In this project, we will investigate a new smart system made from a hardware component and a software approach that will enable the creation of programmable matter. The hardware component is a mass-producible, sub-mm, MEMS, namely a micro-robot using computationally controlled actuators used for power distribution, communication, adhesion, and locomotion. The software approach aims to provide a language enabling scalable, real-time, efficient, expressive and at the same time safe programming of an ensemble of micro-robots and making this ensemble interacting with others communicating things through the IoT. This research focuses on the main challenge to programmable matter: Scaling. For hardware the challenge is to scale down the size of the individual unit. For software the challenge is to scale up the number of elements that can be effectively controlled with a single, easy to understand program. Moreover, the simulation framework must scale up in the number of simulated micro-robots. We tackle the former by using a true 3-dimensional LSI/MEMS chip integration with deformable substrate as our main manufacturing method and a single effector, for all the necessary functionality of the unit. The latter is tackled creating Foxel a recursively scale-invariant functional shape description language implemented in a logic programming language, Meld, to create programs which are inherently concurrent, distributed, fault-tolerant, and also amenable to formal proofs. This project is a follow-up of the Claytronics project initiated by Intel and Carnegie Mellon University and then co-leaded with FEMTO-ST.

The first and main objective of the SIBI project, is to develop a fully original “all-electrical” sensing scheme enabling the detection and counting of single electron transfer events, a so far elusive goal in electrochemistry. The innovative methodology proposed here exploits the few millivolts voltage shift induced by a single electron stored across an attofarad range nanocapacitor. We intend to show that this electron counting technology could be the basis of ultimate sensing in bioelectrochemistry. We will therefore implement it in a configuration mimicking an electrochemical conformational DNA sensor (E-DNA). A redox-labelled single stranded DNA chain, used as a capturing (probe) strand, will be end-attached to a 10 nm size gold nanodot, fabricated by e-beam lithography, and aligned on top of a nanocapacitor. For proof of concept experiments, the nanoelectrode-tip of an atomic force electrochemical microscope, operated in molecular touching mode (Mt/AFM-SECM), will be used to address electrochemically the redox-reporter of the DNA probe strand. Discrete charging steps of the nanocapacitor, corresponding to the cycling motions of the redox label being oxidized at the tip and reduced at the nanodot, will be detected using state of the art silicon transistor nanotechnology that provides stability and elementary charge sensitivity. Observation of these steps will be the first ever demonstration of single electron counting in electrochemistry. Beyond this unprecedented detection, measurement of the frequency of the charging steps will yield unique access to quantitative information regarding the conformational dynamics of the end-anchored DNA chain. These information will be deciphered by designing cutting edge numerical simulations of the dynamics and electrochemical behaviour of end-grafted redox DNA chains. More precisely, as the second objective of this project, we propose to mimic theoretically the full biomolecular system using a realistic molecular dynamics model for DNA end-attached to electrode surfaces. Quantum transport master equations based on the Quantum Marcus theory that accounts for redox-molecule/electrode distance dependence of the electronic coupling and reorganization energy will be computed based on the position of the redox label to compute the probability of transferring an electron every picosecond. The experimental and theoretical results will be directly compared, and the numerical code adjusted. In addition, as the recent progress in machine learning for signal processing are perfectly adapted for the signal to be measured, we propose to use recurrent neural networks (RNN), to optimize the analysis of the single-electron transfer traces, in the perspective of analysing a large amount of data from sensor arrays. We could for example envision the automatic extraction of the sizes of a population of DNA strands, with single-molecule resolution Beyond its fundamental interest, our modelling approach will help gaining a quantitative understanding in the transduction mechanism of E-DNA sensors, which is still lacking to date. As a last goal of the project, the experimental and theoretical knowledge acquired will be invested in developing an actual conformational electrochemical E-DNA nano-biosensor, capable of single analyte molecule sensitivity, thanks to single electron counting and miniaturization. A sensing platform, consisting in an array of individually addressable nanocapacitors, forming an open nanogap with patterned nanoelectrodes, will be developed. The sensing chip will be employed to assay the Epithelial cell adhesion molecule (EpCAM), a prognostic marker in cancer, by making use of an EpCAM-specific redox-labeled aptamer as the capture DNA probe. We thus expect to demonstrate actual single molecule sensing, made possible by electron counting, for an assay of actual biological relevance.

Many molecules dedicated to therapy are nowadays lately removed from the pipeline of drug discovery, in the clinical phases, because their toxicity or lack of efficacy has not been evidenced in preclinical studies. Researchers and pharmaceutical companies outline the critical need in the development of advanced in vitro trials, looking for human cells based organoids able to improve screening’s efficiency and to reduce the number of animal trials, following the 3R rule. Recent progresses in bioengineering and microtechnology pave the way to “organ-on-chip” devices ensuring 3D cell culture in conditions close to physiology. However, such models are still simplistic and not strongly validated. Based on this analysis, MimLiveronChip proposes a bioinspired approach to mimic the mandatory steps allowing the study of xenobiotic’s toxicity and metabolism in the liver. The original choice is to focus on and to reproduce two key events in series: not only the biotransformation by hepatocytes of substances transported by the blood flow, but also (and upfront), their transfer (hindered or not) across the monolayer of endothelial cells, that are very specific in the liver: they are fenestrated in healthy conditions, and lose these properties in early stages of the pathology (steatosis, fibrosis, …). In this project, we will investigate the relevance of several hypotheses regarding the effect of the mechanical and biochemical micro-environment on this fenestrated status, so as to maintain it or in contrast alter it to mimic pathological cases. Our project also addresses technological issues. As end users, we know that microfluidic devices might appear complex compared to classical 2D culture. Therefore, we propose to develop a fully integrated platform, equipped with flow and pressure controllers, as well as sensors to monitor the cell culture and allow mid throughput assays for drug screening. The newly developed tools will be benchmarked with up-to-date technology employed in CRO or pharma companies. Our goal is to position the organ-on-chip platform in the pipeline of preclinical trials, combining toxicity and metabolism evaluation. This multi-disciplinary project relies on a close collaboration among four teams bringing complementary expertise: hepatic tissue engineering at different scales (UMR UTC-CNRS Biomechanics and Bioengineering); microsystems and biology of endothelial cells (SMMIL-E UMI CNRS), as far as academia is concerned, to which can be added : a SME leader in devices for microfluidic experiments (Fluigent) and a startup dedicated to the development of advanced in vitro trials based on high throughput imaging (HCS Pharma). Preliminary common studies have demonstrated the relevance and the feasibility of the proposed methodology. Regarding socio-economical features, the work performed in our project should allow the reduction of attrition rate for drug candidates in the late clinical stages. This will thus decrease the development costs and duration in the pharmaceutical industry, but also for chemicals, in the framework of REACh regulation, for the assessment of the effect of pesticides, and in agro-industry (nutraceutics). Mid throughput cell culture devices developed in MimLiveronChip can also be used for more fundamental research (system biology, regeneration studies), as they better mimic liver structure and functions. Both companies involved in the project will benefit from direct positive feedback since their portfolio will be enriched with new equipment and screening methods, respectively. Finally, demonstrating the benefit of coupling the endothelial hepatic barrier with the 3D hepatocyte transformation unit will offer many innovative strategy integrating other organs on chip, developed in the academic labs involved in the project.

Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease in western countries. It can be dichotomically divided into steatosis, associated with a benign outcome, and steatohepatitis (NASH), characterized by progression to fibrosis, hepatocellular carcinoma and increased mortality. It is estimated that 24% of European adults develop NAFLD, 10-30% of them will evolve to NASH, among which 10-15% develop fibrosis/cirrhosis. Currently, lifestyle adjustment remains the cornerstone of NAFLD management in the absence of any effective pharmacologic therapy, and the liver transplantation is the only option for NAFLD-related end-stage liver diseases. NAFLD is a complex and dynamic disease whose pathogenesis is poorly deciphered. Understanding of mechanisms involved in NAFLD development and progression is a key step for a better risk stratification. In parallel, pharmaceutical industries face bottleneck to develop new therapies because of the lack of human liver relevant models. Many animal models have been developed to reproduce NAFLD progress but they are not fully relevant due to their specific metabolisms. Simplistic 2D in vitro models are also limited when investigating NAFLD, which involves complex interactions between different cells and cells-extracellular matrix (liver is a complex organ with multicellular organization and most of cells are involved/affected during NAFLD). We hypothesize that the smart design of a complex multicellular NAFLD model will provide an unprecedented tool to investigate this disease. In this project, we plan to address 3 major goals to offer a relevant tool for the follow-up of NAFLD progression and treatment i) development of NAFLD-on-chip model from multicellular liver-on-chip mimicking in vivo liver architecture/physiology using human primary cells ii) identification of mechanisms and biomarkers involved in NAFLD progression and benchmark potential drugs iii) development of an integrated NAFLD mathematical model.

Fusion offer the promise of limitless, clean and safe energy. Numerous designs for fusion reactors have been proposed but most have been plagued by instabilities and turbulence which degrade confinement. The main route to improve confinement has been to make reactors larger and larger. This has culminated into an international project of pharaonic proportions, ITER, with a diameter of 30 m. But the economics becomes unfavorable at this gigantic scale, and it is not even sure if ITER will ever be completed. These setbacks have opened a gap for private initiatives. So far, these public and private races toward fusion have been defined by size: How big does a reactor need to be to confine a fusion plasma? Here we ask the opposite question: What is the smallest reactor that can confine a fusion plasma? A natural scale for a plasma is the Debye length ?, which is the length scale over which the charge of a particle is felt before it gets screened by opposite charges. For a fusion, the Debye length is on the order of ?~100 µm, which suggests that the microscale is a natural scale for fusion plasmas. This immediately raises a question: Can we confine a plasma in a micrometric reactor long enough to achieve fusion? This is the fundamental question that we aim to address. If microscale fusion was possible, it would open a new route as microsystems are much cheaper to build and replace, and are more portable than gigantic infrastructures. Dimensional analysis suggests that confinement should be possible with electric potentials achievable in microsystems (~1-100 kV). A major goal of this project will be to validate this dimensional analysis and estimate the confinement time achievable in actual microsystems. We will do so by combining physical simulations tailored for the regime of microscale and by fabricating prototype microsystems to generate, trap and diagnose plasmas.