We present the principle of a fast magnetic field enhanced colloidal agglutination assay, which is based on the acceleration of the recognition rate between ligands and receptors induced by magnetic forces. By applying a homogeneous magnetic field of 20 mT for only 7 s, we detect CRP (C-reactive protein) in human serum at a concentration as low as 1 pM for a total cycle time of about 1 min in a prototype analyzer. Such a short measurement time does not impair the performances of the assay when compared to longer experiments. The concentration range dynamic is shown to cover 3 orders of magnitude. An analytical model of agglutination is also successfully fitting our data obtained with a short magnetic pulse.

We present the principle of a fast magnetic field
enhanced colloidal agglutination assay, which is based on the
acceleration of the recognition rate between ligands and
receptors induced by magnetic forces. By applying a
homogeneous magnetic field of 20 mT for only 7 s, we detect
CRP (C-reactive protein) in human serum at a concentration as
low as 1 pM for a total cycle time of about 1 min in a prototype
analyzer. Such a short measurement time does not impair the
performances of the assay when compared to longer experiments.
The concentration range dynamic is shown to cover 3 orders of
magnitude. An analytical model of agglutination is also successfully fitting our data obtained with a short magnetic pulse.

Colloids are used in many applications such as contrast agents in MRI or magnetic immuno-agglutination in diagnosis. To improve their performance, the interaction between particles and biological media – in particular proteins – has been studied for more than 150 years; nevertheless this phenomenon is still not well described.
The first part of this thesis is dedicated to an application of colloids: detection of target proteins. Based on orientation of anisotropic magnetic aggregates, a new method has been developed, which generates a signal of aggregation, directly proportional to the amount of protein in the sample: lowering the detection limit. This technique has been validated on a real system: C-reactive protein.
The second part is dedicated to protein adsorption on colloids. The first objective was the development of a protocol to obtain reliable measurements; adsorption could then be characterized in a model system: bovine serum albumin and silica. Results obtained were thus able to dismiss certain paradoxes and to propose a new model.
Ultimately the studies reveal how proteins, by adsorbing between two particles, aggregate colloids. In sufficient quantities, they can also protect against aggregation, enabling a new method of particle stabilization.

Les tests de diagnostic ont pour but de détecter et de quantifier un ou plusieurs analytes en milieu complexe. Depuis les années 80, leur évolution s'est concentrée sur la détection de très faibles quantités d'analytes. Actuellement, cette ligne directrice subsiste mais en parallèle les développements se portent sur de nouveaux objectifs comme la mise au point de tests simples, rapides, transportables et sensibles afin de répondre au mieux aux attentes des nouveaux marchés : le point of care et les pays émergents. Au cours de ce travail, nous nous sommes confrontés à ces nouvelles problématiques. Nous avons choisi d'améliorer la sensibilité du test d'agglutination magnétique précédemment développé au laboratoire en raison de sa simplicité (test homogène) et de sa rapidité (manipulation de billes magnétiques). Nous avons montré que deux paramètres conditionnaient le seuil de détection de ce test à savoir le bruit (variabilité du signal sans antigène) et le signal émis par les billes. Le premier facteur dépend principalement du nombre de billes et le deuxième nous a amené à travailler avec de grandes billes nécessitant une nouvelle méthode de détection, la cytométrie en flux. Le gain en sensibilité n'a pas été clairement démontré en raison de la très faible réactivité des billes utilisées (500 fois moins réactives que les billes "habituelles"). Une chose est sûre, de grandes billes améliorent le signal, la cinétique et la thermodynamique au détriment d'une plus faible mobilité des objets sous champ. Sous couvert d'une bonne réactivité, nous nous sommes aperçus que la première étape de notre test d'agglutination devenait thermodynamiquement et cinétiquement limitante aux faibles concentrations en analyte. Dans le cadre du projet européen DetectHIV, nous avons essayé d'une part de franchir ces barrières à travers un système microfluidique et d'autre part de mettre au point un test complètement intégré. De nouvelles technologies ont vu le jour à l'EPFL (plug magnétique) et à DTU (cytométrie sur puce). Le système développé par l'EPFL est thermodynamiquement favorable, valide l'idée de préconcentration mais nécessiterait un nouveau dimensionnement fluidique pour réellement supplanter les tests en volume. L'intégration, quant à elle, a permis d'acquérir un certain savoir faire dans le domaine de la microfluidique, d'identifier les verrous technologiques et de développer un nouveau mode de circulation des fluides. Pour conclure, ce travail a permis de cerner l'évolution possible des tests d'agglutination magnétique. La sensibilité du test originel c'est-à-dire en volume couplé à une détection turbidimétrique est fixée par le bruit de l'appareil de mesure. Ce paramètre étant peu optimisable, seule une augmentation de la taille des billes permettrait d'augmenter significativement la sensibilité du test (un facteur 10 en théorie). Au delà, une détection individuelle des objets via la cytométrie est nécessaire. D'une part, elle permet de travailler avec de gros objets afin d'amplifier grandement le signal et d'autre part, elle permet de s'affranchir du bruit de l'appareil de mesure. Le bruit mesuré est uniquement dû à la comptabilisation des populations, bruit inhérent au test de diagnostic. Avec cette configuration, un gain de sensibilité passe par une diminution du non-spécifique (relatif au comptage des populations) et par une augmentation de la réactivité des "grosses" billes. En dessous de la dizaine de femtomoles, la formation de complexe n'est plus assurée et devient longue, un changement de format fluidique est alors nécessaire. Le nouveau format fluidique testé, la microfluidique, s'avère très prometteur à condition d'intégrer une méthode de détection (cytométrique ou autre) et d'augmenter les débits.

We discuss the theory of ligand receptor reactions between two freely rotating colloids in close proximity to one other. Such reactions, limited by rotational diffusion, arise in magnetic bead suspensions where the beads are driven into close contact by an applied magnetic field as they align in chainlike structures. By a combination of reaction-diffusion theory, numerical simulations, and heuristic arguments, we compute the time required for a reaction to occur in a number of experimentally relevant situations. We find in all cases that the time required for a reaction to occur is larger than the characteristic rotation time of the diffusion motion tau_rot. When the colloids carry one ligand only and a number n of receptors, we find that the reaction time is, in units of tau_rot, a function simply of n and of the relative surface alpha occupied by one reaction patch alpha = pi r_C²/(4pi r²), where r_C is the ligand receptor capture radius and r is the radius of the colloid.

We introduce a general methodology based on magnetic colloids to study the recognition kinetics of tethered biomolecules. Access to the full kinetics of the reaction is provided by an explicit measure of the time evolution of the reactant densities. Binding between a single ligand and its complementary receptor is here limited by the colloidal rotational diffusion. It occurs within a binding distance that can be extracted by a reaction-diffusion theory that properly accounts for the rotational Brownian dynamics. Our reaction geometry allows us to probe a large diversity of bioadhesive molecules and tethers, thus providing a quantitative guidance for designing more efficient reactive biomimetic surfaces, as required for diagnostic, therapeutic, and tissue engineering techniques.

When ligands and receptors are both attached on surfaces, because of the restriction of configurational freedom, their recognition kinetics may be substantially reduced as compared with freely diffusing species. In nature, this reduction may influence the efficiency of the capture and adhesion of circulating cells. Here we show that similar consequences are observed for colloids grafted with biomolecules that are used as probes for diagnostics. We exploit Brownian magnetic colloids that self-assemble into linear chains to show also that the resulting one-dimensional confinement considerably accelerates the recognition rate between grafted receptors and their ligands. We propose that because confinement significantly augments the colliding frequency, it also causes a large increase in the attempt frequency of the recognition. This work gives the basis of a rapid, homogeneous, and highly sensitive bioanalysis method.

A new general methodology based on magnetic colloids to study the recognition kinetics of tethered biomolecules is introduced here. By tuning the architecture of the tethers that link molecules to surfaces, we explain the large diversity of observed association times. We thus studied the association kinetics between a streptavidin and a biotin grafted on two opposite surfaces. We also considered the case of cellular adhesion molecules, the cadherins, for which we measured an association kinetic constant smaller than that of the strepatividin-biotin interaction.

We present a new technique to measure the mechanical properties of small biomolecules. This technique uses long range repulsive colloidal forces together with magnetic attraction as a force probing tool. The biomolecules are grafted between superparamagnetic particles, which are regularly spaced within long chains maintained by an external magnetic field. Varying the magnetic field results in compression or extension of the molecules between the particles. In order to demonstrate this technique we use, as a size controlled model molecule, a short double stranded DNA (151 base pairs) for which the force-extension law is determined and found in agreement with existing predictions.