Cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP) are second messengers that relay external signals to downstream effector proteins. In cardiomyocytes and smooth muscle cells, cyclic nucleotides control several key functions such as the relaxation tone or the proliferative / hypertrophic phenotype. The modulation of cyclic nucleotides levels thus appears as an appealing therapeutic objective. Signalling events triggered by extracellular stimuli arise from an ingenious regulation of intracellular cyclic nucleotide levels, which result from a balance between their production and their elimination. Classically, cyclic nucleotide elimination has been attributed to hydrolysis mediated by cyclic nucleotide phosphodiesterases (PDEs). However, PDEs may not be the sole regulator of cyclic nucleotide pathways. Our first results have indeed established that efflux of cyclic nucleotides out of cells represents an alternative or complementary mechanism, in addition to PDEs, ensuring intracellular cyclic nucleotide homeostasis. We have indeed identified the expression of an energy-dependent efflux pump called MRP4 (multidrug-resistance-associated protein MRP4, ABCC4) in smooth muscle cells as in cardiomyocytes and demonstrated that MRP4 is responsible for an active transmembrane efflux of cyclic nucleotides thereby limiting the activation of mediated signal transduction. Specific inhibition of MRP4 in cells modified the intracellular content of cyclic nucleotides and markedly enhanced their signalling effect in vitro and in vivo. We thus show that MRP4 is an important and independent regulator of endogenous cAMP and cGMP levels. The modulation of MRP4 activity or expression might therefore have therapeutic potential by modulating cAMP- and cGMP-mediated signal transduction. The following project will aim to further specify the function of MRP4 in the cardio-vascular system and certify the therapeutic impact of its inhibition. In addition to investigations on the prevention of coronary smooth muscle cells proliferation, we will focus on other cardio-vascular diseases where cyclic nucleotides may play a key role. Specifically, we will study the physiopathological and pharmacological role of MRP4 and MRP5 in the prevention of pulmonary artery hypertension where the inhibition of smooth muscle proliferation may represente a therapeutic challenge. Furthermore, we will study the role of MRP4 in the heart and more specifically on the hypertrophic remodelling and on the inotropic response, both phenomenon which are at least partly controlled by cyclic nucleotides-mediated signalling pathways. Our research strategy will be based on the use of RNA interference inserted in viral vectors (as already performed for our previous experiments), and on the investigations of knock-out mice (MRP4-/-). On top of cellular experiments in isolated cardiomyocytes and/or smooth muscle cells, several pathological models (aortic constriction and isoproterenol infusion for the heart; chronic hypoxy for pulmonary hypertension) will be performed. This scientific program will allow validating the function of MRP4 in the cardio-vascular system.

From isolated case reports, the scale of the human immunodeficiency virus (HIV) epidemic has become a global pandemic, and HIV the most extensively studied and notorious pathogen in history. 35 million people are estimated to live with HIV by the end of the 2008. Based on compelling evidence that CD8+ T cells are key players in the immune response to HIV, the scientific community has focused enormous efforts on the development of therapeutic and prophylactic vaccines that elicit T cell responses with the premise that such interventions will be sufficient to confer protective effects against HIV. To date however, the development of effective T cell vaccines remains elusive, which is highlighted by the recent failure of the Merck STEP study, the most ambitious clinical trial performed to date to tackle the HIV pandemic. Consensual opinion is that our general understanding of T cell efficacy against viruses in humans is actually still limited. Without such knowledge, vaccine design strategies will remain largely empirical and further failures are likely. Although it is clear that CD8+ T cell efficacy depends on qualitative rather quantitative attributes, the exact nature of effective CD8+ T cells remains poorly understood. The study of CD8+ T cell immunity is the central theme of my research. During the last 10 year period, my work has focused on the characterization of CD8+ T cells in different contexts, infectious or cancerous. The research program I am proposing to perform here is in the direct continuity of this work and represents the synergy of my expertise acquired in Oxford, Lausanne and Paris. The aim of the present proposal is to better understand the factors that rule the development and maintenance of effective CD8+ T cell responses against HIV. I have recently reported that CD8+ T cells associated with superior control of HIV replication present a high sensitivity for HIV antigens. The level of antigen sensitivity determines the strength of the stimulus received by T cells on exposure to defined densities of cognate antigen. Since the functional outcome of antigen engagement depends on the derived stimulus, the degree of sensitivity has consequences beyond those for simple T cell activation kinetics. Based on the premise that antigen sensitivity is central for CD8+ T cell efficacy, I propose to return to basic science to explore in details various aspects of CD8+ T cell generation and function in HIV infection. I plan to use a variety of approaches based on well established and novel technologies, together with the privileged access to a variety of biological samples available at the Hôpital Pitié Salpétrière in Paris, where I am established. I propose to investigate several issues that are central for our understanding of the establishment of effective CD8+ T cell responses in humans; the project is divided in 4 parts accordingly: 1. Importance of naïve HIV reactive CD8+ T cell precursor frequency for efficacy of CD8+ T cell responses. The frequency of naïve T cell precursors reactive for specific HIV antigens may be directly related to the qualitative attributes of CD8+ T cell memory populations, for instance by providing a more diverse clonotypic repertoire, from which selection of high quality clonotypes (i.e. with high antigen sensitivity) may be more probable. 2. Significance and interplay of the different attributes of CD8+ T cell efficacy. The interplay between the major factors important for CD8+ T cell efficacy in HIV infection (HLA restriction, target epitope and antigen sensitivity) and their respective weight for anti-viral efficacy remain unclear. 3. Determinants of antigen sensitivity at the effector-target interaction level. Beyond the affinity of the expressed T cell receptor (TCR) for antigen, other factors related to the CD8+ T cell itself or to the target cell and viral factors may be important for antigen sensitivity and ultimately affect efficacy to suppress HIV replication. 4. Exhaustion of immune resources and maintenance of effective CD8+ T cells. Continuous renewal of exhausted antigen specific T cells is crucial for the maintenance of effective immunity against persistent viruses. However, the renewal of effective CD8+ T cells may be compromised in HIV infected patients, due to the exhaustion of global primary immune resources during the infection.

The proof of concept of correlative microscopy has been demonstrated for the observation and the characterization of biological cells involving Electron Microscopes (EMs). Nevertheless, up to date, correlative techniques involving the EM suffer from the fact that they are not yet able to provide a precise 3D rendering and topography informations. The ambition of the project is to provide a complete robotic instrument integrated in an EM to give to an operator without engineering skills a precise 3D rendering of intercellular structures with a typical size between 10 nm and 500 nm. We argue that embedding an AFM probe on a high-speed multi-dof nano-robotic system able to ensure a precise positioning of the probe in the three-dimensional space inside EMs will deeply modify the way of how topography images and force characterization are obtained at the nano-scale.

Fab labs are workshops which make powerful fabrication machines available to a wide audience to create physical and computational artifacts. The machinery available in fab labs was previously solely available to experts, partly because it is expensive and partly because their use requires a certain level of expertise. Yet no prior knowledge is required to become a fab lab user, and the main ways in which users can acquire knowledge is through tutorials offered by fab lab staff, through guides and documentations found online, and through free exchange between users. However, documentation is mostly seen as a burden and thus commonly neglected by fab lab users. Recognizing both the importance of sharing knowledge and the overhead that documentation puts on users, we aim with this project to integrate knowledge documentation into the fabrication process such that (1) the extra effort required to document is considerably reduced, and (2) users experience the benefit of documenting their activities more directly by being able to make immediate use of prior knowledge during their fabrication activities. Our approach consists of applying the principles of information visualization in the physical context of fab labs. More specifically, we apply the concept of situated visualization. Simply showing information on a regular computer screen is maladaptive to the requirements of physical environments where people tend to move around to accomplish different tasks. However, with the use of situated visualization techniques that show data when and where they are most relevant, data can be shown in close proximity to the place where a user is working thereby adapting gracefully to the requirements of fab lab activities. Our proposal is the first attempt to combine two emerging and growing research areas in information visualization – situated analytics – and in human computer interaction – interaction techniques for fabrication. Situated data visualization show data directly near the physical space, object, or person generating the data. They have many potential benefits such as making information in environments visible such that the data can analyzed in the physical context that generated the data. Furthermore they make it easier to adapt one's actions in the physical world in real-time based on the shown data. Finally, they support collaborative analysis by people sharing the same environment. All of these properties make them a well-adapted choice for the use in physical fab lab environments. Our work program is broken down into five work packages: * WP0: Coordination and management to coordinate with the collaborating fab labs and to organize communication of our results through websites and public presentations. * WP1: Situated tools and techniques to capture activities in fab labs and to reflect the captured data back into physical spaces. * WP2: Knowledge representation leading up to situated visualization techniques augmenting fab lab activities. * WP3: Evaluation of results and hypotheses from the previous work packages through rigorous user studies. * WP4: Tracking and automation to extend the system to include less well defined activities that require explicit tracking of user activities within the fab lab space. This project will have an impact on multiple domains. Scientifically, the project will advance our understanding of how people address and solve challenges in fab labs through technology. Technologically, it will produce prototypes which will be released as open-source software and hardware and which can be expanded into startups by the students involved in this project. Socially, the project will make fab labs more accessible to the general public. Locally, the project will amend the fab labs at UPMC and Digiscope by integrating emerging technologies. The project will fund one PhD student, one engineer, and 2 MSc internships.

The GORILLA project aims to improve our understanding of the security level provided by cryptographic schemes used to secure sensitive data and communications, be them widely used or only in development. Its main goal consists in improving the concrete efficiency of some cryptographic "attacks" capable of breaking security guarantees offered by these cryptographic schemes. Public-key cryptographic schemes, including digital signature and key-exchange, are required in the TLS protocol to establish secure connections all over the internet. Their security relies on the assumption that well-known computational problems are hard, including integer factoring, computation of discrete logarithms modulo a prime or in elliptic curves, solving systems of Boolean quadratic equations, etc. Brutal, unpredictable algorithm improvement and regular progress in computational power may both allow cryptanalysts to break larger and larger instances of these problems that have become "security assumptions" in modern cryptography. This calls for periodically reevaluating the sizes of parameters of the corresponding cryptographic schemes. For this reason, 1024-bit RSA keys, that were once common, are now deprecated. It is therefore necessary to continuously reassess the hardness of some hard computational problems, both in theory and in practice. The project fits into this perspective, through three main objectives: first, improve some algorithms used in cryptographic attacks or invent new, more efficient ones. Second, produce high-quality software implementations of some cryptographic attacks and put them into the public domain. Third, run these program on large parallel computers. The existence of public reference implementations of cryptographic attacks is invaluable to allow the public (or other researchers) to evaluate the security of cryptographic mechanisms and elaborate their own opinion. They may be used to break weak cryptographic schemes (or weak parameters in good cryptographic schemes, such as in the LOGJAM attack), but this is then a public service. The project intends to run the attack software it will develop on a record scale using a large computing infrastructure in order to demonstrate their quality but above all to make a public statement about the scale of feasibility of the attacks. More precisely, the project focuses on a few specific algorithmic problems, which play an important role in the security of either current or future cryptographic schemes. The project first intends to improve some algorithmic aspect of the Number Field Sieve, which is the state-of-the-art algorithm to factor large integers or compute discrete logarithms modulo a prime. The planned improvements concern both the sieving phase and the linear algebra phase; they will be incorporated into CADO-NFS, the reference open-source software developed at INRIA. The, the project will focus on the parallel collision search algorithm of van Oorschot and Wiener, which has many applications in cryptanalysis; it is notably the best algorithm to compute discrete logarithms on elliptic curves. The project will produce a public high-quality implementation and use it to break the Certicom ECC2K-130 challenge, which consists in computing a discrete logarithm over an elliptic curve defined over a 130-bit binary field. The project will then use the same implementation to demonstrate a large "meet-in-the-middle" attack against double encryption, for instance by breaking the double-DES in practice (this has never been done). Lastly, the project will tackle the issue of solving systems of Boolean quadratic equations. This problem is relied upon by "post-quantum" digital signature schemes currently competing for standardization. The point is to better understand the behavior of a recent algorithm of Joux-Vitse and to compare its practical efficiency with that of exhaustive search (or to combine both). Lastly, a significant computational record is sought.