Detection of Bacteria and Antibiotic Resistance
Using a Bi-Material Microfluidic Cantilever
- Fig. 1: The microfluidic microcantilever sensor with three orthogonal signals readout. (a) Is the silicon nitride cantilever with a thin layer of gold (300 nm) deposited on the under-side to serve as a second element of the cantilever. (b) Shows the microchannel of the cantilever when filled with bacteria. Bacteria trapped by inner receptors were immobilized on the microchannel. (c) A cross-section image of an inlet of the cantilever. (d) a scanning electron microscopy image of the tip of the cantilever. (e) Resonance frequency shift as bacteria bind to the immobilized receptors (mAb or AMP). (f) Bacteria inside the cantilever absorbs infrared light, local heat is generated resulting in an expansion of the cantilever elements causing the micro-cantilever to bend corresponding to the mismatch properties of elements extension coefficients. (g) The nano-mechanical deflection of the cantilever shows the specific wavelengths where bacteria absorb infrared light. This orthogonal signal offers excellent selectivity in a complex mixture.
Hashem Etayash1,2, Faheem Khan2, Kamaljit Kaur3 and Thomas Thundat2
Recent advances in micro and nanofabrication of sensors are enabling integration of multi-modal sensor signal generation techniques into a single device. Such integration has the potential to make diagnostic tools more efficient, highly sensitive and selective. Here we describe a multi-modal technique using a microfluidic microcantilever sensor and show its potential applications for detection of bacteria and bacterial drug resistance.
An urgent need exists for the development of a real-time, portable sensor platform for detecting bacteria and their susceptibility to antibiotics with high sensitivity and selectivity . Current methods for detection are time-consuming and suffer from lack of selectivity and stability. Due to the prevalent and widespread misuse of antibiotics, resistant strains of bacteria spread quickly with growing resistance mechanisms, frightening our capacity to treat even common infections and increasing the mortality and disability rate among patients . As reported in Nature Commun.  we have recently developed a multi-modal bi-material microfluidic cantilever sensor that can detect bacteria and suggest their susceptible antibiotics in real-time with sensitivity and selectivity.
Micro-cantilever sensors have attracted substantial attention as a highly sensitive platform for chemical and biological detection due to their simplicity, sensitivity and their ability for label-free and real-time in situ monitoring [4-5]. In the last two decades, a number of versatile sensors based on micro-cantilevers have been developed for the detection of microorganisms and biomolecules such as proteins, DNA, RNA, etc. . The sensing mechanism in the micro-cantilever is based on adsorption of the target molecules on immobilized receptors on the cantilever surface which changes the mechanical properties of the cantilever. Molecular adsorption results in the cantilever bending due to adsorption-induced forces while the resonance frequency changes due to mass loading.
Selectivity in detection depends on the selectivity of the immobilized receptors. Despite the many advances in the development of cantilever sensors, multiple drawbacks exist that limit their translation to clinical applications. For example, the liquid environment surrounding the cantilever limits the accuracy of the resonance frequency measurements. The damping caused by the liquid lowers the quality factor (Q-factor) of the cantilever which decreases resolution in the frequency measurements. While in situ mass detection is severely limited by the damping of the resonance frequency, the adsorption-induced cantilever bending is independent of the presence of a liquid environment. Both readout modes; however, are still affected by the liquid flow rate, the volume of the cantilever cell and the placement of the cantilever in the cell . At present, stabilization of a cantilever response in a liquid environment can take a few hours due to drift. The delivery of the sample into the cell may also create a higher noise level during measurements .
A suspended microchannel resonator (SMR) overcomes all these limitations since the liquid is contained inside the resonator . The SMR measures the buoyant mass of objects that pass through the resonator. The system is able to weigh a single nanoparticle, a single cell and sub-monolayers of proteins with extraordinarily mass resolution [6-7]. Though the SMR overcomes the limitations of liquid damping, it still suffers from lack of specificity (selectivity) which is critical in biomedical diagnosis. To circumvent the selectivity challenge, we combined a multi-modal detection approach into the microfluidic cantilever by instantaneously monitoring orthogonal signals.
A Bi-Material Microfluidic Cantilever
The micro-cantilever is a rigid diving board fabricated from silicon nitride with a thin layer of gold (300 nm) on one side to make it a bi-material structure (fig. 1). A microfluidic channel loop embedded in the cantilever transports the liquids for mass measurements. The inner surface of the microfluidic channel is functionalized with biomolecular receptors in order to trap and pre-concentrate the bacteria to enhance the selectivity of detection . As the bacteria pass through the microfluidic channel they get trapped inside the channel by the immobilized receptors, resulting in a change in the resonance frequency of the cantilever (1st signal). In addition, adsorption of bacteria induces a surface stress, resulting in the deflection of the cantilever. The mismatch in the thermal expansion coefficients between the materials of the cantilever (silicon nitride and gold) makes the cantilever sensitive to changes in temperature due to the bi-material effect. The third orthogonal signal comes from irradiating the cantilever with infrared radiation (IR). When the captured bacteria absorb specific IR wavelength vibrational modes, specific groups are resonantly excited generating heat energy that bends the cantilever. By scanning the wavelength over a spectral region and monitoring the resultant cantilever bending it is possible to create a mechanical IR spectrum of the sample. In our approach, we tested the ability of the device to selectively detect Listeria monocytogenes using anti-listerial, monoclonal antibodies (mAb) and antimicrobial peptide (AMP) from class IIa bacteriocins immobilized on the inner interface of the microfluidic channel on the cantilever. We further tested the applicability of the sensor to identify antibiotic efficacy against bacteria .
Detection of Listeria Monocytogenes
Samples contaminated with various concentrations of Listeria monocytogenes were introduced into microfluidic cantilevers pre-functionalized with anti-listerial mAb or with AMP Leucocin A, which works against Listeria monocytogenes. When bacteria samples pass through the cantilever system, three signals are generated. First, the resonance frequency decreases dramatically. Second, the amplitude of cantilever deflection (∆h) changes significantly compared to a reference cantilever. Third the masses of captured bacteria measured from the resonance frequency shift show 24.5 and 24.9 ng for AMP and mAb-coated cantilevers, respectively as shown in figure 1. Illuminating the trapped bacteria with IR light from a tunable source deflects the cantilever further when bacteria absorb specific wavelengths of IR. Monitoring the nano-mechanical bending of the cantilever as a function of illuminating wavelength, resembles the infrared absorption spectrum of the captured bacteria as shown in figure 1. Unlike the activated cantilevers, the deflection, resonance frequency shift and the IR-induced nano-mechanical signals from the control samples do not show any measurable, well-defined signals. A direct relationship between cantilever’s response and the number of bacteria in the injected solution was observed during the injection of a serial concentration of Listeria monocytogenes. We have detected cell concentrations as low as 100 cells per 100 µL (a single cell per µL) using these sensors with a signal-to-noise ratio of 3 .
In order to determine the selectivity of the sensor towards Listeria monocytogenes, a number of strains including gram-positive and gram-negative bacteria were introduced into the sensor. The results show minimal response of the cantilevers to the other strains of bacteria in both mAb-coated cantilever and the AMP-coated cantilever compared to that observed against Listeria monocytogenes. Both cantilever deflection and resonance frequency variation exhibited substantial discernment patterns towards Listeria monocytogenes. The nano-Infrared spectra also revealed selective signature of Listeria monocytogenes compared to other tested strains. It is clear that the immobilized receptors selectively adhere to Listeria monocytogenes while allowing the others to flow through without any attachment. This selective binding permits the cantilever to generate three specific orthogonal signals .
In these experiments, we tracked the response of the bi-material microfluidic cantilevers coated internally with Escherichia coli (E. coli) while exposing to a growth medium enriched with ampicillin and/or kanamycin antibiotics. E. coli (DH5α) was chosen for this test as it has a positive response to ampicillin while repelling the bactericidal effect of kanamycin. Monitoring the cantilever bending and frequency responses for ~30 minutes after the introduction of ampicillin showed a reduction in the cantilever deflection (6-7 nm) and an increase in the resonance frequency. These signals also did not recover after flushing the channel with a freshly prepared bacterial growth medium. An enhanced deflection and a downward shift in resonance frequency were observed within ~30 minutes of exposure to kanamycin. Flushing the channel with liquid growth medium recovered the signals. These signals of the sensor are associated with bacterial response to antibiotics and indicated that ampicillin killed E. coli while kanamycin did not. Confocal microscopy imaging, where life/dead staining kit was used, confirmed the results and differentiated the live from dead bacteria when exposed to the antibiotics. Furthermore, the time-dependant fluctuation of the cantilever, which is statistically measured as a variance, was used to investigate bacterial response to the antibiotics. It appears that the fluctuation decreases dramatically with ampicillin, while it stays consistently high during kanamycin injection . Previously, Longo and colleagues suggested that changes in bacteria metabolism might be the cause of the observed fluctuations of the cantilever .
To validate this concept, we introduced a medium enriched with glucose (5%) to the E. coli after exposure to antibiotics and recorded the orthogonal signals. We found that ampicillin exposed E. coli did not respond to the extra glucose medium, while the kanamycin exposed cells responded very well. It appears that kanamycin exposed cells had indeed a full metabolic recovery, resulting in a decrease in the resonance frequency, an increase in the cantilever’s deflection and an increase in the noise of the nano-mechanical oscillations. Moreover, by performing the multivariate analysis of the infrared spectra obtained from both experiments, we defined the difference between intact and killed bacteria. As expected, the analysis showed unique infrared absorption features for the bacteria exposed to ampicillin, which differ from that exposed to kanamycin, showing a discernable difference between intact and injured bacteria .
The microfluidic cantilever sensor integrated with simultaneous multiple signal generation techniques, offers a unique approach for diagnosis with exceptional specificity and sensitivity. Unlike other traditional micromechanical sensors, the bi-material microfluidic cantilever abolishes the liquid damping effect and enables in situ monitoring, adsorption-induced cantilever deflection with high sensitivity. The cantilever can also generate photo-thermal signatures of the sample inside the microfluidic channel calorimetrically. The thermal-based nano-mechanical spectroscopy is a direct method for measuring IR absorption that offers overtone-free signals that is ideal for molecular recognition. The ability of the sensor to generate multiple orthogonal signals makes it an ideal potential tool for point-of-care applications.
1 University of Alberta, Faculty of Pharmacy and Pharmaceutical Sciences, Edmonton, Canada
2 University of Alberta, Department of Chemical and Materials Engineering, Edmonton, Canada
3 Chapman University, Chapman University School of Pharmacy (CUSP), Harry and Diane Rinker Health Science Campus, Irvine, California, United States
B.Sc. M.Sc Hashem R. Ali Etayash
Faculty of Pharmacy and Pharmaceutical Sciences
University of Alberta
Edmonton, AB, Canada
More articles on Medicine & Diagnostics: http://www.laboratory-journal.com/science/medicine-diagnostics
Antibiotic Resistance: Facing the Challenges of Bacterial Infections: http://www.laboratory-journal.com/science/pharma-drug-discovery/antibiot...
 Farahi, R. H.; Passian, A.; Tetard, L.; Thundat, T., Critical Issues in Sensor Science To Aid Food and Water Safety. ACS Nano 2012, 6 (6), 4548-4556.
 Alanis, A. J., Resistance to antibiotics: are we in the post-antibiotic era? Arch Med Res 2005, 36 (6), 697-705.
 Etayash, H.; Khan, M. F.; Kaur, K.; Thundat, T., Microfluidic cantilever detects bacteria and measures their susceptibility to antibiotics in small confined volumes. Nature Communications 2016, 7, 12947.
 Etayash, H.; Thundat, T., Microcantilever Chemical and Biological Sensors. In Encyclopedia of Nanotechnology, Bhushan, B., Ed. Springer Netherlands: 2015; pp 1-9.
 Arlett, J. L.; Myers, E. B.; Roukes, M. L., Comparative advantages of mechanical biosensors. Nat Nano 2011, 6 (4), 203-215.
 Burg, T. P.; Godin, M.; Knudsen, S. M.; Shen, W.; Carlson, G.; Foster, J. S.; Babcock, K.; Manalis, S. R., Weighing of biomolecules, single cells and single nanoparticles in fluid. Nature 2007, 446 (7139), 1066-1069.
 Godin, M.; Delgado, F. F.; Son, S.; Grover, W. H.; Bryan, A. K.; Tzur, A.; Jorgensen, P.; Payer, K.; Grossman, A. D.; Kirschner, M. W.; Manalis, S. R., Using buoyant mass to measure the growth of single cells. Nat Meth 2010, 7 (5), 387-390.
 LongoG; Alonso Sarduy, L.; Rio, L. M.; BizziniA; TrampuzA; NotzJ; DietlerG; KasasS, Rapid detection of bacterial resistance to antibiotics using AFM cantilevers as nanomechanical sensors. Nat Nano 2013, 8 (7), 522-526.