Frédéric Mallard

Direct identification of clinically relevant bacterial and yeast microcolonies and macrocolonies on solid culture media by Raman spectroscopy.

Abstract. Decreasing turnaround time is a paramount objective in clinical diagnosis. We evaluated the discrimination power of Raman spectroscopy when analyzing colonies from 80 strains belonging to nine bacterial and one yeast species directly on solid culture medium after 24-h (macrocolonies) and 6-h (microcolonies) incubation. This approach, that minimizes sample preparation and culture time, would allow resuming culture after identification to perform downstream antibiotic susceptibility testing. Correct identification rates measured for macrocolonies and microcolonies reached 94.1% and 91.5%, respectively, in a leave-one-strain-out cross-validation mode without any correction for possible medium interference. Large spectral differences were observed between macrocolonies and microcolonies, that were attributed to true biological differences. Our results, conducted on a very diversified panel of species and strains, were obtained by using simple and robust sample preparation and preprocessing procedures, while still confirming published results obtained by using more complex elaborated protocols. Instrumentation is simplified by the use of 532-nm laser excitation yielding a Raman signal in the visible range. It is, to our knowledge, the first side-by-side full classification study of microorganisms in the exponential and stationary phases confirming the excellent performance of Raman spectroscopy for early species-level identification of microorganisms directly from an agar culture.

Micro-organism extraction from biological samples using DEP forces enhanced by osmotic shock

On the road towards efficient diagnostics of infectious diseases, sample preparation is considered as the key step and remains a real technical challenge. Finding new methods for extraction of micro-organisms from a complex biological sample remains a major challenge prior to pathogen detection and analysis. This paper reports a new technique for capturing and isolating micro-organisms from a complex sample. To achieve the segregation of pathogens and blood cells, dielectrophoretic forces applied to bioparticles previously subjected to an osmotic shock are successfully implemented within a dedicated microfluidic device. Our device involves an electrode array of interdigitated electrodes, coated with an insulating layer, to minimize electrochemical reactions with the electrolyte and to enable long-time use. The electric field intensity inside the device is optimized, considering the insulating layer, for a given frequency bandwidth, enabling the separation of bioparticles by dielectrophoretic forces. Our predictions are based on analytical models, consistent with numerical simulations (using COMSOL Multiphysics) and correlated to experimental results. The method and device have been shown to extract different types of micro-organisms spiked in a blood cell sample. We strongly believe that this new separation approach may open the way towards a simple device for pathogen extraction from blood and more generally complex samples, with potential advantages of genericness and simplicity.

Viability of 3 h grown bacterial micro-colonies after direct Raman identification

Clinical diagnostics in routine microbiology still mostly relies on bacterial growth, a time-consuming process that prevents test results to be used directly as key decision-making elements for therapeutic decisions. There is some evidence that Raman micro-spectroscopy provides clinically relevant information from a limited amount of bacterial cells, thus holding the promise of reduced growth times and accelerated result delivery. Indeed, bacterial identification at the species level directly from micro-colonies at an early time of growth (6h) directly on their growth medium has been demonstrated. However, such analysis is suspected to be partly destructive and could prevent the further growth of the colony needed for other tests, e.g. antibiotic susceptibility testing (AST). In the present study, we evaluated the effect of the powerful laser excitation used for Raman identification on micro-colonies probed after very short growth times. We show here, using envelope integrity markers (Syto 9 and Propidium Iodide) directly on ultra-small micro-colonies of a few tens of Escherichia coli and Staphylococcus epidermidis cells (3h growth time), that only the cells that are directly impacted by the laser lose their membrane integrity. Growth kinetics experiments show that the non-probed surrounding cells are sometimes also affected but that the micro-colonies keep their ability to grow, resulting in normal aspect and size of colonies after 15h of growth. Thus, Raman spectroscopy could be used for very early (<3h) identification of grown micro-organisms without impairing further antibiotics susceptibility characterization steps.

Optimization of dielectrophoretic separation and concentration of pathogens in complex biological samples

Sample preparation is a key issue of modern analytical methods for in vitro diagnostics of diseases with microbiological origins: methods to separate bacteria from other elements of the complex biological samples are of great importance. In the present study, we investigated the DEP force as a way to perform such a de-complexification of the sample by extracting micro-organisms from a complex biological sample under a highly non-uniform electric field in a micro-system based on an interdigitated electrodes array. Different parameters were investigated to optimize the capture efficiency, such as the size of the gap between the electrodes and the height of the capture channel. These parameters are decisive for the distribution of the electric field inside the separation chamber. To optimize these relevant parameters, we performed numerical simulations using COMSOL Multiphysics and correlated them with experimental results. The optimization of the capture efficiency of the device has first been tested on micro-organisms solution but was also investigated on human blood samples spiked with micro-organisms, thereby mimicking real biological samples.

Fluorescence coupling into structured waveguide as platform for optical portable sensors

Optical chemical sensors and biosensors are attracting research interest in applications such as environmental monitoring and biomedical diagnostics. Structured Integrated Optical Waveguide is one solution to reduce the readers cost and size. The principle is the capture of fluorescence emitted by Qdots at the surface of a rib waveguide, which collects then guides it at the end-face of the chip to be detected. However, fluorescence coupling into a waveguide is still not easy to predict as it depends on fluorophores environment and dipoles orientation and location. We report here the validation of a simple theory concerning optimization of optical waveguides thickness considering a fluorophores position. Optimisation of coupling power between a dipole and a guided mode can be simplified by the optimisation of the guided modes intensity ratio integrated in the 5 nm region over the guides core surface (where QDots are supposed to settle) divided by the whole guided intensity. A model has been developed from the work of Marcuse1: coupled power is proportional to the square of the electrical field of the guided wave. As a result, this model gives an optimal cores thickness and efficiency of coupling depends on polarisation. Moreover, FDTD simulations do complete this study. Three thicknesses have been therefore experimentally deposited: 100 nm, 125 nm and 150 nm. To conclude, experimentation corresponds to the model. A new, sensitive and potentially low cost portable transducer for the analysis of all kinds of biomolecular affinity systems has been developed and validated.

Towards an easy-to-use tuberculosis diagnosis through exhaled breath analysis: a liquid fluorimeter with an excitation at 265 nm

The struggle against tuberculosis is one of the World Health Organization priorities. Identifying in a short time, patients with active tuberculosis, would bring a tremendous improvement to the current situation. Recovering from this infectious and deadly disease (2 million of death per year) is possible with a correct diagnosis to give an appropriate treatment. Unfortunately, most common tuberculosis diagnoses have few drawbacks: - skin tests: not reliable at 100% and need an incubation of 2 days before the diagnosis, - blood tests: costly and sophisticated technology, - chest X-ray: the first step before the sputum tests used for a bacterial culture with a final diagnosis given within 2 weeks. A tuberculosis test based on exhaled breath analysis is a prospective and noninvasive solution, cheap and easy to use and to transport. This test lies on a fluoregenic detection of niacin, a well-known mycobacterium tuberculosis specific metabolite. In this paper, it is assumed that the selected probe is specific to niacin and that exhaled breath does not contain any interfering species. To address this problem, a fluorimeter is developed with a cheap and cooled CCD ( 2k

Culture-free Antibiotic-susceptibility Determination From Single-bacterium Raman Spectra

) as a sensor, to easily determine the suitable “fluorescent zone”. In comparing aqueous solutions with and without niacin, 250 pM of niacin have been detected. With a commercial fluorimeter (Fluorolog from Horiba), only 200 nM of niacin are detected. The present detection remains 10 times above the estimated targeted value for a tuberculosis test. The excitation source is a LED, which typically emits 20 W at 265 nm through an optical fiber. The emission signal is detected around 545 nm. A typical light exposure lasts 700 seconds. Analysis of biomarkers with a liquid fluorimeter is generic and promising as health diagnosis.

Fast Raman single bacteria identification: toward a routine in-vitro diagnostic

Raman spectrometry appears to be an opportunity to perform rapid tests in microbiological diagnostics as it provides phenotype-related information from single bacterial cells thus holding the promise of direct analysis of clinical specimens without any time-consuming growth phase. Here, we demonstrate the feasibility of a rapid antibiotic-susceptibility determination based on the use of Raman spectra acquired on single bacterial cells. After a two-hour preculture step, one susceptible and two resistant E. coli strains were incubated, for only two hours, in the presence of different bactericidal antibiotics (gentamicin, ciprofloxacin, amoxicillin) in a range of concentrations that included the clinical breakpoints used as references in microbial diagnostic. Spectra were acquired and processed to isolate spectral modifications associated with the antibiotic effect. We evidenced an “antibiotic effect signature” which is expressed with specific Raman peaks and the coexistence of three spectral populations in the presence of antibiotic. We devised an algorithm and a test procedure that overcome single-cell heterogeneities to estimate the MIC and determinate the susceptibility phenotype of the tested bacteria using only a few single-cell spectra in four hours only if including the preculture step.

Opto-electronic DNA chip: high performance chip reading with an all-electric interface.

Timely microbiological results are essential to allow clinicians to optimize the prescribed treatment, ideally at the initial stage of the therapeutic process. Several approaches have been proposed to solve this issue and to provide the microbiological result in a few hours directly from the sample such as molecular biology. However fast and sensitive those methods are not based on single phenotypic information which presents several drawbacks and limitations. Optical methods have the advantage to allow single-cell sensitivity and to probe the phenotype of measured cells. Here we present a process and a prototype that allow automated single-bacteria phenotypic analysis. This prototype is based on the use of Digital In-line Holography techniques combined with a specially designed Raman spectrometer using a dedicated device to capture bacteria. The localization of single-cell is finely determined by using holograms and a proper propagation kernel. Holographic images are also used to analyze bacteria in the sample to sort potential pathogens from flora dwelling species or other biological particles. This accurate localization enables the use of a small confocal volume adapted to the measurement of single-cell. Along with the confocal volume adaptation, we also have modified every components of the spectrometer to optimize single-bacteria Raman measurements. This optimization allowed us to acquire informative single-cell spectra using an integration time of 0.5s only. Identification results obtained with this prototype are presented based on a 65144 Raman spectra database acquired automatically on 48 bacteria strains belonging to 8 species.