Nanoparticle sample preparation and mass spectrometry for rapid diagnosis of microbial infections

In vitro diagnostics encompasses a wide range of medical devices and assays, which aim to provide reliable and accurate diagnosis of disease. This can be achieved by detecting a target, for example, a protein biomarker or a pathogen bacterium, and/or host factors such as cytokines induced in an inflammatory response. Detection involves an assay to capture target molecules and distinguish them from other substances in an ex vivo sample matrix. Selective capturing can be achieved using affinity probes, such as antibodies or small molecules, often coupled to a label, for example, an enzyme or a particle, to facilitate detection in complex matrixes (Figure 1). Today, the combination of nanoparticle approaches for sample preparation/concentration, with high information content, rapid analysis by mass spectrometry, is changing the way we detect and identify pathogenic bacteria in the diagnosis of microbial infection.

Figure 1. Standard configurations for a nanoparticle sandwich immunoassay. The capture (black) and reporter (red) antibodies in the example bind on different sites of the target forming the sandwich. The capture probe is attached to a solid phase surface in panel (a) and a magnetic particle in panel (b). The reporter probe can be linked to an enzyme (e.g. enzyme-linked immunosorbent assay) or another particle (e.g. cluster assay). (c) Schematic of a cluster immunoassay where both the solid phase and the label are nanoparticles (either magnetic or fluorescent).

Most diagnostic assays are still carried out in centralised laboratories with equipment operated by dedicated laboratory personnel. Decentralised and rapid testing is important for time-critical diagnoses, where time affects morbidity and mortality outcomes for the patient, particularly in the case of bacteria sepsis. The trend towards decentralisation poses demanding constraints on assay technologies. Target molecules are present in body fluids in minute concentrations. To achieve high sensitivity detection, the physics, biology and chemistry of the sensing device, or biosensor, needs to be well integrated and controlled to transduce the presence of specific analytes into a measurable signal. This challenge is even harder when trying to detect bacteria, due to the incredible diversity of the genotype and phenotypes between species and within species. Detection of most pathogens is still carried out by culturing a clinical sample, a process susceptible to contamination, with a long lead time to diagnosis.

Matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI-TOF MS) has been adopted as a fast and reliable method for bacterial detection and identification, based on the characteristic mass spectrometry profiles for each bacterial species¹. Databases have been developed covering commonly encountered pathogenic microorganisms for use in routine bacterial identification with results in <2 h. MALDI-TOF MS can also be used to identify microorganisms directly from clinical specimens, especially urine and blood^2,3. Recently, a combination of a nanoparticle immunoassay with MALDI-TOF has been used to identify dengue and influenza viruses^4,5. Agrawal recently reported on a wide range of particle-based nanotechnologies for rapid detection of protein biomarkers in food and beverages⁶. This review provides an overview on recent developments in nanotechnology that address key challenges in modern microbiology, with an emphasis on nanoparticle-based diagnostics and mass-spectrometry profiling of bacteria.

Magnetic nanoparticles can be decorated with a high density of capture probes due to their extremely high surface-to-volume ratio⁷. Additionally, they can be externally manipulated by controlled magnetic fields in a background of a complex biological matrix. The bioactive coating on the surface captures and magnetically ‘tags’ the target, enabling magnetic enrichment⁸, washing and re-suspension into a clean matrix or buffer to facilitate detection. Sensitive particle-based diagnostic nanotechnologies require efficient suppression of the background arising from non-specific interactions originating from the sample matrix. As a consequence, a great deal of research has focused on the development of robust surface architectures to extend assay applicability to complex matrixes. Hydrophilic polymer coatings (e.g. polyethylene glycol) are extensively exploited to reduce biofouling and to space the bioactive probes from the surface of the particle, which ameliorates risk of steric hindrance for analyte capture^9–11. Once tagged, the target can be detected magnetically or optically (Table 1). Magnetic detection encompasses the measurement of stray magnetic fields induced into the particles by an external field, for example, by means of giant magneto-resistance (GMR) and diagnostic magnetic resonance (DMR)^12,13. The assay architecture is similar in most of the recent literature and is based on several steps^14–17. The capture probe is immobilised on the sensor surface and binds the target. Subsequently, a biotinylated reporter probe binds the sequestered targets. Finally, streptavidin-coated magnetic particles label the captured targets and enable detection. With this approach, Osterfeld et al. demonstrated multiplexed detection of bacterial protein biomarkers in buffer at a concentration of 1 pg/mL and Xu et al. reported detection of synthetic DNA at a concentration of 10 pM^14,16. By combining magnetic tagging and detection with PCR amplification, Koets et al. demonstrated detection of 4–250 pM of genomic double-stranded bacterial DNA¹⁸. With a similar approach, Escherichia coli strains have been detected with a lower limit of detection (LoD) of 10–1000 CFU/mL¹⁷.

Table 1. Current state-of-the-art in nanotechnologies based on magnetic nanoparticles, biochemical-based test and mass spectrometry for the detection of biomarkers, bacteria and activity of antibiotic-resistant enzymes.

Diagnostic magnetic resonance takes advantage of the change in opto-magnetic properties of the particles upon target-induced aggregation. In particular, clusters of magnetic particles have increased efficiency to dephase the spins of surrounding water molecules, thus inducing a shift in the spin–spin relaxation time (T2), which is measurable using nuclear magnetic resonance (NMR). A major advantage of this technology is the ability to work with turbid or opaque matrixes such as whole blood. This approach has been widely exploited by T2 Biosystems to sensitively capture protein biomarkers, nucleic acids and bacteria¹⁹. By combining PCR amplification with detection using NMR, they recently demonstrated sensitive detection (0.5 pM of amplified DNA) and phenotyping of bacteria within 2 h²⁰.

Magnetic particles can also be combined with fluorescent reporter probes (e.g. quantum dots, particles or enzymes) with measurement of scattering and fluorescence²¹. Scattering of light occurs when electromagnetic radiation encounters matter: the radiation couples to the electronic cloud in the particle, which results in an induced electric dipole. The oscillating induced dipole moment is manifested as a source of electromagnetic radiation, thereby resulting in scattered light. The phenomenon has been exploited to demonstrate solution-based detection of 1 pM of ovalbumin in buffer by turbidimetry in 3 minutes²²: clusters of particles scatter more light than single particles, therefore target-induced aggregation results in a measurable decrease in transmitted light. The strong background arising from single particles limits the sensitivity of the approach²³. Magnetically controlled rotation of linear clusters overcomes this limitation: rod-shaped clusters exhibit a time-dependent cross-section during one period of rotation, thus modulating the scattered light intensity²⁴. Single spherical particles expose the same cross section to an incoming light beam, suppressing the amplitude of the modulation. With this approach, 0.5 pM of prostate specific antigen (PSA) in undiluted plasma could be detected in less than 15 minutes in one step²⁵. Bruls et al. report detection of 1 pM of cardiac troponin and biomarkers of illicit drugs in less than 5 minutes²⁶. The target is rapidly captured by magnetic nanoparticles and magnetically focussed on the sensor surface where secondary antibodies against multiple targets are immobilised. The unbound particles are magnetically removed and the light scattered by the bound magnetic particles is measured using frustrated total internal reflection. Optical scattering from magnetic particles was recently shown to improve the sensitivity in detecting and sorting bacteria. Sorting bacteria using flow cytometry enables sensitive and selective analysis and sorting of subpopulations. However, it has always been a difficult task due to the small size of bacteria, relative to mammalian cells. For such small objects, a light scatter signal is on the threshold detection level. The use of magnetic particles allows for enrichment of the bacteria from a large sample volume and consequent signal amplification²⁷.

Several research groups have exploited magnetic labelling for enriching or extracting targets from a complex matrix followed by fluorescent detection aided by an enzyme or another fluorescent label (e.g. fluorescent nanoparticle or quantum dot). Amplification-free detection of 60 pM of nucleic acids has been demonstrated in a total assay time of 45 minutes in a three-step format¹⁰. Magnetic nanoparticles sequester the target sequence and, after a washing step, fluorescent nanoparticles decorated with the complementary sequence bind to the target. The final washing step removes unbound fluorescent labels and enables detection of the fluorescent intensity. With a similar approach, Wang et al. report magnetic tagging and chemiluminescent detection of a cancer biomarker in serum at a concentration as low as 4 pg/mL in less than 10 minutes²⁸. Gu and co-workers developed vancomycin conjugated magnetic nanoparticles to capture several Gram-positive pathogens (e.g. vancomycin-resistant enterococci)^29,30. In a further refinement, fluorescent vancomycin stained the enriched bacteria for the detection of 10 CFU/mL from blood samples within 2 hours using fluorescence microscopy³¹.

Mass spectrometry has been traditionally used for identification of bacteria and yeast¹, and also recently with viruses. Notably, virus detection is usually used in combination with an immunocapture approach using nanoparticles such as those reviewed above⁵. For bacterial identification, MALDI-TOF can reduce the turnaround time to up to 1 day earlier to obtain accurate diagnostic results with cost savings of ~50% against culture-based and PCR-based protocols³². MALDI-TOF is capable of providing fragmentation data at a rate of less than 1 second per sample. Samples can also be processed in batch mode; for example, 192 samples in one batch for the Vitek MS system in 1 hour, with a further hour required for analysis and identification (Table 1).

The current approach to determine the bacterial susceptibility to antibiotics in a clinical diagnostic setting is by using the disk susceptibility test (DST) and minimal inhibition concentration using Vitek (Biomerieux). Recently, matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS) has been explored as a fast and reliable method for bacterial identification, based on the characteristic protein profiles for each bacterial species^33,34. Databases have been developed that include commonly encountered pathogenic microorganisms, thus allowing the use of this method in routine bacterial identification in <2 hours. However, the antibiotic susceptibility patterns of the causative pathogens, which are essential to ensure appropriate and accurate antibiotic prescription, are usually only available 24–48 hours after collection of a specimen. Several recent studies explored the use of mass spectrometry in identifying carbapenemases and other β-lactamases by measuring the degradation of the antibiotic, with results produced in 4 hours^35,36. However, this approach does not allow for identification of the specific carbapenemase involved, because it only measures the hydrolysis of carbapenems, which occurs with any carbapenemases^37–39.

Sample preparation is crucial in obtaining the best spectra of microorganisms using MALDI-TOF. The use of nanoparticle-based techniques could give the most favourable detection limit by extracting the microorganisms from complex materials, such as blood. We foresee integrated workflows with sample preparation using a combination of nanoparticle-based systems coupled to MALDI-TOF to allow simple, rapid and high-throughput application in IVD.