Field of Invention

The present invention generally relates to a method and system of detecting a target biological entity, for example within a biofluid, by performing electrical measurements with a sensor in the presence of the entity-containing biofluid.

Background

In the field of diagnostics in the medical industry it is often desirable to be able to detect a biological target entity, which may be symptomatic of some underlying condition, efficiently and using a non-invasive procedure. Although an invasive procedure may typically be more reliable in providing concrete evidence for an underlying condition, it may be costly, inconvenient for a patient, and time consuming. Invasive procedures may include such methods as a tissue biopsy. Therefore, this method is unlikely to be the most efficient method of detection. Moreover, it would typically be necessary for a highly trained medical professional to carry out such an invasive procedure.

A more preferred method of detection of some target biological entity would be a so called point-of-care procedure. Point-of-care testing is known in the medical and diagnostic industry, and has the advantages of being quicker, non-invasive and not confined to medical laboratory. Such a procedure may take the form of a liquid biopsy. In other words, a bodily fluid may be quickly and comfortably extracted from a patient. For example, the fluid may be a blood sample. The necessary tests would then ideally be performed upon the biopsy fluid in a portable device able to receive the fluid. Further advantages of this method include that it is cheap, has a quick turnaround for receiving a test result, and may not require the supervision of a highly trained medical professional to the same extent as an invasive procedure.

The target biological entity to be detected in the point-of-care procedure may typically be a biological entity or other molecule which is symptomatic of an underlying condition or state. Therefore, it is important that the device should be both highly selective in detecting only the relevant target entity, and produce a strong and reliable signal which is indicative of only said target entity. In other words, the device should be able to produce a high signal to noise ratio which reproducibly indicates the presence of the target entity. The combination of these two desirable features should therefore produce a device which is able to detect the presence of the target entity, even when it is present in only trace amounts.

Microfluidic devices are well known in the diagnostic and sensing industry, and may comprise modules or be integrated with various materials, forming “lab-on-a-chip” devices. These devices have the advantage that they may be mass-produced, and may be utilised as a highly portable and conveniently-sized sensing devices. Furthermore, due to their small or even micro-scale dimensions, the amount of fluid which such a sensing device needs to function is very small, which again improves the convenience of a liquid biopsy procedure.

One important area in the medical diagnostic field is that of early cancer detection, screening, prognosis and therapy follow-up. Current methods of diagnosis may involve the detection of nucleic acids indicative of a particular cancer. Such methods use amplification and/or sequencing technologies which effectively create a profile of an entire DNA or RNA sequence. However, these methods require large reagent consumption, are not suitable for point-of-care integration and are likely to be time- consuming and leave significant environmental footprints. A notable alternative to the time-consuming sequencing and/or amplification technologies is a method which, instead, targets the proteins which are the mediators responsible for regulating the relevant activities for cancer cell proliferation. Such protein mediators may be extracted via a liquid biopsy without the need for an invasive procedure.

Hence, there remains a need in the state-of-the-art to provide a method having the ability to selectively target and detect the presence of such protein mediators or entities, in a convenient and reliable point-of-care device.

Summary

Aspects and preferred features are set out in the accompanying claims. According to one aspect there is provided a method of detecting a target biological entity in a biofluid. The method may comprise introducing the biofluid to a suspension to provide a precursor mixture. The biofluid may comprise a plurality of target biological entities. The suspension may comprise a plurality of nanoparticles, wherein each of the plurality of nanoparticles is functionalized so that it is able to bind with the target biological entity to produce a bound nanoparticle-entity assembly. The method may further comprise treating the precursor mixture to separate the bound nanoparticle-entity assemblies from nanoparticles not comprised in one of the bound nanoparticle-entity assemblies to provide a treated precursor mixture.

The method may further comprise characterizing the treated precursor mixture with a sensor. The sensor may comprise a substrate bearing a pair of electrodes separated by a lateral distance of less than 100 nm, and may be separated by lateral distance of less than 20 nm, 10nm or 5nm. A region between the electrodes may define a sensing region.

The characterizing may comprise applying an electric field to the treated precursor mixture to concentrate the bound nanoparticle-entity assemblies in the sensing region. The characterizing may further comprise applying a nanoparticle sensing voltage between the electrodes. The characterizing may further comprise characterizing a response of the sensing region to the nanoparticle sensing voltage to determine treated precursor mixture characterizing data. The method may then further comprise detecting the presence of the target biological entity from the treated precursor mixture characterizing data.

Advantageously, by functionalising the nanoparticles themselves, the transport of the target entities and subsequent concentration around the sensing region may be efficiently facilitated. The nanoparticles may have a high affinity with the electrodes themselves, for example, if both the electrodes and nanoparticles are made of the same material, e.g. gold. Additionally, the use of nanoparticles allows an electric field to be applied which may more easily attract the nanoparticles bound to the entities, and concentrate them around the sensing electrode. Thus, there is provided an efficient method of sequestering and concentrating target entities, which may possess an advantageous sensitivity to detecting the presence of target entities. In an alternative approach, the‘naked’ nanoparticles not yet bound to a target entity may be sequestered or selectively removed from the mixture, such that only bound nanoparticles-entity assemblies remain. Here‘naked’ is used to refer to a nanoparticle which is functionalized but not bound to a target. The suspension may be a liquid suspension containing heterogeneous species, including the nanoparticles. Furthermore, the nanoparticles may be conductive nanoparticles, for example, metallic or semi-metallic nanoparticles.

The concentrating of the bound nanoparticle-entity assemblies in the sensing region may occur by dielectrophoresis (DEP). That is, a dielectrophoretic force may act on the nanoparticles, which may be conductive nanoparticles, where the force acts on the nanoparticles in the direction of the sensing region. The method may further, or alternatively, comprise: measuring a current flowing between the electrodes due to the nanoparticle sensing voltage; detecting a current between the electrodes, the current indicating potential binding of a bound nanoparticle-entity assembly between the electrodes; and differentiating, based on the current between the electrodes, between a binding event defined by at least one bound nanoparticle-entity assembly becoming bound in the sensing region of the electrodes, and a non-binding event, to detect the target biological entity.

The treatment and corresponding separation may comprise two separate streams: one designed to transfer the un-bound nanoparticles, and one to separate and transfer away the bound nanoparticle-entity assemblies. The treatment step may alternatively comprise a mechanical filter designed to separate the assemblies based on relative size of all particles in the precursor mixture. The treatment may further comprise separating target biological entities not comprised in one of the bound nanoparticle- entity assemblies.

The method may further comprise applying the electric field to the treated precursor mixture, comprising applying an AC voltage to a pair of treatment electrodes. These electrodes may be external electrodes in addition to the sensor electrode pair. For example, these may be optimally positioned to induce concentration of the nanoparticle-entity assemblies around the sensor. As further example, a frequency of around or less than 1 - 10 MHz may be used, driven by an RMS voltage of < 1.5 V. This alternating electric field induces the DEP and attractive force between the source of the electric field (sensing region) and the nanoparticle-entity assembly.

Furthermore, the pair of treatment electrodes may be the pair of electrodes on the substrate separated by a lateral distance of less than 100 nm. Advantageously, using the existing electrodes allows for a simplified device structure, where the pair of electrodes serves the purpose both of attracting the nanoparticles, and sensing the nanoparticles. Further advantageously, the potency of an attractive force may be increased by using the existing sensor electrodes. That is, an increasingly narrow lateral separation of electrodes may increase the dielectrophoretic (DEP) force without the need to increase a voltage to power the applied electric field.

Any of the above methods may comprise characterizing a response of the sensing region to the nanoparticle sensing voltage to determine treated precursor mixture characterizing data comprises identifying whether, after concentrating the bound nanoparticle-entity assemblies in the sensing region, the sensing region exhibits ohmic behaviour. Ohmic behaviour may be classified as a behaviour exhibited by a typical resistor. For example, ohmic behaviour may be probed/determined by measuring either resistance by a DC current, or impedance by AC. For example, this step may comprise identifying a current/voltage profile characteristic of a resistor, or in the case of impedance, a characteristic Nyquist Plot.

The substrate may bear a group of at least 10 or 100 pairs of electrodes each with electrodes separated by a lateral distance of less than 100 nm, to define a group of the sensing regions. This can help to provide an improved signal-to-noise ratio. Thus, applying the electric field to the treated precursor mixture may concentrate the bound nanoparticle-entity assemblies between or around each pair of electrodes, and more broadly the group of sensing regions, wherein the characterizing comprises characterizing the response of each pair of electrodes of the group of sensing regions and combining the responses to determine the treated precursor mixture characterizing data.

A conductivity or permittivity of the plurality of nanoparticles may be higher relative to a conductivity or permittivity of the precursor mixture. Metallic nanoparticles (such as gold, for example) may be used which have a higher conductivity than the surrounding medium of the precursor mixture. Such an increased relative difference in conductivity may advantageously increase the concentrating of the bound nanoparticle-entity assemblies in the sensing region.

Applying the electric field may induce an attractive force acting between the sensing region of the electrodes and the nanoparticle. The particles may be metallic/conductive in this scenario. Nanoparticles bound to a target to form a nanoparticle-entity assembly may also be subjected to the same force, and thus the bound assembly is attracted to the sensing region. The force may be a dielectrophoretic (DEP) force. Alternatively, dielectric nanoparticles may be used such that a repulsive force acts on the nanoparticles in the presence of an alternating electric field. This repulsive force may be used to drive the nanoparticles, or preferably nanoparticle-entity assemblies, to the sensing region in special circumstances.

The nanoparticle sensing voltage may be a constant voltage which induces a direct current between the electrodes. Advantageously, when an increasingly small nanogap is used, an increasingly small corresponding sensing voltage need be applied in order to identify the presence of target entities, for example, identifying ohmic behaviour in the sensing region. Alternatively, AC may be used as a sensing voltage, which further may be a continuation of the AC voltage used to apply the electric field used to concentrate the nanoparticle-entity assemblies. In other words, the same electrodes, and same AC voltage, may be used to create both the electric field and to characterize a response of the sensing region determine treated precursor mixture characterizing data

Each of the plurality of nanoparticles may be functionalized with a binding element which provides a capability to bind with the target biological entity to produce the bound nanoparticle-entity assembly. For example, the binding element may be an aptamer/antibody which favourably binds to the biological target entity.

In some implementations each nanoparticle has less than on average (e.g. median number) 5, 3, or 2 aptamers or antibodies attached, or on average just a single aptamer or antibody attached. This allows just one or a few target entities e.g. extracellular vesicles such as exosomes, to be bound to each nanoparticle. This can be useful as the DEP forces on the target entity and nanoparticle may be of opposite sign (depending upon the frequency e.g. outside 1-10MHz), and in general can increase controllability of the process. Also, the target entities, e.g. exosomes or proteins, tend to have a low electrical conductivity and it in implementations the electrodes should be bridges mainly by the nanoparticles rather than the target entities, to improve the signal-to-noise ratio.

Known techniques may be used to ensure that just one or a few aptamers or antibodies are attached to each nanoparticle. For example they may be attached by combining the nanoparticles and aptamers/antibodies in a solution/suspension (allowing physical attachment to the nanoparticles), and then nanoparticles with the target number of aptamers/antibodies attached may be separated by centrifugation.

A surface of each of the electrodes adjacent to the sensing region may be functionalized by providing the surface with linker molecules which enhance the binding the bound nanoparticle-entity assembly between the electrodes. These linker molecules may comprise a thiol linker. In an example where the nanoparticles are gold, thiol linkers may be used such that a high affinity is achieved between the sensing region and the nanoparticles, due to a relative strength of a thiol-gold interaction.

Linker molecules such as molecules which comprise a thiol linker are optional. If present they may be provided on either or both of a surface of the electrodes and a surface of the nanoparticles. In the latter case a nanoparticle e.g. a gold nanoparticle, may have a majority of its surface e.g. 90% covered by the linker molecules and the remainder e.g. 10% covered by aptamers/antibodies.

An average maximum dimension of the nanoparticles may greater than about 20 nm and/or less than about 100 nm. Furthermore, a standard deviation may be small such that the plurality of nanoparticles is monodisperse, i.e. having relatively uniform maximum dimensions.

The plurality of nanoparticles may comprise gold nanoparticles. Other inert metals with similar conducting properties may also be used, which would readily occur to the skilled person. The treating of the precursor mixture may comprise selective separation of the bound nanoparticle-entity assemblies from the nanoparticles not comprised in one of the bound nanoparticle-entity assemblies according to a property of any of: density, size, permittivity and conductivity.

Treating according to density in the precursor mixture may comprise centrifugation of the precursor mixture.

Treating the precursor mixture according to size may comprise passing the precursor mixture through a microfluidic array comprising a mechanical filter. For example, the mechanical filter may take advantage of laminar flow properties in order to separate particles by size, for example, in order to isolate the larger nanoparticle-entity assemblies. In other approaches, a physical mechanical filter may be used such as a nanoscale mesh or molecular sieve which filters components of the precursor mixture by size.

Treating according to permittivity and/or conductivity in the precursor mixture may comprise applying an alternating electric field to the precursor mixture. That is, an additional Dielectrophoresis (DEP) step may be performed, prior to concentrating the assemblies around the sensing region, in order to separate the nanoparticle-entity assemblies. In this way, the size of the components of the mixture may play a factor: that is, diffusion and mass transport properties of particles caused by a DEP force may be dependent on particles size, and viscosity of the medium.

In all above systems and methods, the target biological entity may be an extracellular vesicle e.g. an exosome. The target biological entity may be a protein.

According to a further aspect, there is provided a microfluidic system for detecting a target biological entity in a biofluid. The system comprises: a first microfluidic input to receive a biofluid, wherein the biofluid comprises a plurality of target biological entities; a second microfluidic input to receive a suspension of nanoparticles, wherein each of the plurality of nanoparticles is functionalized so that it is able to bind with the target biological entity to produce a bound nanoparticle-entity assembly; a mixing chamber or channel to mix the biofluid with the suspension of nanoparticles to form a precursor mixture. The system further comprises a precursor mixture treatment chamber or channel configured to treat the precursor mixture to separate the bound nanoparticle-entity assemblies from nanoparticles not comprised in one of the bound nanoparticle-entity assemblies to provide a treated precursor mixture; a precursor characterization chamber or channel, which may be the same chamber/channel as the treatment chamber or channel, to characterize the treated precursor mixture, the precursor characterization chamber or channel comprising a sensor with a substrate bearing a pair of electrodes separated by a lateral distance of less than 100 nm, wherein a region between the electrodes defines a sensing region.

The system further comprises a treated precursor mixture characterization system configured to: apply an electric field to the treated precursor mixture to concentrate the bound nanoparticle-entity assemblies in the sensing region, apply a nanoparticle sensing voltage between the electrodes; characterize a response of the sensing region to the nanoparticle sensing voltage to determine treated precursor mixture characterizing data. The system further comprises a detector configured to use the treated precursor mixture characterizing data to detect to the presence of the target biological entity in the biofluid.

The biofluid may be provided in a solution e.g. it may flow in the solution over the sensing region. The solution may have an ionic strength of at least 1 % (0.01x) PBS or 0.1x PBS (phosphate-buffered saline); the solution may have an ionic strength of less than 10%, 5%, 2% or 1 % PBS. If proteins e.g. antibodies, or extracellular vesicle(s) e.g. exosomes, are bound on the nanoparticle or electrode surface, they can become more stable and more resistant to change in pH/ionic strength compared to when they are (alone) in the solution. This can be advantages in facilitating use of a low ionic strength solution e.g. 0.1x or 0.01x PBS, which can facilitate DEP control.

According to another aspect, there is provided a method of detecting a target biological entity in a biofluid, comprising flowing the biofluid over a sensor, the sensor comprising a substrate bearing a pair of electrodes separated by a lateral distance of less than 200 nm. A region between the electrodes defines a sensing region, and a surface of each of the electrodes adjacent to the sensing region is functionalized by providing the surface with target recognition molecules to bind the target biological entity. The method further comprises: applying a voltage between the electrodes; measuring a current flowing between the electrodes due to the applied voltage; detecting a change in the current flowing between the electrodes as the biofluid is flowing over the sensor, indicating potential binding of the target biological entity to one or more of the target recognition molecules; and differentiating between a binding event in which the target biological entity is bound in the sensing region by one or more of the target recognition molecules and binding/adsorption of a non-target substance between the electrodes using the detected change in current, to detect the target biological entity.

In a related approach, the surface of each of the electrodes in the sensing region is functionalized by providing the surface with target recognition molecules to bind the target biological entity.

The flowing of the biofluid over a sensor may comprise flowing the biofluid in a solution over the sensor. Furthermore, the solution may have an ionic strength of at least 0.01x PBS or 0.1 x PBS and the lateral distance between the pair of electrodes may be less than twice the Debye length of the solution. In some implementations, the lateral distance between the pair of electrodes may need to be sufficiently small in order that the Debye length of the solution overlaps. Furthermore, this feature is equivalent to requiring that the electrical double layer produced over the electrodes overlaps. The width of electrical double layer may be a function of at least the ionic strength of the fluid. In implementations, a gap between the electrode pair may need to be less than ~ 20 nm in order for the Debye length to be equal or less than half the lateral distance between the pair of electrodes.

The target recognition molecules may be aptamers, antibodies, fragments of antibodies, single-chain variable fragments or nanobodies, which are chosen so as to selectively bind to the target biological entity. Hence, the detection of the target biological entity is been achieved using the detected change in current. The measuring of the change in current may be in respect of either: an absolute change in the value of current measured, a proportional increase in the magnitude of the current relative to the background current measured when no target molecule is bound between the electrodes, or wherein a current is measured to be above an absolute predetermined threshold current. The lateral distance between the pair of electrodes may be less than 200 nm in implementations. The measuring of the current and/or detecting the change in the current may be performed either directly or indirectly and may comprise measuring one or more of resistance, conductance, impedance (which may be a complex impedance, that is with real and imaginary components), or capacitance between the electrodes; optionally one or more as a function of frequency.

Detecting the change in the current flowing between the electrodes may comprise detecting an increase in the current, and wherein the differentiating using the detected change in current comprises differentiating using a magnitude of the increase in current.

The differentiating may comprise detecting a change in the current which is greater than a detection threshold.

The method may further comprise detecting a sequence of changes in current to detect a sequence of the binding events, wherein the sequence of changes in current may comprise a sequence of stepwise increases in the current.

The method may further comprise adapting the voltage applied between the electrodes and/or adapting detection threshold, dependent upon an isoelectric point (pi) of the target biological entity. This may correspond to an implementation where the target biological entity is a protein.

The method may further comprise adapting the voltage applied between the electrodes and/or, adapting the detection threshold, dependent upon a predicted orientation of the target biological entity when the target biological entity is bound to one or more of the target recognition molecules.

The method may further comprise modifying the target biological entity to increase the detected change in current when the target biological entity is bound by attaching one or more further entities to the target biological entity, wherein the one or more further entities comprise entities which specifically bind to the target biological entity. The method may further comprise adapting the lateral distance between the electrodes dependent upon a size of the target biological entity, and/or adapting a shape or configuration of the electrodes adjacent to the sensing region of the electrodes dependent upon a predicted likelihood of capture of the target biological entity.

The method may further comprise differentiating between the binding event in which the target biological entity is bound in the sensing region by one or more of the target recognition molecules and a binding event in which a non-target entity is bound in the sensing region.

According to another aspect, there is provided a system for detecting a target biological entity in a biofluid, comprising a sensor comprising a substrate bearing a pair of electrodes separated by a lateral distance of less than 100 nm. A region between the electrodes defines a sensing region, and a surface of each of the electrodes adjacent to the sensing region is functionalized by providing the surface with target recognition molecules to bind the target biological entity. The system further comprises: at least one channel for providing the biofluid to the sensor; a sensing system configured to apply a voltage between the electrodes to cause a current to flow between the electrodes due to the applied voltage, wherein the sensing system is configured to measure a signal dependent upon the current flow; and a detection system configured to detect a change in the current flowing between the electrodes as the biofluid is flowing over the sensor, indicating potential binding of the target biological entity to one or more of the target recognition molecules. The detection system is further configured to differentiate between a binding event in which the target biological entity is bound in the sensing region by one or more of the target recognition molecules and adsorption of a non-target substance between the electrodes using the detected change in current, to detect the target biological entity.

The system may comprise a plurality of sensors, each sensor comprising a substrate bearing a pair of electrodes. The system may also comprise a microfluidics chip or microfluidic system bearing the plurality of sensors and at least one channel for providing the biofluid to the plurality of sensors.

According to any of the above systems or methods, except when the target biological entity is an exosome, the target biological entity comprises a protein. Brief Description of the Figures

These and other aspects of the invention will now be further described, by way of example only, with reference to the accompanying figures in which:

Figure 1 shows the profile view of the electrode pair with a bound target biological entity;

Figure 2 shows a model microfluidics system and a general flowchart of a method for the detection of biological entities;

Figures 3a and 3b show, respectively, a plan view image of an array of electrode pairs, and a plan view image tilted 60 degrees of an array of electrode pairs;

Figures 4a and 4b show, respectively, a plan view image of a single electrode pair, and a plan view image of an electrode pair in operation, comprising a group of bound exosomes coupled to magnetic nanoparticles for visualization by scanning electron microscopy;

Figure 5 shows a current vs. voltage graph, showing signals derived from aggregate bindings of exosomes (indicative of Figure 4b) and non-binding events;

Figure 6 shows a variety of potential binding/adsorption scenarios for target entities and non-target entities;

Figure 7 shows a representative current-time graph illustrating binding/adsorption scenarios;

Figure 8 illustrates a profile cross-sectional view of an electrode pair with a nano-gap, and an electrical double layer in the surrounding fluid;

Figure 9 illustrates a method of using Dielectrophoresis (DEP) and current measurements to capture and detect a target molecule using gold nanoparticles; Figures 10a and 10b show, respectively, a plan view scanning electron microscope (SEM) image of a single electrode pair with nano-gap, having single and aggregate gold nano particles adsorbed onto the electrodes, and a plan view scanning electron microscope (SEM) image of an array of five electrode pairs with a nano-gap; and

Figures 11a-11c show, respectively, a plan view scanning electron microscope (SEM) image of a single electrode pair with nano-gap, having gold nano particles bridging the nano-gap between the electrodes, a higher resolution image of the region around the nanogap in Figure 11a, and voltage current plots of a response of the bridged nanogap in 11a and 11b (left-hand side) and a response of a non-bridged nanogap.

Detailed Description

There are now described methods and systems for selectively detecting the presence of a target biological entity within a fluid medium.

Figure 1 shows a system in which a biological entity 100 has been bound by target recognition molecules 106 between the pair of electrodes. The initial electrodes 110, 112 represent metallic or otherwise conductive materials which are disposed onto an insulating platform 114 which may comprise Si0 ₂ . The lower layer 116 underneath the upper platform may comprise Si. The target recognition molecules may 106 be aptamers, which have been specifically chosen to only bind to the target biological entity 100. The target biological entity 100 may be a protein, or more specifically a vesicle such as an exosome 404.

The distance 104 between the two electrodes 110 is 200 nm or less, or preferably 100 nm or less. However, it is possible that a much smaller distance may be achieved between the electrodes. In certain implementations, the distance between the electrodes 110 may be only 5 nm. Over the electrodes, a further conducting layer 108 is disposed, which may be made of gold. The fabrication process of disposing the conductive layer over the electrodes involves a selective growth step where the distance between the resulting gold layers may be accurately controlled (Serdio et al., Nanoscale, 2012, 4, 7161). This particular technique also provides a smooth surface, which facilitates functionalizing the surface with target recognition molecules, and which can also beneficially reduce the surface potential.

Between the electrodes 110 there may be provided a region which defines a sensing region. An example of this sensing region can be seen in figure 1 , where the sensing region is characterised as the part having the target recognition molecules 106 disposed onto it. These target recognition molecules 106 serve to functionalize this sensing region of the electrodes 110, and as such are designed or chosen to selectively bind to the target biological entity 100. The molecules 106 are named as such because they are designed to specifically recognise the target entity. The distance 102 between the sensing regions of the two electrodes 110 is generally smaller than the lateral distance 104 between the initial electrodes. However, it will be understood by the skilled reader that the gold layer 108 does not necessarily protrude over the initial electrodes 110, 112, and therefore lateral distance 102 may be equivalent to or smaller than lateral distance 104. This distance 102 is nevertheless defined by the distance between the tips of the disposed conducing layer 108, which may be as little as 5 nm, but possibly much greater, up to 100 nm or even 200 nm.

The device is designed to detect a target biological entity 100 in a biofluid by measuring or detecting a change in current. Device 120 is at least able to apply a voltage or potential difference across the two electrodes 110 via the circuit 118. The device 120 is further able to at least measure a current, as a result of the applied voltage applied to the electrodes 110. Some implementations are configured to measure the electron transport flowing through the target entity 100 due to the applied voltage. The voltage applied will generally be between -2V and 2 V. A different voltage may be predetermined based on characteristics (such as conductivity, permittivity, size, charge, or isoelectric point) of a specific target entity. It will be understood by the skilled person that electron transport refers to the absolute flow of current through an electrically conductive medium such as the target entity trapped within the pair of electrodes described. Electron transfer, by contrast, can be defined as the phenomenon when a charged entity merely transfers or deposits its charge onto a medium, material or entity.

It will be appreciated by the skilled person that depending upon the implementation device 120 (or characterization of the sensing region) may not be limited to applying a constant steady voltage. Thus in some implementations, the device is able to apply a time-dependent potential difference resulting in either a variable direct current or even an alternating current. Implementations of the device are not limited to measuring merely a direct (or alternating) current. Thus in other implementations, which are still aimed towards detecting a target entity 100, the device 120 is able to measure any or a combination of: resistance, conductance, impedance, or even capacitance. For example impedance spectroscopy may be performed in order to characterize the sensing region e.g. distinguish a target molecule. This list is not exhaustive.

A method may be provided for differentiating between a binding event 604 in which the target biological entity 100 is bound in the sensing region by one or more of the target recognition molecules 106, and the binding event 606 or adsorption event 610 of a non-target substance between the electrodes. Hence, in one implementation the detection of the target biological entity 100 may be achieved by measuring a detected change in current. The measuring of the change in current may be in respect of either: a predetermined absolute change in the value of current measured; a predetermined proportional increase or decrease in the magnitude of the current relative to the background current measured when no target molecule is bound between the electrodes; or, wherein the current increases or decreases and is measured to be above, between or below an absolute predetermined threshold current. For example, an increase in the magnitude of the current, when a target binding event has occurred, may produce a change in the magnitude of the absolute current by a factor of one thousand, or 10 ³ . Further, the threshold for detection may be a current range, e.g. where the current range has upper and lower bounds), e.g. within which a current signal is be measured, for a defined period of time, in order to indicate/detect binding of a target entity 100.

Figure 2 illustrates one example procedure 200 for the detection of biological entities, where a microfluidics chip 202 is employed to direct the flow of the biofluid. Advantageously, hundreds and preferably even hundreds of thousands of sensors comprising electrode pairs 110 may be disposed onto a sensing area 204 bearing, or adjacent to, such a microfluidics chip 202. In one implementation, a microfluidics chip such as 202 is able to separate and disperse the biofluid prior to being passed over the plurality of sensors. In this way, a very large surface area of sensing region 204 is able to be covered with only a small amount of biofluid. In another implementation of the microfluidics chip, the chip bears at least one channel 206 which guides the biofluid directly over an array of sensors in the first instance. In other words, the array of sensors bears the microfluidics chip overhead, and is exposed to the direct flow of the biofluid. In implementations, however, the device is able to utilise multiple electrode pairs 110 simultaneously in order to optimise current detection, and improve the quantification of target binding events. For example, figures 3a and 3b show three electrode pairs 300 disposed in parallel at two different visual perspectives.

The microfluidics chip may be made of a polymer such as PDMS or PMMA. As described, in one implementation the channels 206 may be configured as a fractal arrangement 202, but may also be configured as a continuous path in order to guide flow of the biofluid over the sensors. Again, in either scenario, the described techniques can provide an advantageously high signal-to-noise ratio in detecting target entity 100 binding events (406, 604), resulting from a large plurality of binding events which are likely to occur on the array of sensors disposed on the chip.

Using microfluidics has the advantage, especially in the medical diagnostics industry, that only a small amount of liquid biopsy needs to be extracted (for example, 100 pl_ of fluid such as blood, as in Figure 2). This has the effect that the device may be utilised as an efficient and highly portable sensing device. Furthermore, the device can be easily utilised in a point-of-care manner, where the indicator of successful detection is a quickly measurable property of the output current. The microfluidics and electrode sensor array may be mass-produced according to a predetermined and highly adaptable fabrication specification.

Figure 4a and 4b show an electrode pair in an un-bound 402 and bound 406 (with target entity 100) state, respectively. It is contemplated that more than one binding event 406 may occur at a single electrode-pair 110 sensor. Thus Figure 4b, for example, shows an experimental setup after which multiple target entities 100 have been sequentially bound in-between an electrode pair. In other words, a sequence of target molecules 100 may become bound to the one or more recognition elements in succession over a single electrode pair. In some implementations the visible particles 404 may be magnetic nanoparticles onto which the target entities have been disposed, where the picture of 4b was taken after having dried the sensor and introduced it into a vacuum. Generally, when carrying out a method as described above, for example as shown in the scenario of Figure 4b, a group of target entities may become interdigitated or bound adjacent to one another. The entities will hence complete the circuit and facilitate (or difficult) electron transport across the electrodes after the voltage has been applied. With this scenario in mind, some implementations are able to detect such a sequential binding event using the same method of detecting a change in current. Upon each successive binding event, the device will be able to detect stepwise changes either in the absolute current, or a proportion of the existing current as previously described.

In some implementations, an increase in current may be detected, which may be indicative of a target entity. Figure 5 shows a current/voltage graph 510 which plots a current signal where a binding of a target entity 100 (in this case exosomes) has occurred. The number of sensors utilised simultaneously in this scenario is roughly 100. A suitable threshold for detection may be predetermined to be a certain proportional increase in current. In this example 510, it can be seen that an electrode with bound target entities (i.e. exosomes) may cause the current to increase in magnitude by roughly a factor of 1000 (or 3 orders of magnitude), relative to the control group 516, or unbound electrodes 402. Therefore, the threshold for indicating a binding event may, for example, be set at a minimum increase factor of 500.

Figure 4 depicts an example in which the exosomes are suspended in a phosphate- buffered saline (PBS) solution with pH 7.4. This may be an example of a synthetic biofluid, or alternatively, the PBS may be used as an additive to a physiological biofluid with a similar pH. The various ionic salts dissolved within the PBS solution give it an improved conductivity. Therefore, when the Exosomes 404 become bound to the electrodes 110, the displacement of the PBS from between the sensing region may cause the conductivity in the region between the electrodes to decrease. In this way, the threshold for detecting the target entity may be predetermined to be a certain decrease in absolute current or relative current magnitude.

Figure 5 shows, schematically, a current vs. voltage graph 510, showing signals derived from aggregate bindings of nanoparticles/exosomes and non-binding events. Figure 5B relates to a system in which the exosomes are initially suspended in a phosphate-buffered saline (PBS) solution with pH 7.4. This example comprises one hundred sensors, where each sensor comprises an electrode pair. Each electrode pair of the e.g. one hundred electrode pairs used to measure the signal determined in Figure 5 is capable of binding to the target biological entity, and measuring a current/impedance across the nanogap. In implementations each electrode pair is functionalised to capture nanoparticles/exosomes.

The upper grouping of signals 512, with current range of about 10 ⁶ to 10 ⁸ Amps, corresponds to groups of individual measurements 514 where a successful binding event has occurred between the nanogap of the electrodes. In the context of these exemplary results and group 512, a binding event 514 corresponds to either an individual nanoparticle/exosome bridging the gap or an aggregate of nanoparticles/exosomes bridging the electrode gap. Aggregations of target entities, e.g. with a large numbers of nanoparticles/exosomes (between 10 and 50), binding the gap result in a higher value of current to be measured. Nevertheless, it will be appreciated that even the binding events 518 with low current values (i.e. less than 10 ⁸ Amps, corresponding to fewer, or a single, nanoparticle/exosome bridging the electrode gap) still provide a signal a factor one thousand greater than a signal corresponding to a non-binding event 518. The lower group 516 of measurements corresponds to such non-binding events, for example, mere adsorption of targets onto a surface of an electrode, but where the gap 102 is not bridged. Thus, a distinction may be also made between the size of aggregate which bridges the sensing region nanogap 102.

Figure 6 illustrates a variety of successful binding events and non-binding events. Scenario 608 illustrates an electrode pair (which may have sensing aptamers attached) but no bound or adsorbed biological entity. Event 604 represents a successful binding event of the target entity between two aptamers/antibodies attached to each electrode of the electrode pair. Scenario 606 illustrates an electrode pair where a non-target entity is bound to the recognition molecules.

Scenario 610 illustrates an aggregate of various target and non-target entities or bio complexes, which are not bound to the electrodes, but have become merely adsorbed onto the surface. It should be noted that this aggregate 610 of biomolecules is easily washed, and adsorption is thus merely a transient event. In other words, because of the weaker nature of adsorption compared to being (chemically) bound to the aptamers/antibodies, the aggregate possesses a higher dissociation constant. Further, the aggregate binding is in contrast to the state where a dual, permanent, binding occurs of both a target entity and a non-target entity in parallel 602. In 602, both molecules are bound, and therefore possess lower dissociation constants.

Event 612 represents merely adsorption of an entity onto a wall of an electrode, but which does not bridge the electrode gap. Therefore, no‘ohmic’ current representative of a resistor (or, a very low or indiscernible signal) will be generated due to event 612. Nonetheless some implementations of the system measure current changes due to electron transport changes resulting from the trapping of a biological entity between two conductive surfaces 604, rather than binding on a single surface 612.

There can be many mechanisms involved in current flow, particularly where AC current is concerned, and in general the techniques described herein should not be taken to be limited to any particular underlying mechanism for a detected change in DC and/or AC current (or, equivalently, conductance, resistance, capacitance, or impedance).

Various currents may be associated with the potential binding scenarios depicted in Figure 6. In a case where target entities are detected directly, without using nanoparticles, a current signal indicative of a successful binding event of the target entity is may be distinguishable from other binding events, as only this event may lie within a detection threshold. Other current levels may indicate events lacking binding, or a dual binding of a target and non-target entity (which may be indicated by a stepwise increase in the current), the binding of only a non-target entity (which may show a lower current flow), and the aggregate and non-stable adsorption of various bio-complexes, which may be transient events: Since the aggregate adsorption is easily washed away the detection of this may be verified by a sudden drop in current upon washing. Generally, an adsorption or physisorption (non-chemical bond) event may be indicated by only a fleeting change in measured current.

One advantage of some implementations is that the functionalized electrode pair can provide stability for binding (and subsequently detecting) target entities. In other words, capturing a protein at two points, each anchored to a separate electrode, may provide a far more stable binding mechanism than binding at a single point, therefore facilitating detection of entities at a sensor, where the entities are comprised within a fluid which flows over the sensor.

In implementations the sensor may be adapted in order to anticipate certain predetermined properties of the target biological entity. For example, when the target entity is a protein, or other molecule which is susceptible to becoming charged, the isoelectric point (pi) of the protein may be known. The pi will then be accounted for in the current detection when the sensing method takes place. The isoelectric point (pi) is defined as the pH at which a particular molecule carries a net electric charge of zero in the statistical mean. It will be understood by the skilled person, therefore, that the anticipation of the pi of an entity (or pH of a fluid) is only relevant to a sensing method which takes place in a fluid such as serum, as pH and pi have no meaning in a solid or crystalline medium.

In some implementations, in place of a real liquid biopsy or biofluid being used, a synthetic biofluid may be used. This synthetic biofluid will contain the target entity, and may be used for the purposes of testing or calibrating the sensor device. For example, a range of synthetic fluids having various pH levels and a range of target entities with various pi values may be passed over the sensor in different combinations. The current response of the device for each combination will be measured. In this way, it may be possible to more accurately select an appropriate detection threshold for a particular combination of serum pH and target-entity pi.

When the pi of the protein is known, the applied voltage of the sensor may be adjusted during the sensing procedure. In one implementation the pH of the biofluid and the pi of the protein will be predetermined, and from these two parameters it will be possible to calculate the likely charge of the protein. For example, if the pH of the serum is lower than the pi of the protein, the protein is likely to be positively charged. In anticipation of the probable electrical charge of the protein, the voltage applied to the sensor may be altered (e.g. increased or decreased) in order to improve the sensitivity of the system to detecting a change in current upon binding.

In some implementations, it is possible to synthetically adjust the pH of the serum before being disposed onto the sensor. For example, a pH buffer solution may be added in order to adjust the pH of the serum to match the pi of the protein. By matching the pi to the pH, the charge of the protein may be controlled to be kept (on statistical mean) electrically neutral. However, it can be preferred to maintain the pH of the biofluid, which in most physiologies would typically be a pH of 7.4. It is preferred to maintain the pH because the 3-dimensional folded structure of the protein may be susceptible to an unfavourable change in conformation in a different (unnatural) pH. It will be understood by the skilled person that the pi may relate to any target biological entity, and not necessarily a protein.

In addition to the pi of the target entity, the size and dimensions of the target biological entity may be known in advance. In this way, when the disposed layer 108 (which may comprise gold) is chemically grown or otherwise disposed onto the electrodes, certain control parameters may be introduced in order to control the growth of the layer 108 to match, or correlate to, the size of the target entity. In this way, the resultant gap 102 between the sensing regions will be correlated to the dimensions of the target entity 100. Advantageously, by anticipating the dimensions of the target entity during fabrication of the sensor, the selectivity of the sensor may be further improved, which will in turn increase the reliability of the device in detecting the target entity 100. As previously mentioned, the reliability may correspond to an increased signal-to-noise ratio in detecting the change in current. It is important to appreciate that the recognition molecules 106 may be of a similar size or even larger than the target entity. Therefore, the combined dimensions of the recognition elements 106 when bound to the target entity are accounted for when predetermining the electrode gap 102.

Furthermore, the predicted or most likely orientation of the target entity 100 when it becomes bound to the sensing region may be anticipated. In this way, the gap size 102 will be tuned to further account for the most likely bound sate of the target entity 100. In other words, the distance across the entity in its predicted bound orientation will be predetermined, and the gap dimensions 104 and 102 of the sensor will be fabricated in order to correlate to this length.

By predetermining the various electrode size parameters, it should be appreciated that the described system can attain a high level of sensitivity. The sensitivity in this context relates to the fact that the electrode-pair dimensions are tuned with the aim that only a single entity becomes bound or adsorbed in the sensing region. In other words, in some implementations the system may achieve single target detection per sensor and the detected current change will then be indicative of only a single entity/molecule. This has the advantage that different entities may be easily distinguished from one another.

In some implementations, more than one type of target molecule may be present, where electrodes can be designed in such a way as to ensure single-molecule binding. Some implementations thus enable differentiation between these two target molecules, once bound to a sensing region, based on their respective detected current changes. Therefore, some implementations also achieve a high level of specificity in that the presence of multiple target entities, and any irrelevant non-target entities, does not unfavourably convolute the resultant detected change in current. Advantageously, the specificity can make it possible to mitigate for a false-positive result, since precise current thresholds indicative of the binding of individual target molecules are possible to measure.

It is possible to enhance the conducting properties of the target biological entity 100 prior to the sensing routine. By modifying the target entity 100 in such a way, the signal-to-noise ratio of the current detection will be further increased. For example, where the target entity is a protein, modification may include attaching molecules 122 such as antibodies or antibody fragments to the surface of the protein. The molecules 122 may be selectively attached to the target entity 100 in order to coat its outer surface. Other molecules 122 such as nanoparticles and other co-factors may alternatively be attached to the surface of the protein. The nanoparticles may be specifically chosen to functionally increase the electrical conductivity of the proteins. Modification may be performed in a separate vessel and prior to the biofluid being introduced to the sensor. It will of course be understood by the skilled person that multiple different entities 122 may be simultaneously attached to the target entity 100 in order to improve its electrical or conducting properties. These enhancing molecules 122 are chosen such that they specifically attach to the target entity 100.

Alternatively, the enhancing molecules 122 may also be chosen to modify the surface of the target 100 entity for purpose of promoting the binding of the target entity to the target recognition molecules 106. For example, the surface enhancing molecules 122 may be selected because they increase the strength of the chemical bond formed to the target recognition molecules. The dimensions or configuration of the electrode pair may be further adapted in anticipation of the predetermined dimensions of the target entity 100. Specifically, width, shape or curvature of the electrode pair 112 tips 108 may be altered so as to increase the area or volume of the sensing region interposed between the electrodes. For example, in certain implementations, the tip of the electrode surface 108 may be flattened and/or widened such that more target entities may be bound in parallel in the sensing region. This may facilitate a sequence of target entities 100 becoming bound and a series of stepwise changes in current may optionally then be measured. Furthermore, this electrode tip fabrication may be altered in anticipation of a predicted likelihood of the orientation of the target molecule 100 upon its capture.

Alternatively, adapting the electrodes in this way may simply be done to increase the likelihood of a desired method of capture. In some implementations, the width of the electrode tips (corresponding to horizontal distance axis in Figure 6) may be reduced to approximately 10 to 40 nm, in order to ensure that there is only sufficient space for the recognition elements 106 and a single 604 target entity 100 to become bound. In this way, the likelihood of the parallel binding scenario (602, 406) is significantly reduced, and the likelihood of singular target binding 604 is promoted.

In addition to providing an indication of a target entity based on its successful binding to the sensor, some implementations provide a further method of distinguishing between the possible binding of a non-target molecule and a target molecule. It is possible in some scenarios that a non-target entity will bind (602, 606) to the electrode pair and produce some change in current, but which is not indicative of the current change expected from the binding of the target entity 100. The current change threshold may be selected carefully to be indicative of only the target entity, i.e. the current threshold may be given upper and lower bounds.

A further method of detection is possible with the system, in which a time-dependent voltage is applied over the electrode pair and the resultant transient current (in arbitrary units) is measured. For example a constant voltage may be applied, and the current may be measured over a period of time. Moreover, the voltage may be cycled through a charging and discharging routine wherein the voltage is increased from 0 up to a maximum value and then returned back to zero. The rate of change of voltage in the cycling procedure may be fixed. Alternatively, a variable rate of change of voltage may be applied in order to observe a capacitive difference. Such a method may exploit the history dependence of the system, i.e. hysteresis. In this way, a more sophisticated transient current profile may be used as an indicator for the binding of a target entity 100, and may help to further distinguish between the binding and/or adsorption of merely a non-target entity.

In some implementations the device may comprise a substrate or chip having a lateral dimension of <1mm, potentially <100pm, <10pm, or <1 pm, optionally in a microfluidics system. In some implementations the sensor comprises a pair of electrodes configured to detect a protein in a bodily fluid. The separation of the electrodes defines a gap having nanometre dimensions, which defines a sensing region, which may have a lateral separation of <50 nm, <20 nm, <10 nm, or in implementations ~5 nm or less e.g. ~2nm.

Figure 7 shows, schematically, an example current-time graph illustrating various potential binding/adsorption scenarios. The upper signal 604 corresponds to the target entity binding event 604 in Figure 6, corresponding to a single target entity bridging the electrode gap 102. Other non-target-binding events (606, 610, 602, 608, 612) cause a much lower signal, in implementations with a lower magnitude than the signal produced by a binding event by a factor of e.g. approximately 1000.

Further method and corresponding systems for selectively detecting the presence of a target biological entity, such as a protein or extracellular vesicle, e.g. exosome, within a biofluid are described below.

Figure 8 illustrates a sectional view of an electrode pair with a nano-gap, and an electrical double layer in the surrounding fluid. In this sensor example 800, individual ions 802 persist in the medium, and concentrate over the electrodes’ 808, 810 surface to form a double-layer 804. This double layer can be seen to overlap in the sensing region between the electrodes 806.

It may be beneficial both for capturing of a target entity, and subsequent sensing/measurement of a target entity, to ensure that the electrical double layer (EDL) on each respective electrode of the electrode pair overlaps in the sensing region. That is, the EDL overlaps in the nanogap 806, 102 between the sensor 800. In other words, it may be favourable for the lateral distance between the pair of electrodes may be less than twice the width of the EDL in the surrounding solution. In some implementations the lateral distance between the pair of electrodes may need to be sufficiently small in order to ensure this.

Generally, an EDL is a structure that appears on the surface of a charged surface object when it is exposed to a solution with dissolved ions. The double layer refers to two parallel layers of charge surrounding the surface (e.g., the electrode surface). The first layer comprises ions adsorbed onto the surface due to either chemical interactions and/or a charged electrode surface. The second layer Is composed of ions attracted to the first layer via the Coulomb force, which electrically screen the first layer. In this way, the width of the EDL may be seen as equivalent to the Debye length in an ionic solution. The Debye length is generally a property of an ionic solution which is a measure of a charge carrier’s net electrostatic effect, and to what extent said effect persists in the solution. Every Debye-length T _D [ ], the electric potential will decrease (i.e. be screened) in magnitude by 1/e.

The width of electrical double / Debye layer is a function of at least the ionic strength of the fluid. Moreover, the Debye length may be on the order of a few nanometres on an electrode having an applied potential. In implementations, a gap 102 between the electrode pair may need to be less than ~ 20 nm in order for the Debye length (i.e. EDL width) to be equal or less than half the lateral distance between the pair of electrodes. In some implementations, even narrower gaps 102 may be created, e.g. as low as 5 nm or even 2 nm gaps.

The Debye length is related to ionic strength of the fluid, according to:

where z represents ionic species, and q their corresponding charge values. In implementations, an ionic solution comprising at least 0.01x or 0.1x PBS is used to maintain the 3D conformation of certain biological targets. Moreover, this concentration of PBS leads to a Debye length of 2.4 nm at standard conditions. Therefore, a lateral distance 102 between electrodes would be required to be at most ~4.8 nm. Advantageously, by tuning the width of the nanogap and the ionic concentration (i.e. with PBS) in this way, noise due to the‘screening effect’ and due to charges 802 moving in the bulk above may be mitigated. Furthermore, when a target entity is captured in the nanogap having an overlapping EDL, charge transport across the gap may be supported. For example, electron tunnelling may be encouraged. Thus, a higher signal-to-noise ratio is achievable, in particular with an increasingly narrow electrode gap.

Various other advantages may be associated with narrower gaps on the order of 20 nm, 10 nm, or even ~5 nm or ~2nm, despite the difficulty associated with reproducibly fabricating such narrow electrode gaps. For example, other non-classical electron transport effects (in addition to tunnelling) may be encouraged, such as Flickering resonance. Generally, this mode of electron transport only occurs in the range of about 1-2 nm. However, fabrication of small nanogap around 5 nm may encourage Flickering resonance which in turn will improve sensitivity of a measurement when a target entity becomes being between sensing region of the electrodes.

Entity Characterization with Alternating Electric Fields and Nanoparticles:

In some implementations, the sensors bearing electrode pairs, or array of sensors/electrodes, may not need to be functionalised with target recognition molecules (e.g. aptamers). We describe below a further example which does not rely on mass transport/diffusion in order for the target entity to reach the sensing region of the electrode. Alternating electric fields may be used to induce an electrophoretic force on conductive nanoparticles bound to target entities, and bound nanoparticle-entity assemblies may be actively transported to, and concentrated around, a sensing region or nanogap of an electrode.

In detail, the time averaged DEP force of a spherical particle with radius R and in a solution with a dielectric permittivity of s _m is provided below:

FDEP (w) = ne _m R ³ Re(f _CM (o ))V\E\ ²

The‘real’ part of the above Clausius-Mossotti factor (CMF), Re(/ _CM (<y)), determines the direction of the DEP force based on the dielectric permittivity and conductivity (inside the CMF term) of the solution and particle. With conductive nanoparticles, this force is generally attractive between the nanoparticles and the source of the electric field. However, with dielectric particles, this DEP force may become repulsive. The gradient of the E-field in solution squared, n|£Ί ² , correlates with the supply voltage applied across the DEP electrodes. The DEP electrodes may also be the same electrodes 910, 912 as used to characterise and measure the sensing region to detect the presence of the target entity. Additionally, the magnitude and the sign (i.e. attractive or repulsive) of the DEP is also a function of permittivity of a particle (not only conductivity), especially for biological targets. Furthermore, the DEP force can be a function of the frequency of the applied E-field, especially for biological targets. However, for metallic particles, DEP force is generally invariant with respect to the frequency of the applied E-field.

Advantageously, dielectrophoresis (DEP) offers rapid concentration and isolation of nanoparticulate matter that does not depend on specific chemical binding or alterations. The DEP process commonly utilizes two electrodes in solution that are subjected to an alternating electric field (E-field). The force on the particles derives from the fact that the alternating electric field induces local dipoles within the particles. These local dipoles cause a net force toward, or away from, the E-field gradient depending on: the frequency of oscillation, and the relative dielectric permittivity of the particle and surrounding medium.

For example, particle attraction may be increased with a high conductivity of particles relative to the surrounding medium. For example, conductive metal nanoparticles (e.g. gold) may be used in a substantially non-conductive medium (e.g. ultra-purified de ionised water such as milli-Q water, which may have a resistivity of around 18 MQcm).

In summary, both the sign and the magnitude of the DEP force acting on particles due to an alternating electric field may be a function of any of: particle size, particle permittivity, particle conductivity, applied frequency, electric field gradient (which may be influenced by tuning the size of the nanogap), and/or relative conductivity of the particle relative to the medium (which may be influenced by the ionic strength of surrounding fluid, e.g. by tuning a concentration of PBS).

Detecting Extracellular Vesicles (EVs), by known techniques can require purification by ultracentrifugation or precipitation, and detection through Nanoparticle Tracking Analysis (NTA) or Dynamic Light Scattering (DLS). Both these measure the Brownian motion of nanoparticles, whose speed of motion, or diffusion constant, is related to particle size through the Stokes-Einstein equation. These techniques typically require a concentration in solution of 10 ^{6 _} 10 ^s EVs and a volume of at least 300 pL to be injected into a measurement chamber. Therefore, such devices are not sensitive to small concentrations of extracellular vesicles such as exosomes and provide almost no information relating to proteins or other biomarkers decorating the surface of EVs.

There is now described a much more sensitive system and method, which can produce a high signal to noise ratio not only in detecting e.g. exosomes but also in transporting relevant biomarkers.

Figure 9 illustrates an example of a method of using dielectrophoresis (DEP) to influence and accelerate the transport of nanoparticle-entity assemblies/pairs. Figure 9 further illustrates the capture and characterization of the target entity 908 with the aid of gold nanoparticles 906. The electrode 910, 912 are gold in the illustrated example, wherein a nano-gap of approximately 40 - 50 nm exists, although this may be as low as 10 nm or even 5 nm. The electrodes are provided on a substrate 14 as in other implementations.

Target entities such as exosomes 908 forming part of a biofluid may be sequestered by functionalised nanoparticles 906. In the example of figure 9, the nanoparticles 906 are gold nanoparticles. They are functionalised with an aptamer 918 or other suitable linker which interacts to bind with the exosome 908. Thus, a nanoparticle-exosome pair is formed. In implementations, greater numbers may aggregate to more generally form assemblies. For example, two nanoparticles 906 may sequester a single exosome 908.

Step 900 shows such a binding/sequestering event taking place in order to form the nanoparticle-entity assembly. It may be advantageous to use an excess of GNPs compared to exosomes, in order that all of the exosomes become sequestered/bound, and no naked exosomes remain. This will of course require that naked GNPs be filtered/separated removed from the mixture in a separate step. Additionally, factors such as mass transport and rates of diffusion of particles may be accounted for in the mixing step. Sufficient time is allowed for the exosomes 908 to interact with the GNPs 906 and form a bond. Larger nanoparticles have much slower passive diffusion through a medium and therefore form bound assemblies at a slower rate. However, larger nanoparticles also respond better to DEP (i.e. possess a stronger attractive force in the presence of an alternating E-field). Thus a balance of GNP size may be struck; GNPs between about 20 nm - 100 nm provide a good balance between rates of passive diffusion, and rates of concentration during the DEP step 902.

In between mixing step 900 (comprising introduction of nanoparticles to exosomes to encourage binding an formation of nanoparticle-entity assemblies) and concentration step 902 (concentration of assemblies around sensing region) a further treatment step may be implemented in order to filter or separate the bound assemblies from the unbound GNPs 906, and/or the unbound exosomes. Such a separation treatment step is described in detail below.

Step 902 depicts an electric field 916 being applied to the medium such that a force is exerted on the nanoparticle(s) 906. In implementations, for a force to be exerted on the nanoparticles 906, a relative difference should exist between the conductivity of the nanoparticles and the surrounding medium as described above. Thus, in implementations, gold nanoparticles may be used which are more conductive than the dilute ionic medium in which they are suspended. Since the nanoparticles are bound to the exosomes, the exosome also become attracted to the electrode sensing region. An alternating current should be applied to induce an alternating electric field. In some implementations a frequency of up to about 1.5 MHz may be used.

The electrophoretic force causes any nanoparticles 908 present in the surrounding medium to be concentrated around the sensing region. Advantageously, the electrodes 910, 912 do not need to be functionalised with aptamers or any kind of target recognition molecule. Gold nanoparticles (GNPs) have a natural affinity, i.e. a thermodynamically favourable interaction, with gold electrodes. Therefore, when the nanoparticles 908 become concentrated around the medium due to DEP, the GNPs 918, as part of the nanoparticle-entity assemblies, may become bound to the region 102 in between the electrodes. This can be seen in step 904.

In implementations, the electrode 910, 912 may be functionalized with linker molecules with further encourage/enhance an interaction between the gold nanoparticles and the region 102 of the gap around the electrodes. For examples, thiol linkers can be used. Advantageously, sulphur in the thiol linkers provides a thermodynamically favourable interaction with the gold nanoparticles.

An increasingly narrow lateral separation of electrodes may increase the dielectrophoretic (DEP) force without the need to increase a voltage to power the applied electric field. However with DEP there can be adverse effects on the sample solution and its analytes due to large trapping voltages. In order to overcome the thermal motion of particles with dimensions under 100 nm, conventional DEP electrodes with micrometre-scale gaps typically require an unfavourable high trapping voltage, for example, a minimum of 10 Vpp. Such large trapping voltages can cause Joule heating, bubble formation, and unfavourable electrochemical reactions.

Implementations of the described system/method have small gaps between the electrodes, e.g. less than ~ 10 nm, to facilitate bridging this gap with one or a few nanoparticles. However this can provide an additional advantage of increasing the DEP trapping force without the need to raise the trapping voltage, mitigating these unfavourable effects.

Further, by reducing the width between DEP electrodes, the gradient of the E-field can be increased substantially, which provides a greater force on the nanoparticles and thus an improved ability to efficiently concentrate the GNPs/nanoparticle-entity assemblies around the sensing region. Ultimately, this can result in a method more sensitive to the presence of target entities, since the nanoparticle-entity assemblies may be collected more efficiently.

As mentioned above, with certain particles having dielectric properties (that is, electrica!!y insulating, but which may be polarized by an applied eiectric field), the DEP force may become repulsive. Exosomes 908 are an example of particles which may exhibit a repulsive force (i.e. relative to the sensing region) when subjected to an alternating E-field 916. The magnitude of this force may depend on the size of the exosome, which may vary for example between about 50 - 150 nm. Thus, the exosomes may generally be larger than the GNPs 906 which may range from about 20 nm - 100 nm. Thus, in an unfavourable scenario, a repulsive force acting on an exosome 908 due to an alternating E-field 916 may be greater than an attractive force acting on a nanoparticle 906. In other words, the force on the exosome 908 (due to its size and/or relative conductivity) might be higher than a force acting on the gold nanoparticle 906. Thus, the bound assembly may be pushed away from the sensing region 102 (i.e. along a path of decreasing gradient of the electric field).

Exosomes 908 comprise a lipid bilayer which has a low conductivity, and thus may exhibit dielectric properties. Therefore, when an alternating E-field 916 is applied to a medium, exosomes 908 may be repelled in said medium, and (gold) nanoparticles 906 may be attracted. Thus, the binding due to the linker/aptamer 918 may break. In another unfavourable scenario, the bound nanoparticle-exosome assembly may be repelled from the sensing region.

In order to mitigate against the above scenario, and ensure that the assemblies become concentrated around the sensing region, it is desirable to increase the conductivity of exosomes relative to medium, and thus and induce an attractive DEP effect (i.e. an attractive force on the exosome in addition to the nanoparticle). In an implementation which achieves this, a solution of e.g. about 1% PBS (-0.018 S/m) may be used, with E-field 916 frequencies in the range of 1 MHz - 10 MHz. Hence, exosomes 908 may be positively attracted (or at least neutrally affected) by the alternating E-field (DEP force) and thus binding around the nanogap sensing region 102 is maintained.

In Figure 9, an AC current is applied to the electrodes 910, 912 which induces an alternating E-field in the fluid suspension containing the nanoparticle-entity assemblies in step 902. After the attraction and concentrating 902 of the nanoparticle-entity assemblies around the sensing region 102, the assemblies bridge the electrode nanogap 102. A direct current is generally then applied, which is used to characterise/measure a response of the sensing region in order to identify whether the assemblies (and thus exosomes) are present. A voltage of around 1 V may be applied to produce a direct current in 904. A baseline current (where no bridging occurs, e.g. as in Figure 10A) may be around 1 - 20 pA. A current produce from a bridged gap (e.g. as in Figure 11 A, described below) may be around 1 - 100 nA. Thus, a signal to noise ratio of over a thousand may be achieved in implementations of the described method.

An ohmic response (i.e. indicative of classical resistor behaviour) generally indicates that the nanogap sensing region 102 is bridged by the assemblies. Therefore, a characteristic linear relationship may be seen when a direct current is applied across the electrodes. However, in other implementations, an AC current may be used to characterise a response of the sensing region. In implementations, the same AC current as used to induce a DEP force in the nanoparticles may be used to characterise an impedance response. In this way, an ohmic response may still be identified for example, by identifying a characteristic impedance profile of a classical resistor.

As a further example, as mentioned above, when very narrow electrode gaps are fabricated on the order of 5 nm, non-classical electron transport effects may be seen upon application of a constant voltage to an electrode pair bridged with nanoparticles. Thus, if effects such as tunnelling and flickering resonance are to be expected, the identification of nanoparticle-entity assemblies may comprise identifying a non-classical current-response profile for a resistor.

Treatment and separation:

As mentioned above, in implementations a further treatment step is used in order to filter/separate/sequester the bound assemblies from the unbound GNPs 906, and/or the unbound exosomes 908. A separation step is used so that unbound nanoparticles, e.g. GNPs 906, do not persist in the mixture, become attracted to the sensing region, and produce a‘false positive’ signal. The treatment/separation step exploits a variety of property of the particles to be separated, for example: particle density, size, conductivity, and/or permittivity.

Treating according to density in the precursor mixture may comprise centrifugation of the precursor mixture. For example, the precursor mixture containing a suspension of: unbound nanoparticles; unbound target entities (e.g. exosomes); and bound nanoparticle-entity assemblies, may be passed through a centrifuge, or subjected to centrifugation. This will separate all particles in the suspension according to density, and thus may allow extraction and isolation of only the bound nanoparticle-entity assemblies.

Treating the precursor mixture according to size may comprise passing the precursor mixture through a microfluidic array comprising a filter. This filter may be a mechanical filter, or may exploit laminar flow properties in order to separate particles by size. Thus, in this way, a portion of the suspension or fluid may be isolated which contains only the larger nanoparticle-entity assemblies, but not any unbound nanoparticles. In other implementations, a physical mechanical filter may be used such as a nanoscale mesh or molecular sieve which filters component the precursor mixture by size. For example, a microfluidic array may be used which contains micrometre-sized or nanometre sized pores, which may be used to collect bound nanoparticle-exosome assemblies.

Treating according to conductivity/permittivity in the precursor mixture may comprise applying an alternating electric field to the precursor mixture. That is, an additional dielectrophoresis (DEP) step may be performed, prior to concentrating the assemblies around the sensing region, in order to separate/isolate the nanoparticle-entity assemblies. In this way, the size of the components of the mixture may play a factor: that is, rates of active diffusion of particles caused by a DEP force may be dependent on particles size and medium viscosity.

Examples

Figure 10A shows a plan view scanning electron microscope (SEM) image of a single electrode pair with nano-gap, having single and aggregate Gold nano particles adsorbed onto the electrodes.

The resultant structure shown in Figure 10A was generated using the method described above in conjunction with Figure 10A. However, no exosomes are present in the SEM image or method used in conjunction with Figure 10A. The electrodes 910, 912 shown are un-passivated.

In detail, example experimental parameters are as follows:

- A gap size of ~ 40-50 nm is used;

- An electric field is generated with an AC amplitude of 1 5V, is equivalent to an RMS voltage (Vrms) of -1.06V.

- The frequency of the electric field is driven at 1 MHz, and held constant for 30 seconds.

- PBS solution of 100x dilution is used;

- Gold nanoparticles of stock concentration are used (2.6 x 10 ¹⁰ per ml) are used; - The total volume of the resultant fluid is 45 mI_

- Control (i.e. no DEP applied) showed that no attraction occurred between GNPs and electrode.

Figure 10B shows a plan view scanning electron microscope (SEM) image of an array of five electrode pairs with a nano-gap. Each of the sensing regions 102 shown is equivalent to the sensing region/nanogap in figure 10A. Again, each of the electrodes 910 and 912 are gold and grown such that each end approaches the other in order to produce a nano gap of less than 100 nm. An array 1000 of sensors as shown here may comprise far more than 5 electrode pairs. For example, more than 10 or 100 sensors may be used, for example as used to produce the results in figure 5B.

When used, for example a method to attract and characterise nanoparticle-entity assemblies as in figure 9, each of the sensing regions 102 may be used to apply an alternating electric field, and subsequently attract/concentrate assemblies. Each sensing region 102 may then be used to apply a direct current to characterise a response of each of the regions to identify the presence of target entities.

Figure 11A shows a plan view scanning electron microscope (SEM) image 1100 of a single electrode pair with nano-gap 102, having gold nanoparticles 906 bridging the nano-gap between the electrodes. The nanogap 102 in Figure 11A shown is approximately 40 - 50 nm. The conditions for providing an alternating electric field to the medium to concentrate the GNPs 906 using DEP are: 1.5 V amplitude AC at a frequency of 1 MHz; PBS used at 1 % dilution; DEP conditions were sustained for 1 minute.

It should be appreciated that, in the examples, no bound nanoparticle-assemblies have been created prior to DEP and characterization. Figures 11A - 11C merely show an example with ‘naked’ unbound gold nanoparticles. However, the skilled person will appreciate that results described here may be equivalent to a result where bound nanoparticle-entity (e.g. nanoparticle-exosome) assemblies are used. Thus, these results relate to an example where no filtering step has been provided to remove the naked/unbound GNPs. Figure 11 B shows a higher resolution image 1102 of the region around the nanogap in Figure 11A. It can be seen that the GNPs 906 form a contiguous bridge between each end of the electrode either side of the nanogap 102. Thus, a conductive path is formed which may behave as a resistor, and thus can be characterised by either applying a DC or AV voltage of approximately 1 - 1.5 V.

Figure 11C shows voltage current plots of a response 1104 of the bridged nanogap in 11A and 11 B (left-hand side) and a response 1106 of a non-bridged nanogap. During application of a DC voltage used to characterise a response of the sensing region (to identify presence of the GNPs bridging the nanogap), a signal of 33 nA was recorded. This relates to a signal to noise ratio of over 10,000.

The derivation of this signal-to-noise ratio can be appreciated from the current-voltage plots 1104 and 1106. The current-voltage profile in 1104 is linear, and thus represents an ’ohmic’ response representative of a classical resistor, corresponding to the structure of Figures 11A and 11 B. In this sense, the bridged gap seen in Figure 11 B behaves as a classical resistor. The signal in 1106 shows no ohmic response whatsoever, as it corresponds to a nano-gap which is not bridged by any GNPs. Thus, graph 1106 corresponds e.g. to figure 10A, where adsorption of the GNPs has occurred on the surfaces of the electrodes away from the sensing region 102.

Figure 11 C is thus an example of characterising data 1104, 1106 produced from a sensing region when a direct current is applied between the electrodes. This further demonstrates the characterizing of a response of the sensing region to the nanoparticle sensing voltage to determine treated precursor mixture characterizing data. For example, characterization methods (for example, including pattern recognition methods, which may be learned by a machine) may be used to characterise 1104 as data representative of a bridged gap, or the presence of the target entity. Data 1 106 is an example of data which may be characterized as an absence of a target biological entity.

The above embodiments have been described by way of example only, and the described embodiments are to be considered in all respects only as illustrative and not restrictive. It will be appreciated that variations of the described embodiments may be made without departing from the scope of the invention.