, interest in this class of molecule has led to the generation of a large number of less toxic derivatives of GA. One classical example is 17AAG which has gained clinical importance over the years and has entered in phase III trial. Intrigued by the clinical success of 17AAG, we were interested in determining its ability to modulate parasite growth. Indeed, 17AAG was able to inhibit parasite growth in a manner similar to that of GA. We further extended our study to parasites isolated from patient samples. Here too, we found 17AAG to be effective in inhibiting growth of the parasite. Finally, we examined the efficacy of 17AAG at a pre-clinical level using a mouse model of malaria. Using Peters’ four-day test we found 17AAG, to be effective in attenuating parasite growth and prolonging the survival of parasite infected mice (n=4, p=0.00692; n=10, p=0.001). Clinical relevance of heat shock proteins of Plasmodium falciparum A recent study using in vivo expression profiles of parasites derived from blood samples of infected patients has revealed previously unknown physiological diversity in the biology of malaria parasites. According to gene expression profiles, parasites were clustered into three different physiological states – starvation, glycolysis dependent active growth and environmental stress response. In order to examine the clinical relevance of molecular chaperones in malaria, we reanalyzed the previously published gene expression data of clinical parasites from 46 patients. Our analysis of this data showed that organellar chaperones were up-regulated upon starvation (cluster1) while cytosolic chaperones such as Hsp90 were up-regulated in active growth conditions (cluster2) indicating up-regulation of distinct group of Hsps in response to different environmental cues. Interestingly, Hsp90 and its co-chaperones, previously implicated as drug targets in malaria, clustered in the same group. Further, some patients of cluster 3 (environmental stress response) showed higher expression of Hsp90 while others showed lower expression. In general, cluster 3 group of patients were heterogeneous in terms of expression of chaperones. Using non-negative matrix factorization (NMF), cluster 3 was sub-clustered into two groups 3a and 3b. Cluster 3b showed up-regulation of cytosolic chaperones similar to cluster 2 indicating these two clusters are inter-related. Most of the Hsp90 dependent pathways such as trafficking, signaling, anti apoptotic and pro-survival found to be most active in cluster 2 indicating the dependence of this group of parasites on Hsp90. The two main outcomes of our chaperone analysis are (1) the up-regulation of molecular chaperones in parasites are not a general response to hostile conditions as perceived previously, but is largely determined by the host factors and may differ from one host to another (2) the disease specific pathways may exist in natural condition by the up-regulation of specific chaperone and its interactors as a response to different host environment. Clinical proteomics of human malarial parasites Much of our understanding about the life cycle of parasites and importance of parasite proteins have been gleaned from the studies in laboratory strain or with the laboratory adapted clinical parasites. Although, these studies provide us first hand information about the functionality and the importance of these proteins, but they often fail to mimic the actual disease environment. In the patient, parasites are exposed to host factors such as hormones, metabolites, inflammatory mediators which can influence the expression of proteins and thus parasite biology. Further, instead of parasite exposure to 37°C temperature throughout the erythrocytic cycle in vitro, it is exposed to several rounds of febrile episodes inside human, which can also influence the parasite life cycle. Furthermore, clinical analysis is important to validate the presence and expression of drug targets in actual disease environment. Therefore, analysis of malaria parasite from clinical settings has become an important component in our laboratory and this thesis. Proteomic analysis of clinical samples has emerged as an important tool to understand the proteins dynamicity as response to disease environment. We have initiated clinical proteomic study of P. falciparum, the cause of most common and fatal malaria in humans and extended it further to the neglected malaria parasite P. vivax. The study of P. vivax has largely been over-shadowed by the enormous attention devoted to P. falciparum. Notably, the drugs which have been discovered against P. falciparum are not as effective against P. vivax. Further several unique features of P. vivax such as dormant hyponozoites, reticulocyte host preference and formation of specialized caveolae vesicle complex structure distinguish its biology from P. falciparum and warrant concerted effort directed at this parasite. A major limitation in studying this parasite is the absence of a long-term culturing system. Therefore, research on this parasite requires samples obtained directly from patients. In spite of the inherent difficulty in obtaining such samples, this method provides us an opportunity to study this parasite in its real environment which has a huge effect on the expression as well as function of parasites and host proteins. Our current knowledge about the life cycle of this parasite has been gained from the recently published transcriptome study. Even though transcriptome analyses provide useful understanding at the level of gene expression, they do not reflect the active protein component of a cell. In other words, most of disease outcome is a result of interaction of the protein component with the environment. We therefore attempted to understand the protein component of this parasite in the disease environment to shed light on its pathogenicity. Despite facing several challenges in the way of proteomic analysis of this parasite such as availability of samples, low parasitemia, contamination of parasite proteins with abundant host proteins etc, we were able to identify 154 P. vivax proteins abundantly expressed in clinical environment using mass-spectrometry based approach. We found many proteins unique to this parasite along with known drug targets. This study is the first of its kind and could prove to be a very important step towards gaining insights into the physiology of this parasite.This study serves as a proof-of-principle method which in future is likely to help in identifying many more potential drug targets, vaccine candidates and diagnostic markers from clinically relevant samples as opposed to cultured samples. Summary Despite the importance of PfHsp90 in malaria biology, it has not been characterized in terms of its biochemical properties and its interaction with the inhibitor. In this study, we have successfully cloned, expressed, purified and characterized full length PfHsp90. We found that PfHsp90 exhibits a hyper-ATPase activity and is more sensitive to GA mediated inhibition as compared to human Hsp90. We have also shown that its sensitivity towards GA is dependent on its acetylation status as treatment of infected erythrocytes with HDAC inhibitors increases its sensitivity to GA. Using a pull-down assay, we have determined, unequivocally, that GA specifically binds to Hsp90. Most importantly, we have demonstrated that 17AAG, a clinically well-established inhibitor of Hsp90, inhibits parasite growth in a laboratory strain, field isolates and an in vivo mouse model of malaria. Overall, our biochemical characterization and drug interaction studies underscore the importance of PfHsp90 as a potent drug target and its inhibitors as a candidate drugs for the treatment of malaria, one of the deadly human infectious diseases. Our efforts to understand the importance of molecular chaperones in parasites isolated directly from patient samples (clinical setting) has revealed conspicuous association of Hsps with previously defined parasite physiological states. In particular, parasites obtained from a specific group of patients exhibited heightened dependence on Hsp90-dependent pro-survival pathways, indicating an increased response to host stressors in this group of parasites. Thus, parasite encoded chaperones, in particular PfHsp90, play a major role in defining the pathogenesis of malaria. A disease is an outcome of interaction between pathogens and its host, therefore it is important to study parasite in its real environment to understand disease pathogenesis. Our lab has previously reported the first ever proteomic analysis of P. falciparum from malaria patients. In this study, we have made an attempt to understand the unexplored biology of another important malaria parasite P. vivax. We have used a mass-spectrometry based approach to identify the protein content of this parasite. This technically challenging attempt has enabled us to identify many proteins. This study is an important step towards understanding the biology of this parasite in dearth of any information available on the proteins involved in this parasite’s pathogenicity.