Patent Application
20240253033
This disclosure incorporates by reference the sequence listing XML file submitted via the USPTO patent electronic filing system titled “2023-11-09 APP—Sequence Listing.XML” which was created on Nov. 9, 2023 and has a size of 38.4 KB.
This disclosure relates to the field of microfluidics and detection systems.
Optical signal transduction techniques (e.g. colorimetric, light, spectroscopy) and other detection techniques such as electrical sensing, have been extensively applied for the detection of biological analytes. The developments of point of care devices that utilize these techniques is on the rise. In the case of optical techniques, this can majorly be attributed to key features of colorimetric readouts, such as high sensitivity, ease of analysis and interpretation, minimal training requirements, and affordability. However, one of the major drawbacks with colorimetric or electric readout systems is the challenge with the interpretation of results, the major reason being variability in the sample (e.g. varying saliva compositions), and the requirements for sophisticated instrumentation. To offset these drawbacks, recently, there have been increasing efforts to develop portable setups for interpreting the readout with smartphones and open-source technologies at the core of automation and data transmission units. Implementing smartphone and open-source technologies, can allow miniaturization and would ease the process of data collection and analysis. In the case of colorimetric readouts, most of the previously reported imaging setups were designed for assays where the color change is driven by the assay and imaging setups act as a proxy to the human eye to either reduce user-to-user variability, facilitate quantification and/or enable automation. Improvements in portable detection platforms (e.g. colorimetric, electric and the like) are therefore desired, particularly with respect to user friendliness, user-to-user variability, assay speed, signal quantification and automation.
In one aspect, there is provided a detection system for detecting an analyte in a sample, comprising:
In some embodiments, the detection apparatus is a spectroscopy detection apparatus or an electrical detection apparatus.
In some embodiments, the microfluidic chip further comprises a filter barrier fluidly connected to the incubation chamber, downstream of the incubation chamber inlet. In some embodiments, the detection apparatus is a light detection apparatus and comprises a light source for providing an epi illumination on the sensing chamber of the microfluidic chip, a condensing lens for condensing light from the sensing chamber of the microfluidic chip, and an image sensor receiving the light condensed by the condensing lens, the image sensor adapted to register the light as an electronic signal and to send said electronic signal to a processing device. In some embodiments, the microfluidic chip is part of a microfluidic cartridge comprising an inlet apparatus connected to the microfluidic chip, and covering the inlet, the incubation chamber and the filter barrier of the microfluidic chip, the inlet apparatus comprising: a receptacle in fluid communication with the inlet of the microfluidic chip, the receptacle being adapted to receive the sample, a storage chamber having a rupturable membrane and housing a colorimetric sensor, the storage chamber is in fluid communication with the microfluidic chip upstream of the sensing chamber and the filter barrier of the microfluidic chip, an outlet apparatus connected to the microfluidic chip and covering the outlet, the outlet apparatus is in fluid communication with the outlet of the microfluidic chip, the outlet apparatus comprising: the actuator, wherein the actuator is a screw adapted to be released in order to drive the flow of the sample to the incubation chamber, a second suction membrane adapted to be released in order to flow the sample past the filter barrier and to drive a flow of the colorimetric sensor released from the storage chamber. In some embodiments, the outlet apparatus comprises a second crew, and wherein the screw is adapted to be released in order to drive the flow of the sample to the incubation chamber, and the second screw is adapted to be released in order to flow the sample past the filter barrier and to drive the flow of the colorimetric sensor released from the storage chamber. In some embodiments, a closing lid is provided for closing the receptacle of the inlet apparatus. In some embodiments, a second storage chamber is provided in the inlet apparatus, the second storage chamber housing lysis reagents and having a rupturable membrane. In some embodiments, a second piercing actuator is provided in the inlet apparatus to pierce the rupturable membrane of the second storage chamber, and the second piercing actuator is connected to the actuating motor. In some embodiments, a heating actuator is provided in the inlet apparatus comprising a heating element for a lysis of the sample, and the heating actuator is connected to the actuating motor.
In some embodiments the detection system further comprises an imaging box, wherein the imaging box comprises the suction membrane, the detection apparatus and the actuator, and wherein the suction membrane is positioned below a support which is adapted to releasably bind to the microfluidic chip, and wherein the detection apparatus is a light detection apparatus.
In some embodiments the flow of the sample is bidirectional.
In some embodiments the detection apparatus is a light detection apparatus and wherein the sensing chamber comprises a plasmonic nanosurface, the plasmonic nanosurface including nanostructures protruding from the plasmonic nanosurface, the nanostructures having a size that is smaller than that of the diffraction limit of light, the nanostructures having a metallic layer that is plasmon-supported on top of a back reflector layer.
In some embodiments the detection apparatus is an electrical detection apparatus and wherein the sensing chamber comprises a dimeric DNA aptamer gold nanostructure.
In some embodiments the microfluidic chip comprises a plurality of the sensing chamber and multiple parallel channels each leading to one of the sensing chambers.
In some embodiments the detection system further comprises a motorized platform connected to the detection apparatus for moving the detection apparatus between the plurality of the sensing chamber.
In some embodiments the analyte is selected from nucleic acid, microorganism, a cell of a multicellular organism, or a protein.
In some embodiments the detection system further comprises a heating plate.
In some embodiments the actuating motor is connected to a controller and the controller is coupled to the processing device.
In some embodiments the processing device is selected from a smart phone, a tablet, or a computer.
Many further features and combinations thereof concerning the present improvements will appear to those skilled in the art following a reading of the instant disclosure.
The present detection systems leverage a controlled flow of fluid in a microfluidic chip with the use of a suction membrane at the outlet of the microfluidic chip. In operation, when the sample is provided in the inlet of the microfluidic chip, the suction membrane controls the air pressure in the channels by expanding or contracting to thereby drive the fluid flow towards the outlet or back towards the inlet. Conventional systems that use a screw or similar means directly at the outlet of the microfluidic chip are associated with leakage and inaccurate control of the fluid flow. The addition of the membrane which itself can be actuated by a screw for example, allows an improved control over the fluid flow (including allowing bidirectional flow) and does not lead to leakage. One advantage of the suction membrane is that it can controllably halt the flow of the sample above a reaction zone and/or a sensing region of the microfluidic chip without allowing the sample to leak or uncontrollably continue to flow. The bidirectional flow also allows for the sample to be flowed back and forth between the reaction zone and the sensing region to thereby achieve multiple timepoint measurements. A single suction membrane is sufficient to control the flow of the fluid in the microfluidic device however multiple suction membranes can be included to further increase the control over the fluid flow or to facilitate the automation of the assay. A single suction membrane is sufficient to control the flow of a microfluidic chip having 3, 4, 8, 16, 24, 32, 64, or 96 separate channels.
The size of the suction membrane can vary based on the size of the microfluidic chip. More specifically the size of the volume defined inside the suction membrane is correlated to the total volume in the channels controlled by the suction membrane and in fluid communication therewith. Different shapes of the suction membrane are contemplated herein. For example, a hemisphere, an oval shaped hemisphere, a pyramidal shape, or cuboidal shape. Exemplary diameters for a hemisphere suction membrane are in the range of from 0.5 mm to 10 cm with a height of from 0.25 mm to 3 cm. Although a hemisphere is preferred, the function of the suction membrane is independent of its shape as long as the volume contained therein can be controlled by compression or relaxation of the suction membrane.
In the below exemplified embodiments, a mechanically actuated screw was used to perform the compression and relaxation on the suction membranes. However, other mechanical actuators are possible. For example, a piston like structure can be mechanically controlled to push or move away from the suction membrane. It is also contemplated that the mechanical actuation may be passive by the addition of weight on top of the piston or removing weight.
The utility of the suction membrane in controlling the fluid flow is not limited to any specific assay performed by the microfluidic chip. Preferred embodiments pair an assay that can be automated by detection not requiring human inputs with an automation of the control of the suction membrane. In the exemplified embodiments below, the detection is a colorimetric detection or an electrochemical/electric detection. However, other types of detection are possible and broadly spectroscopies are applicable. For example, surface-enhanced Raman spectroscopy (SERS) can be used to detect an analyte using microfluidic chips. In one example, the SERS outputs a label-free spectroscopic finger print for extracellular vesicles (EV) molecular profiling. The SERS is associated with a microfluidic device having an embedded arrayed nanocavity microchips that can detect EV. In one example, an embedded MoS2 monolayer can be used to enable the label-free isolation and nanoconfinement of single EVs due to physical interaction (Coulomband van der Waals) between the MoS2 edge sites and the lipid bilayer; and a layered plasmonic cavity that enables sufficient electromagnetic field enhancement inside the cavities.
In some embodiments, the platform is a colorimetric assay via plasmonic excitation. An opaque metallic nanostructured plasmonic platform is sensitive to characteristics of incident light such as its intensity, spectral profile, uniformity, and the angle of incidence. An epi-illumination imaging setup is preferably used to offer better control over illumination and imaging modalities.
One example of colorimetric techniques is the sensing of nucleic acid sequences using a nucleic acid amplification assay and a colorimetric sensor that changes color based on the amplification. This is particularly useful for the detection of pathogens. While polymerase chain reaction (PCR) technique remains the gold standard technique, isothermal nucleic acid amplification techniques (NAAT), especially Loop-mediated isothermal amplification (LAMP) have gained traction. Several advantages of LAMP include a requirement for constant temperature for amplification (instead of a cycle as per PCR), higher specificity and sensitivity compared to conventional methods, and stability against some amplification inhibitors. Among different colorimetric techniques, naked-eye/dye-based readouts are suitable for integration with LAMP. This technique also allows for easy integration with lab on a chip (LOC) platforms for point of care/need applications, as they require simple imaging setups allowing easy interpretation.
A major feature of a point of care/need diagnostic system is the ability to integrate all of the discreet assay steps in a fashion more suitable for point of care/need settings. In the case of nucleic acid amplification assays, these steps are sample collection, sample processing, reagent mixing, amplification reaction, and detection sub-steps. In addition to this, the sequential nature of a typical RT-LAMP assay, i.e. pathogen lysis, metering of sample and reagents followed by controlled heating for amplification, necessitate not only precise but also a facile setup for the end user in a point of care/need setting. To address these broad needs, a microfluidic setup was developed to implement techniques that allow metering and precise control of fluids and heat transfer, onto a single platform. Indeed, confined reaction volumes allow for fast and high throughput analysis, and enhanced heat transfer. Although, previous research on colorimetric readout-based pathogen detection platforms utilized microfluidic setups to integrate major assay sub-steps i.e., sample collection, sample processing, reagent mixing, amplification reaction, and detection, they often involved user involvement in one of more of these sub-steps, especially in sample collection, sample preprocessing, and/or fluid manipulation steps.
Another significant feature of microfluidic setups is their propensity to incorporate auxiliary components that could help reduce user involvement while achieving features suitable for point of care settings. Specifically, there has been growing interest in leveraging additive manufacturing techniques to fabricate auxiliary components that could offset the need for expensive equipment and/or trained personnel. For colorimetric detection assays, in both 3D printing leveraged and conventional integrated microfluidic systems alike, most of the current technologies lack user independence and autonomy in processing one or more of the sequential assay steps, especially in sample collection, sample preprocessing, and/or fluid manipulation steps.
Making reference to
Now making reference to
The incubation chamber 103 may receive/contain the mixed (or unmixed) sample. When the sample is contained in the incubation chamber 103 with the colorimetric sensor, the sample may be subjected to conditions that promote the colorimetric reaction to change the color of the sample. For example, heating may be applied.
As shown, a barrier 104, which may also be referred to as a barrier filter, is located downstream of the incubation chamber inlet. The barrier 104 may restrict the flow (e.g., reduce the flow rate) from exiting the incubation chamber 103. In the embodiment shown, the incubation chamber 103 includes the barrier 104. After incubation, the sample is pushed (e.g., by pressure differential) through the barrier 104 and into the sensing chamber 105. The barrier 104 may occupy a portion of the incubation chamber 103. The proportion of the incubation chamber 103 occupied by the barrier 104 may vary depending on the embodiments. In some embodiments, the barrier 104 may take up less than 50% of a total volume of the incubation chamber 103. Other proportions, i.e., between 50% and 70% could also be contemplated. The sample may be substantially or entirely retained in the incubation chamber 103 thanks to the barrier 104. In this context, the term “substantially” may be defined as having at least 80%, at least 85%, at least 90%, at least 95% or at least 99% of the volume of the sample being retained in the incubation chamber 103 for an incubation period. The barrier 104 may be enclosed within the incubation chamber 103 in a downstreammost portion of the incubation chamber 103. This is only one possibility. The barrier 104 may be located downstream of the incubation chamber 103, such as in a separate “filter chamber” fluidly connected to the incubation chamber 103 downstream thereof as another example.
The barrier 104 may entrap debris, undesirable particles, cells, such as bacteria, or other microscopic bodies before the flowing sample reaches the sensing chamber 105. In at least some embodiments, the barrier 104 includes an array of protruding microstructures. In an embodiment, the protruding microstructures are substantially cylindrical in shape (such as micropillars), however other shapes may be contemplated. The protruding microstructures are spaced apart to define a minimal distance between two protruding microstructure or between a protruding microstructure and a wall of the incubation chamber 103. The protruding microstructures may all have substantially the same size or there can be a variation in the size between the different protruding microstructures. The protruding microstructures can be arranged such that a flow of sample across the barrier 104 has to go through at least one minimal distance. This minimal distance can be referred to as the filter pore size. The minimal distance may be a range of values as the distance between two protruding microstructures may not be exactly constant. Solid particles, organisms or molecules that are larger than the minimal distance may not go through the barrier 104 and may thus be entrapped in the incubation chamber 103. For simplicity, the minimal distance between protruding microstructures of the barrier 104 will be referred to herein as the pore size of the barrier 104 filter. The pore size can be in the nanoscale range and can prevent the passage of bacteria and certain viruses, molecules and/or polymers. In some embodiments, the pore size is less than 1000 nm, less than 900 nm, less than 800 nm, less than 700 nm, less than 600 nm, less than 500 nm, less than 400 nm, between 100 nm and 1000 nm, between 200 and 900 nm, between 200 and 800 nm, between 200 and 700 nm, between 200 and 600 nm, between 200 and 500 nm or between 200 and 400 nm. In some embodiments, the protruding microstructures have a size of from 5 to 50 μm or from 5 to 40 μm. For example, the protruding microstructures can be micropillars having a diameter from 5 to 50 μm or from 5 to 40 μm. The spacing between the micropillars is preferably sufficiently large so as to allow a substantially laminar flow of the sample across the barrier 104. In some embodiments, the barrier 104 comprises at least two rows of protruding microstructures spaced apart so as to define a minimal distance being the pore size as described above. In some embodiments, the barrier 104 can comprise at least 3, at least 4, at least 5, at least 6, from 2 to 20, from 2 to 15, or from 2 to 10 of rows of protruding microstructure.
As shown in
The sensing chamber 105 has a plasmonic nanosurface, which may be referred to as plasmonic substrate, to allow increased sensitivity to the change in color of the colorimetric sensor. The color-generation strategy of plasmonic color printing involves the patterning of various geometrical metallic nanostructures. The nanostructures have a size that is smaller than that of the diffraction limit of light. The nanostructures and materials thereof are designed to resonate at a specific optical frequency leading to the production of different colors across the visible spectrum. The nanostructures act as nanoantennas when exposed to an electromagnetic field of light that resonate to increase a color gamut between the changes in color of the colorimetric sensor.
The nanostructures protrude from the nanosurface. The nanostructures may be in the shape of nanodisks, ellipses, nanocubes and/or multimers. In one embodiment, the nanostructures may have a diameter between 200 nm and 1000 nm. The interparticle gap spacing (gap between adjacent protruding nanostructures) can be considered to act as a plasmonic nanocavity. The interparticle gap length can for example vary from 20 nm to 500 nm (±5 nm or ±5%) or 73 nm and 340 nm (±5 nm or ±5%) with increasing nanostructures diameter (or maximum transverse dimension if diameter does not apply). With decreasing color change in the assay, smaller gaps (below 200 nm) generate more enhanced electromagnetic field. In a least some embodiments, the nanostructures have a plasmon-supported metallic layer with the ability to provide tunable localized surface plasmon resonance. The plasmon-supported metallic layer may provide high-resolution plasmonic color with a white background. For example, the plasmon-supported metallic surface is one of gold, silver, and aluminum or alloys thereof such as AuAg, AuAl, and AgAl. The plasmon-supported metallic layer may also be a bimetallic of Au, Ag, and Al such as AuPd, AgPd, AuNi, AuCu, AgCu, AuNiCu. In at least some embodiments the plasmon-supported metallic layer is a layer of 5 nm to 100 nm, 10 to 50 nm, 15 to 25 nm or 10 to 25 nm. Under the plasmon-supported metallic layer the nanostructures may have a back reflector layer. In some embodiments, the back reflector layer has a thickness from 10 nm to 400 nm, 50 nm to 130 nm or from 60 nm to 120 nm. The back reflector layer covers nanoparticles that are deposited on the surface of the sensing chamber 105. The back reflector material may advantageously be deposited using sputtering, ebeam deposition and/or spin coating. In one embodiment the back reflector layer comprises or consists of one of ZnO, TiO2, hydrogen silsesquioxane (HSQ), AZ MiRTM, or polymethyl methacrylate (PMMA). In one example the nanoparticles are made of polystyrene and have a diameter of between 200 to 1000 nm. The nanoparticles may be in any suitable shape, preferably a spheroidal shape such as a sphere. Unlike organic-dye color filters, the plasmonic color filters may offer advantages such as high color tunability, sensitive color changing based on medium permittivity and low color degradation rate.
Referring to
As illustrated in
The microchannel 102c′ defines a mixing zone between the incubation chamber 103′ and the sensing chamber 105′ adapted to provide sufficient mixing of the sample with the colorimetric sensor and sufficient time for the colorimetric reaction to occur at a predetermined flow rate. As it is being mixed, the flowing sample may change color before reaching the sensing chamber 105′. The mixing zone, also referred to as mixing channel, may be a serpentine or any other suitable shape as described above, for example with or without mixing sub-chambers, one or more microchannels 102c′, etc. In an embodiment, the walls of the microchannel 102c′ have an outline adapted to promote mixing. Mixing may be performed by creating a turbulent flow. In an embodiment, such as shown, the side walls of the microchannel 102c′ have a toothed outline. The toothed outline may define a series of arrows or serially distributed triangularly shaped sections. Other outlines may be contemplated, such as an outline defining discontinuities, bends, waves, or irregular patterns, for example. As another mixing parameter, surface roughness of the walls may also contribute to the mixing. Other configurations may be contemplated. While such toothed outline is described with reference to microchannel 102c′, it should be understood that any one of the microchannels (102a, 102b, 102c, 102a′ and 102b′) identified may have such configurations described with respect to microchannel 102c′.
Returning to
Although it is preferred that the colorimetric sensor is provided autonomously (e.g. without human intervention) by the microfluidic cartridge, it is possible that the sample is prepared before being provided to the microfluidic chip. For example the sample can be suspended, purified, filtered, or centrifuged after being collected. However, it is an objective of the present colorimetric detection system to minimize the steps taken by individuals to increase autonomy and to decrease human induced errors and inconsistencies. Accordingly, it is an advantage of the present system to provide the colorimetric sensor from the microfluidic cartridge 10 as described below.
The colorimetric sensor may be any appropriate colorimetric sensor that is specific to an analyte of interest so that the colorimetric sensor changes the color of the sample if the analyte is present. The colorimetric sensor may be a salt that reacts with a metabolic enzyme such as the NADH/NADPH cellular oxidoreductase enzymes. Examples of such salts include resazurin (metabolized to resorufin) and tetrazolium salts that are metabolized to formazan due to the disruption of the tetrazole ring. The tetrazolium salt can be selected from the group consisting of MTT (3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyltetrazolium bromide), XTT (2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide inner salt), MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethylphenyl)-2-(4-sulfophenyl)-2H-tetrazolium), and WST (water-soluble tetrazolium salts). The colorimetric sensor may be a 3,3′,5,5′-Tetramethylbenzidine or TMB which is a chromogenic substrate that can used in staining procedures in immunohistochemistry as well as a visualizing reagent exploited in enzyme-linked immunosorbent assays (ELISA). TMB is a white solid that forms a pale blue-green liquid in solution with ethyl acetate. The colorimetric sensor may be a pH sensitive dye in the fluid media that changes color when the pH changes. For example the colorimetric sensor may be phenol red, methyl blue, bromothymol blue, p-nitrophenol (formed by alkaline phosphatase from p-nitrophenol phosphate) and other similar colorimetric acid-base indicators. The colorimetric sensor can also be a H2O2 sensitive media that changes color with the concentration of H2O2 such as iodide or titanium based H2O2 indicators and the Amplex™ Red reagent which reacts with H2O2 to produce the red-fluorescent oxidation product, resorufin. In addition, color-sensitive nanoparticles may be used, including Au, Ag, AuPd and other nanoparticles possessing colors in the range of red to the blue. In some embodiments, the sample is selected from oral fluid, sputum, urine, tears, blood, plasma, nasal fluid, sweat, cerebral spinal fluid, suspended cells or microorganisms.
The inlet apparatus contains a storage chamber having a rupturable membrane and housing a colorimetric sensor as described above. The storage chamber 23 is in fluid communication with the microfluidic chip upstream of the sensing chamber 105, and preferably upstream of the mixing channel 102c. In some embodiments, the storage chamber 23 can fluidly connect to the microfluidic chip 100 before the incubation chamber 103 and barrier 104, in embodiments where it is desired to incubate the colorimetric sensor with the sample. One such example is the MTT assay where the objective is to determine and optionally quantify whether a cell population in the sample is alive or dead. In that example, the colorimetric sensor (i.e. MTT) can be incubated in the incubation chamber 103 with the cells and after the incubation the colorimetric sensor passes through the barrier 104 to get to the sensing chamber 105 but not the cells which are retained at the barrier. In other embodiments, the colorimetric sensor is provided downstream of the barrier. One such example is when the analyte is a nucleic acid sequence and the sample has to first be lysed to release the nucleic acid content therein. In that embodiment, the cell debris may be retained by the barrier 104 while the nucleic acid sequences go through and are then mixed with the colorimetric sensor in the mixing channel 102c.
In one particular example, the sample (e.g. saliva) is provided without any treatment directly into the receptacle 21. In such embodiments, the inlet apparatus 20 has a storage chamber containing lysis reagents 22. However, lysis may not be necessary in some cases as that depends on the analyte and the sample. Lysis is generally needed when the analyte is a nucleic acid sequence. In such embodiments, the storage chamber 23 can further contain the necessary reagents to perform a nucleic acid amplification (e.g. polymerase chain reaction (PCR), reverse transcription loop-mediated isothermal amplification (RT-LAMP) or a rolling circle amplification (RCA)). RT-LAMP and RCA are the preferred amplification techniques for a rapid screening. The reagents for a nucleic acid amplification include the primers targeting the analyte (forward and reverse), nucleotide and an appropriate DNA polymerase.
The inlet apparatus operates autonomously thanks to actuators. In embodiments where the lysis chamber 22 is present in the microfluidic cartridge 10, a piercing actuator 41 can be used to pierce a rupturable membrane of the lysis chamber 22 in order to release the contents of the lysis chamber 22 and bring them into contact with the sample to begin lysis. The lysis can include heating to a temperature of 85 to 98° C. for example around 95° C. This can be achieved by using a heating actuator 42 having a heating element 43. Finally, to release the contents of the storage chamber 23, another piercing actuator 44 can be used to rupture a rupturable membrane of the storage chamber 23 to release its content into the channels of the microfluidic chip 100. The actuators 41, 42 and 44 are connected to an actuating motor 40 which mechanically activates the actuators 41, 42 and 44. The actuating motor 40 can be connected to a controller or is coupled to a processing device 60. In some cases, the actuating motor 40 is connected to a controller and the controller is coupled to the processing device 60.
The flow in the microfluidic chip 100 is driven by a negative pressure created by two screws 31, 32 controlling the compression on the suction membrane in the outlet apparatus. The screws 31, 32 are released by actuators 45, 46 which are also connected to the actuating motor 40. It would be most economical and simplest to have the same motor 40 actuate all the actuators, however, it is of course possible to use multiple motors. To drive the flow of the sample to the incubation chamber 103, the first suction actuator 45 releases the first suction membrane which creates a negative pressure to drive the flow. After incubation, to drive the flow of the sample past the barrier 104 and to drive the flow of the contents of the storage chamber 23 if they are provided after the barrier 104, the second suction actuator 46 releases the second screw 32 to drive the flow and optionally the mixing in the microchannel 102c. The autonomous functioning of the present system is described in more detail further below, particularly in the example section.
When the sample reaches the incubation chamber 105, the light detection is performed with a light detection apparatus 50. An advantage of the present system is that a brightfield microscope is not needed and a more compact system is employed. A condensing lens 51, for example an objective, is used to condense the light detected from the sensing chamber 105 of the microfluidic chip 100. A light source 52 is provided to perform an epi illumination on the sensing chamber 105. This means that the light of the light source 52 is provided on the same side as the condensing lens 51. Preferably, the light source 52 is an illumination column with a light emitting diode (LED). An image sensor 53, for example a camera or a CMOS sensor, receives the light condensed by the condensing lens 51. The image sensor can register the light received as an electronic signal and can send the electronic signal to the processing device 60. Accordingly, the image sensor 53 is coupled to the processing device 60. The condensing lens 51, the light source 52 and the image sensor 53 are all connected optically by appropriate means 54. In some embodiments, the microfluidic chip 100 can have more than one sensing chamber 105. In such embodiments, it is required to have the condensing lens 51 be able to move to be positioned above each of the sensing chamber 105. In these embodiments, the light detection apparatus can be connected to a motorized platform 55 which allows at least the lateral movement with respect to the microfluidic chip and in between the inlet apparatus and the outlet apparatus.
The processing device 60 is shown in
For example, and without limitation, the computing device 63 may be a server, network appliance, a set-top box, embedded device, computer expansion module, personal computer, laptop, personal data assistant, cellular telephone, smartphone device, UMPC tablets, video display terminal, gaming console, electronic reading device, and wireless hypermedia device or any other computing device capable of being configured to carry out the methods described herein.
Each processor 61 may be, for example, any type of general-purpose microprocessor or microcontroller, a digital signal processing (DSP) processor, an integrated circuit, a field programmable gate array (FPGA), a reconfigurable processor, a programmable read-only memory (PROM), or any combination thereof.
Memory 62 may include a suitable combination of any type of computer memory that is located either internally or externally such as, for example, random-access memory (RAM), read-only memory (ROM), compact disc read-only memory (CDROM), electro-optical memory, magneto-optical memory, erasable programmable read-only memory (EPROM), and electrically-erasable programmable read-only memory (EEPROM), Ferroelectric RAM (FRAM) or the like.
Each communication interface enables computing device 63 to interconnect with one or more input/output devices 67, such as a keyboard, mouse, camera, touch screen microphone, display screen and speaker. For example, a display screen may display a symbol or sign that is indicative of the presence or absence of the sample analyte in the sample. In one embodiment, the display screen displays the value of the concentration of the sample analyte in the sample. In a further embodiment, the display screen may display the estimated value of the concentration of the sample analyte in the sample of the subject who provided the sample. The display can be a simple display in black and white or a more modern touch screen able to receive commands. A user interface may contain a button or other physical means for the user to signal to the device to begin the analysis of a sample.
In some embodiments, a network interface enables the computing device 63 to communicate with other components, to exchange data with other components, to access and connect to network resources, to serve applications, and perform other computing applications by connecting to a network (or multiple networks) capable of carrying data including the Internet, Ethernet, plain old telephone service (POTS) line, public switch telephone network (PSTN), integrated services digital network (ISDN), digital subscriber line (DSL), coaxial cable, fiber optics, satellite, mobile, wireless (e.g. Wi-Fi, WiMAX), SS7 signaling network, fixed line, local area network, wide area network, and others, including any combination of these.
The computing device 63 can be operable to register and authenticate users (using a login, unique identifier, and password for example) prior to providing access to applications, a local network, network resources, other networks and network security devices. Computing device 63 may serve one user or multiple users.
In some embodiments, a machine learning algorithm is used to increase the accuracy of the detection. In such embodiments, the memory 62 can store the training data set and may continuously update the training data set. The memory 62 may have various parameters stored therein and can store the readings performed over time. Features of machine learning may include dimensionality reduction of the electrical signal using one of principal component analysis (PCA), locally linear embedding (LLE), multidimensional scaling (MDS), t-distributed stochastic neighbor embedding (t-SNE), and linear discriminant analysis (LDA). In further embodiments, the machine learning is configured to perform the classification task using one of logistic regression, soft Regression, decision Tree, random forest (RF), and an artificial neural network (ANN). For example, the machine learning algorithm is configured to perform the regression analysis using one of linear regression, gradient descent, polynomial regression, regularized linear model, ridge regression, lasso regression, and support vector machine (SVM).
Now making reference to
In some embodiments, as illustrated in
As illustrated in
As explained above and also in greater details below, the suction membrane allows for a bi-directional flow which can be leveraged to pass the sample from the incubation chamber 103 to the sensing chamber 105 back and forth multiple times for multiple timepoint measurements. As illustrated in
In one exemplary embodiment, the automated setup of the present system comprises of two modules, a microfluidic module (named microfluidic cartridge 10) (
A true point of care colorimetric nucleic acid testing device, analogous to qPCR, that employs RT-LAMP and RCA to amplify nucleic acid (RNA and DNA) in one step at a steady temperature was achieved which quantifies the presence of nucleic acid biomarkers via the color reading of the media. Unlike most commercialized technologies and those under research that employ fluorescent read-outs, the present system can combine a label free and easily adaptable colorimetric read-out with a multiplexed microfluidic sample preparation and delivery system. The core principle of the colorimetric readout is based on the plasmonic induced enhancement of nucleic acid amplification generated on the surface of a novel integrated nanostructured fluidic platform, which provides ultrasensitive and quantitative detection of low concentrations of nucleic acids in non-manipulated body fluids. When subjected to light, plasmonic surfaces inject electrons into the assay media resulting in plasmonic catalysis and enhanced amplification rate (less than 10 minutes). This plasmonic amplified color change of the media is observable with the help of a bright field microscope or simply the naked eye, obviating the need for complex equipment or trained personnel. The integration of the nano-plasmonic components with microfluidic sample preparation and delivery allows for multiplexed testing, limited sample/reagent consumption, and ease of automation and operation by untrained personnel.
The fabrication of the microfluidic cartridge is described in detail in the Example section below. Nevertheless, in brief, first, a lithography step is carried out to transfer the heater (e.g. width: 400 μm), and pad (e.g. length: 5 mm, width: 2 mm) features to a photoresist layer through a photomask with the desired patterns. This is followed by a buffered oxide etch to remove native oxide and a potassium hydroxide etch for a 200 nm silicon etch. Next, a lift-off process for selective deposition of the heater elements in the etched grooves is carried out. This starts with a second lithography step followed by an electron-beam deposition of for example 240 nm aluminum. Accordingly, the lift-off is completed by submersion in suitable remover. Next, the last lithography step is carried out to pattern the fluidic device features including inlet/outlet ports (e.g. ¢ 2 mm), lysis chamber (e.g. length: 1.74 mm, width: 1.5 mm, depth: 50 μm), mixing channels (e.g. width: 200 μm, height: 50 μm), and plasmonic window (e.g. length: 1.94 mm, width: 1.5 mm, height: 50 μm) into a SU-8 layer (for example SU-8 2050). Subsequently, the wafer (e.g. 6-inch) is diced into individual chips (e.g. length: 26.5 mm, width: 35 mm) using a dicing saw.
The pathogen detection system developed in the present example has three major steps, (i) sample collection from the user, (ii) amplification assay, and (iii) image capture and data analysis. In the first step, the user spits into the saliva collection funnel (as shown in
The portable reflected-light imaging setup with controlled epi-illumination (PRICE) has two main important modules, (i) an illumination module and (ii) an image capture module. To implement the illumination module, a Koehler illumination optical train was mimicked with off-the-shelf optics. A 5000K 90CRI LUXEON™ LED (Lumileds™ Inc.) was used as the primary illumination source. To collimate the LED, a diffusive aspheric condenser lens (d=25.4 mm, f=20.1 mm, Thorlabs™) was used. A ring-actuated iris diaphragm (Thorlabs™) was used as a field diaphragm with aperture diameters ranging from 8 mm (minimum aperture opening) to 12 mm (maximum aperture opening). The collimated light was then illuminated on the back of an achromatic doublet lens (d=25.4 mm, f=30 mm, Thorlabs™), creating an imaging at the focal length of the lens. A ring-actuated iris diaphragm (Thorlabs™) was used as an aperture diaphragm with aperture diameters ranging from 8 mm (minimum aperture opening) to 12 mm (maximum aperture opening). Finally, the image of the light at the aperture diaphragm is collimated by an achromatic doublet lens (d=25.4 mm, f=30 mm, Thorlabs™). The collimated light is then projected onto the back aperture of the objective (TU Plan Fluor EPI 20×, N.A. 0.45, W.D. 4.5 mm, Nikon™ Inc.) via a beamsplitter (Reflectance: Transmittance—30:70, d=25.4 mm, Thorlabs™) placed at an angle of 45 degrees with the vertical. The reflected light from the sample placed the working distance was collected by the objective. Since the objective was infinity focused, a condenser lens (tube lens, f=200 mm) was used to project the image onto the CMOS sensor (Sony™ IMX477R, 12.3 MP, Raspberry Pi Inc.).
The microfluidic cartridge of the present example has two major modules, (i) a cleanroom fabricated microfluidic chip and (ii) a 3D printed fluid handling attachment. Together these two modules enable the integration of sample collection, sample lysis, reagent mixing and amplification steps as a single platform. All the microfluidic chip and the cartridge components were designed using AutoCAD™ and SolidWorks™ software. The fabrication process 300 is shown in
In this example, 3D printed molds were used for the fabrication of suction cups with PDMS. SolidWorks™ software was used to design the master mold. A high resolution SLA 3D printing (Form 3, Formlabs™, USA) was employed at a layer thickness of 25 μm in the z-axis as per the company's specifications. The post-printing treatment included a wash in isopropyl alcohol (IPA) for 20 min followed by drying and ultra-violet (UV) curing at 60° C. for 20 min. Once the curing was done, the supports were removed. The master mold had two components, male and female molds. Before proceeding to pour PDMS for molding, the 3D printed parts were surface treated to avoid curing inhibition at the PDMS-mold interface. The fabrication of the suction cups is shown in
All the 3D printed components were designed with SolidWorks™ software and fabricated using a SLA 3D printer (Form 3, Formlabs™, USA) at a layer thickness of 50 μm in the z-axis as per the company's specifications. The post-printing treatment included a wash in isopropyl alcohol (IPA) for 20 min followed by drying and UV curing at 60° C. for 20 min. Once the curing was done, the supports were removed. Following this, the collection funnel, lysis funnel, and reagent storage chambers housed in the 3D printed fluid handling attachment, were coated with epoxy resin (Artresin™, Canada) to make the surface biocompatible. Finally, a hollow biocompatible aluminum insert (McMaster-Carr™ Inc, Canada) was incorporated into the 3D printed fluid handling attachment as the lysis chamber.
The automation module had three main components, (i) Linear actuator system, (ii) Heating module, (iii) x-y translation stage and (iv) Imaging and data processing module. Two main microcontrollers were used here, Arduino UNO (Arduino Inc.) and Raspberry Pi 4 (Raspberry Inc.). The Arduino was employed to control components (i), (ii), (iii), (iv) and the Raspberry Pi controlled the imaging and data processing module. Five linear actuators (Actuonix™ Inc.) were employed to facilitate sequential fluid handling steps. This detection platform requires heating for the lysis and amplification steps of the assay, to facilitate this employed two heating components, a portable solder iron (TS-100) for saliva lysis and ceramic thermal heater (Bolsen™ Tech Inc.) for LAMP reaction at 65° C. The two heating elements were controlled with a two-channel relay module (Yizhet™ Inc.). As previously mentioned, here a multiplexed approach with three detection chambers was used. Hence it was needed to have a x-y translation stage to scan and image across multiple intra-chambers as well as inter-chamber regions. To realize this a linearly guided computer numerical control (CNC) stage that was driven by a stepper motor (FUYU™ Inc.) was employed for stage movement in x-direction, with a guide length of 50 mm and a resolution of 0.25 mm. To move the chip in y-direction a single axis manual translation stage was used (Edmund Optics™ Inc.) attached to a 3D printed holder that firms the microfluidic cartridge in place. The screw head of this y-axis translation component was controlled by a knob attached to a continuous servo (SPT digital, SPT5325LV). The image once captured by the CMOS sensor was captured and analyzed on-board the Raspberry Pi 4. Finally, the results were transmitted and communicated via a mobile application created with MIT App inventor.
The RT-LAMP assay employed in this example used the primer set against the ORF1ab gene obtained from Sigma-Aldrich™, USA. The individual oligonucleotide concentrations of the 10× primer mix were, 0.2 μmol/L of forward outer primer, 1.6 μmol/L of forward inner primer and backward inner primer, 0.4 μmol/L of forward loop primer and backward loop primer. The primer mix was mixed with WarmStart™ Colorimetric LAMP 2× Master Mix (NewEngland Biolabs™, MA, USA), and RNase free water (Thermo Fischer Scientific™, MA, USA). The standard reaction volume was 20 μL that consisted of 2 μL 10× primer mix, 10 μL 2× master mix, 7 μL RNase free water, and 1 μL synthetic SARS-COV2 RNA (VR-3276SD ATCC, VA, USA) sample. The samples were incubated at 65° C. for different periods to visualize color change versus time for different samples.
A pressure driven flow induced by suction cups was used to enable fluid flow with minimal user involvement. The suction cups act as actuators on the flexible membrane to control the air pressure and thus the fluid flow in the microfluidic channels. The key features of an amplification assay are, (a) discrete sequential steps involved in the assay, and (b) metering of fluids involved (sample and reagents). To expand, the sequential steps in a typical amplification assay (i) sample collection, (ii) sample lysis, (iii) mixing with amplification reagents (in a fixed volumetric ratio), (iv) amplification reaction, and (v) end point detection. Three key technologies were leveraged to enable integration of all these sequential steps on to a single microfluidic cartridge, they are, (1) mechanically actuated suction cups, (2) microfabrication, and (3) additive manufacturing. The suction cups facilitate power free, userless fluid flow manipulation via mechanical actuating with a 3D printed screw-nut like setup. Microfabricated silicon-based chip housed the channels and chambers required for mixing, amplification, and detection steps. Additive manufacturing (i.e. 3D printing) enabled fabrication of a 3D printed fluid handling module that is coupled with the silicon microchip. This fluid handling module houses sample collection, sample lysis, reagent storage and suction cup actuation components.
One of the key technologies in the integration of all the assay steps onto a microfluidic cartridge is microfabrication. SU-8 microchannels were used to enable fluid handling. COMSOL™ simulation methods were employed for designing the microfluidic chip. One key requirement was the mixing of lysed saliva sample with the reagents in the volumetric ratio of 1:10. To execute this, a Y-junction mixing followed by a serpentine channel was used to ensure complete mixing. The area of cross section of all the channels was the same, the channel widths were manipulated to ensure a flow rate ratio of 1:10. The right combination of channel widths was arrived at by using 2D COMSOL™ simulation, modeled at creeping flow conditions. The central idea was that the area of cross section of the channels was the same and the widths were manipulated to change the overall volumetric flow rate. The change of widths ultimately effects the distribution of the negative pressure produced by the suction release. Given the multiplexed nature of the proposed system the lysed sample should be distributed into the channels leading up to mixing module. Considering the design requirements for the automation, a pressure distribution simulation was performed to ensure uniform pressure distribution (
The PDMS based suction cups were employed to drive the fluid flow with no user involvement or expensive external equipment (like syringe pump, peristaltic pump). The suction cups and the membrane work on the principle of air displacement giving rise to a negative pressure in the microfluidic channels thereby inducing fluid flow along the pressure gradient.
The volume of liquid that can be moved/pumped by the suction cup and membrane is dependent on the volume of air displaced upon compression of the membrane. A mechanical actuation system with a nut and screw mechanism was used herein to apply the pressure on the membrane. However, other actuation systems would be also suitable and the present disclosure is not limited to a nut and screw mechanism. In the context of fluid flow in microchannels, it is important to employ suction cups and membranes that can precisely pump desired volumes. However, the mechanical actuation system is prone to having residual volume when the cup is pressed down. Hence, correlating the experimental suction membrane volume to the theoretical volume is important to account for the residual volume. Specifically, the effect of mechanical actuation in the current system on the residual volume. Initially the suction membrane is assumed to be a hemisphere, hence the initial volume (Vin) is given by
In equation 1, d is the diameter of the suction membrane 34 and h is the central height of the hemisphere (
It was further built upon the effect of mechanical actuation on the volume of fluid suctioned. It was hypothesized that the volume of fluid suctioned/pumped (Vp) can be varied by manipulating the deflection of the PDMS suction membrane (
Where V is the volume stored by the suction membrane, f is the deflection, r is the radius of the suction membrane (
Here E is the elastic modulus of the chamber material, h is the chamber PDMS membrane thickness, and p is the loading pressure. These equations hold true only when 2r>h, which is the case in the present suction membrane. From equations, (2), (3) and (4), it can be inferred that volume of the suction membrane (v) varies cubically with loading pressure (p) in the f≤h and varies non-linearly in the regime f>h. The experimental results correlate with the theoretical profile showed in
The delta volume suctioned is given by first order differential of the equation (4), as follows
This equation 5 shows that change in suction membrane volume (i.e. volume of fluid suctioned), is in quadratic relationship with the deflection of the membrane. The experimental results are in correlation with this since a parabolic profile was obtained. In other words, the angular control allows for precise manipulation of the fluid volume in the microchannel. By incorporating this screw-nut actuation system with a screw of pitch of 1.4 mm, it was demonstrated that the pumping of fluid volumes as low as 0.1 μL can be done (
The silicon based microfluidic chip was bonded to the 3D printed attachment to facilitate automation of all the assay steps. More specifically, the 3D printed attachment houses sample collection, sample lysis, reagent storage and suction membrane actuation components (
The sequential nature of the amplification assays requires the amplification reagents mixing with the sample only after the lysis. To facilitate this, the amplification reagents are stored on the cartridge in the storage chambers. The chambers are sealed off from atmospheric pressure, to impede the amplification reagents from entering the serpentine mixing channel before the flow of the lysed saliva to the y-shaped mixing junction. To realize this, high resolution of SLA 3D printer was leveraged to print membranes in the order of microns in thickness. These rupturable membranes, upon breakage, cause the atmospheric venting leading to the flow of the reagents to the Y-junction and subsequent mixing with lysed sample. In the final step of automation, this membrane breakage step is executed with a linear actuator with sharp tip. Hence the thickness of the membrane plays an important role, as the force required for breaking the membrane should be a value that can be executed by the tip of the linear actuator.
Rupturable membranes were fabricated at different thicknesses starting from 100 μm up to 600 μm. The rupturable membranes were then manually broken to evaluate the required force to inflict damage. It was found that an average force of 6.26N (
The extent of mixing module was evaluated by employing image analysis of images captured by a microscope. The whole cartridge operation shown in
A portable reflected-light imaging setup with controlled epi-illumination (PRICE) was designed to capture the colorimetric change of the assay solution. In the current scheme of SARS-CoV-2 detection, the current platform was dubbed “QolorEX” and it enhanced the colorimetric change via plasmonic excitation. QolorEX is a nanostructured platform fabricated in cleanroom with deposition of multiple layers of metals. This platform elicits structure and material dependent-color upon irradiance of white light, extensively studied in the past and termed as plasmonic based coloring. These metallic surfaces are sensitive to illumination properties which could ultimately affect the colorimetric readout. Hence it was important to ensure to have a uniform, consistent, and controlled illumination and subsequently the use of PRICE as a replacement to a commercial brightfield microscope. The PRICE has two main modules (i) illumination module and (ii) imaging module.
The light detection apparatus used in the present example is shown in
Ensuring uniform illumination is crucial for capturing accurate colorimetric reading. Having a controlled illumination column would also facilitate more control over the illumination parameters, specifically area of illumination and the intensity of illumination. One key issue with achieving uniform illumination is the diverging angle of irradiance of an LED. The 5000K LED employed in this example has a total viewing angle (defined as the angle from the LED centerline after which the luminous intensity drops to half of the maximum) of 60°. Among different illumination methods employed in standard microscopes, Koehler illumination is widely adopted owing to its illumination uniformity. Moreover, Koehler illumination setup is more resilient to external disturbances like dust and optical imperfections, making them suitable for imaging outside of controlled environments. These features also facilitate minimal variations in the images obtained by the camera over long and repeated cycles. Hence, a portable illumination column mimicking the Koehler setup would address the challenges with controlled illumination. Another point to note it that Koehler illumination is generally aimed at trans-illumination which works only for imaging translucent samples. In the present case, an epi-illumination is needed since the detection chamber is opaque in nature. The illumination column 902 is shown in
The source of illumination is thus an important property that can have predominant effect on the final color captured and observed. The platform used herein, QolorEX, was characterized using UV-Vis spectroscopy to obtain the resonance wavelength (
For the successful application of the current setup at the point of need, it is necessary to integrate PRICE (imaging setup) and operation of the cartridge in an automated fashion. In first, the operation of cartridge has sequential steps, with key steps that could require user involvement in absence of an automated setup. These key steps, include (i) breakage of the membrane, (ii) release of the suction cups (by twisting the screws holding them in place), (iii) heating for saliva and amplification steps. To automate these steps, a system of five linear actuators was employed (coded—A1, A2, A3, A4, A5) (as shown in
The actuators, A1 and A3 were attached with sharp ended tips that assist with breakage of the membrane with a force of 6.2N. Actuators, A4 and A5 were fitted with cylindrical shaped tips to push the extension of the screw and actuate the suction cups. The linear actuator A2 was fitted with a portable pocket solder to carry out the sample lysis step. It is desired to reach 95° C. for 3 min to complete the lysis process. The temperature was controlled from overshooting by employing cycles of linear movement of the actuators. The desired temperature was reached in under 1 min as represented by infrared images in
Once the amplification reaction was carried out for 15 min, it was desired to capture the calorimetry by PRICE. The imaging setup was completely controlled by Raspberry Pi 4 microcontroller. The PRICE worked in tandem with the Arduino microcontroller to scan different chambers and different regions in the chambers. This movement was enabled by an X-Y translation stage directly controlled by the Arduino. For the X-translation, a CNC linear stage was employed. For the Y-translation a unique contraption involving a linear manual actuation stage and continuous servo motor. The linear stage moves in a single direction with rotation of a screwhead. This screwhead is attached to the rotating shaft of the servo motor via a 3D printed knob. The cartridge holder is firmly fitted to this linear translation stage. The duration of rotation and the direction of rotation of the servo motor shaft therefore determines the movement of the cartridge. The entire setup is covered with a 3D printed enclosure with a sliding door for insertion of the cartridge (
This concerted interplay between (i) imaging (executed by Raspberry Pi 4), (ii) execution of fluidic steps (operation of cartridge) and movement of X-Y translation stage (executed by Arduino UNO) is centrally controlled by a mobile application. Both the Raspberry Pi and Arduino execute commands via Bluetooth commands (
The image of the assay on the platform was captured with the custom-built PRICE setup. In brief, the CMOS sensor captured the image of the detection platform. The raw image was stored in the raspberry Pi for further processing. In addition to controlled illumination, the parameters of image capture, namely, white balance, gain (analog and digital) exposure time, framerate and ISO become important. Among these, white balance and gain values were changed in this study. All the other parameters were fixed and left changed through the experimental process.
LAMP assay was carried out in Eppendorf tubes for wild type SARS-COV-2 synthetic RNA for different points. Then 1 μL of the droplet was transferred to the detection platform and imaged with the PRICE setup and brightfield Nikon microscope.
It was shown herein that an automated colorimetric setup can perform pathogen detection. The sequential steps of the assay and signal transduction and analysis are automated using three different modules, (i) a portable reflected-light imaging setup with controlled epi-illumination (PRICE), (ii) a microfluidic cartridge and (iii) automation control unit. The QolorEX platform, a specialized nanostructured platform herein developed leverages plasmonic excitation for highly sensitive detection of respiratory infectious pathogen is employed as the key detection technology. First, to capture the colorimetric change, PRICE is designed for imaging the assay chamber. The imaging setup offered superior spatial and spectral control with only a 17% variation in the relative intensity and a resolution and FoV of 4.4 μm and 298 μm, respectively. Next, to eliminate the involvement of the user, a microfluidic cartridge with mechanically actuated PDMS suction membranes is implemented by leveraging additive manufacturing techniques. The flow was shown to be mechanically actuated by a screw-nut mechanism with excellent control over the fluid pumping. This actuation mechanism demonstrated lowest volume of fluid suctioned at 0.1 μl for a 30 degree rotation of the actuating screw. Subsequently, the microfluidic chip also showed perfect extent in mixing lysed sample with the reagents. With the final automation and control module, the cartridge operation was concerted using system of linear actuators and electrothermal heaters connected via Arduino UNO and Raspberry Pi controlled via mobile application. The imaging was implemented in a direct comparison format with a Nikon brightfield microscope, and a quantifiable colorimetric change was recorded in 15 min.
Nucleic acid amplification is the gold standard molecular diagnostic test, but it is not easily deployable due to required lengthy protocols and specialized equipment. The point of care testing to date is limited by versatility and rapidity in reading the amplification signal. Additional testing was performed on the nanoplasmonically boosted nucleic acid amplification. The microfluidic sample collection/preparation to achieve fully automated minute-scale (sample-to answer time of 13 minutes) colorimetric detection of multiple nucleic acid biomarkers at single nucleotide resolution (QolorEX). Here, it is shown that one-step isothermal amplification of RNA/DNA with loop-mediated amplification (RT-LAMP) and rolling circle amplification (RCA) can be boosted three-folds via plasmonic color enhancement generated on the surface of plasmonic nanostructures confined in the microfluidics. This offers a label/probe-free colorimetric approach deployable for the detection of a variety of targets by simple tuning of amplification reagents with a quantitative response as a function of the pathogen load. The versatility of the QolorEX platform was demonstrated through the detection of respiratory viruses such as SARS-COV-2, and Influenza A H1N1, as well as antimicrobial-resistant bacteria such as Escherichia coli (E. coli) and Methicillin-resistant Staphylococcus aureus (MRSA). It was also possible to discriminate between SARS-COV-2 variants by incorporating RCA to detect viral RNA alternatives at the level of single nucleotide polymorphism. The diagnostic capability of QolorEX in clinical setting was also demonstrated by testing 33 saliva samples from COVID-19 patients and achieved quantitative detection of viral RNA in saliva with a detection limit of 5 RNA copies/μL and 95% accuracy on par with qPCR. The simplicity, sensitivity, and robustness of QolorEX's technology is an advantageous platform for the realtime monitoring of pathogenic infections and assist in clinical decision-making.
Accordingly, the QolorEX platform was tested for the detection of Influenza A H1N1, SARS-COV-2 and its Alpha B.1.1.7, Delta B.1.617.2, Gama P.1, Eta B.1.525 and Omicron B.1.1.529 variants. The results with qPCR. To demonstrate the universal application of QolorEX, in addition to the investigations circumambient RNA amplification testing, DNA amplification detection in bacterial infections based on E. coli and MRSA were also studied. For clinical validation, 33 saliva samples from COVID-19 patients and 15 healthy samples were tested with QolorEX. The QolorEX device meets all the ASSURED (affordable, sensitive, specific, userfriendly, rapid, equipment-free, delivered) criteria laid out by World Health Organization (WHO) for developing an ideal biosensor in both developed and developing countries.
The list of materials used is as follows: polystyrene nanobeads (Polystyrene particles (PS-R); Micro Particle GmbH), synthetic SARS-COV-2 RNA (ATCC VR-3276SD; Cedarlane™), SARS-COV-2 B.1.1.7 Alpha variant RNA (ATCC VR-3326D; Cedarlane), MERS-COV (ATCC VR-3248SD; Cedarlane), heat-inactivated SARS-COV-2 (ATCC VR-1986HK; Cedarlane), heat-inactivated Influenza A (0810248CFHI; Cedarlane™), E. coli (#211540, Merlan™ Scientific), colorimetric RT-LAMP master mix and HiFi Taq DNA ligase enzyme (NewEngland Biolabs™), PLPs, RCA primers, and synthetic cDNA SARS-COV-2 targets, MgCl2, KCl, NAD, and Triton X-100 (Sigma Aldrich), LAMP primers, DTT (ThermoFisher Scientific), healthy human pooled saliva (IRHUSL50ML), and healthy human single donor saliva (IRHUSLS5ML) were bought from Innovative Research and stored upon arrival at −80° C., SARS-COV-2 variants (Omicron, Delta, Eta, and Gamma) RNA were obtained from cooperator laboratory at McGill University (Vidal Lab). HCOV 229E RNA was obtained from cooperator laboratory at Lady Davis Institute for Medical Research in Jewish General Hospital (Chen Liang Lab), MRSA DNA was obtained from cooperator laboratory at McGill University Health Centers—MUHC—research institute (Dao Nguyen Lab). All assays were prepared using ultra-pure DNase/RNase-free distilled water (ThermoFisher™ Scientific). Primers' composition and concentrations, and the RCA probes are presented in Tables 1-8.
E. coli LAMP Primer Set (targeting malB Gene)
The SARS-COV-2 RT-LAMP primers used herein to target the ORF1ab gene. H1N1 RT-LAMP primers were designed through the NEB LAMP Primer Design Tool targeting highly conserved sequences of the Hemagglutinin (HA) gene of the H1N1 influenza A virus. The LAMP recognition part was evaluated by Basic Local Alignment Search Tool (BLAST) and no sequence variations have been seen in 100 hits provided by the NCBI website. The nominated LAMP primers were selected according to the optimized parameters offered by the Primer Explorer V5 protocol.
H1N1 RT-LAMP primers were designed through the NEB LAMP Primer Design Tool targeting highly conserved sequences of the Hemagglutinin (HA) gene of the H1N1 influenza A virus. The LAMP recognition part was evaluated by Basic Local Alignment Search Tool (BLAST) and no sequence variations have been seen in 100 hits provided by the NCBI website. The nominated LAMP primers were selected according to the optimized parameters offered by the Primer Explorer V5 protocol.
The P681H and L452R mutation sites were determined according to CoV-GLUE-Viz and GISAID and PLPs recognition parts were designed to be specific to the mutations. In addition, the PLP target site for WT SARS-COV-2 detection was the ORF1ab gene which is the same as the targeting site for SARS-COV-2 LAMP primers. The PLPs are designed in a way that the SNP is distinguished by the upstream of PLP which is providing the 3′-hydroxyl group at the ligation junction base-paired next to the phosphorylated 5′ end on a target strand. The veracity of PLP circularization upon SNP detection has been determined using Thermostable Ligase Reaction Temperature Calculator provided by New England Biolabs (NEB) website. Thereafter, the PLPs were evaluated using the Mfold web server to avoid undesirable secondary structure, especially in the PLP recognition site. RCA forward and reverse primers were designed to be hybridized to the spacer part of PLPs that connects two specific arms of PLP altogether. PLPs selectivity was confirmed through a gel electrophoresis experiment.
For RT-LAMP assays a standard reaction volume of 20 μL was used. It contained 2 μL 10× primer mix, 10 μL 2× master mix, 7 μL Rnase-free water, and 1 μL RNA sample. Heat-inactivated viral samples were first thermally lysed at 95° C. for 3 minutes and then mixed with the assay. This was followed by incubation at 65° C. for different periods to visualize color change versus time for different samples. The PLP ligation reaction was performed in a final volume of 10 μL including 1 μL synthetic complementary DNA (cDNA) of SARS-COV-2 RNA genome, 1 μL of 1 μM PLP, 2 μL UltraPure Distilled Water and 5 μL 2× no-Tris-HCl HiFi Taq DNA ligase ligation solution (20 mM MgCl2, 20 mM KCl, 2 mM NAD, 0.1% Triton X-100, 20 mM DTT, pH 8.50) and 1 μL HiFi Taq DNA ligase enzyme. The ligation mixture was first incubated at 95° C. for 5 min for DNA denaturation and then cooled down to PLPs annealing temperature (which are 60, 58, 55° C. for P681H, L452R, and WT PLPs, respectively) to let PLPs hybridize with cDNA and ligate via HiFi enzyme in a thermocycler (Analytik Jena, Germany). Thereafter, the ligation reaction was 12 μL of WarmStart™ Colorimetric LAMP 2× Master Mix, 1.6 μM RCA reverse, and forward primers in a final volume of 24 μL. The RCA amplification reaction was performed at 65° C. for different periods to visualize color change versus time for different samples.
For bacterial DNA Extraction, E. coli samples were cultured overnight at 37° C. in Luria Broth (LB) media. Bacteria concentration was determined using a Spectronic™ 21D spectrophotometer. Aliquots of different concentrations of 107 CFU·mL−1, 105 CFU·mL−1, 104 CFU·mL−1, 103 CFU·mL−1, 102 CFU·mL−1, and 10 CFU·mL−1 were prepared by suspending E. coli cultures in LB media. E. coli DNA was extracted by boiling cultures at 95° C. for 10 min. Methicillin-resistant S. aureus DNA was obtained from the McGill University Health Centre using the chemical lysis method. All DNA sample concentrations were measured using a Nanodrop™ 2000 Spectrophotometer and suspended in Universal Buffer 48 (Bio Basic Inc., ON, CA) to achieve desired concentrations.
QolorEXLAMP assay For SARS-COV-2 tested spiked solutions (RNase-free water and healthy saliva) of 8×105 RNA copies. μl−1 —5 RNA copies. μl−1 of SARS-COV-2 RNA and 90 PFU. μl−1—0.01 PFU. μl−1 for heat-inactivated SARS-COV-2 to fit in a biologically relevant range. Similarly, a study was performed for Delta B.1.617.2, Omicron B.1.1.529, Eta B.1.525, and Gamma P.1 variants RNA with the concentration of 8×105 RNA copies. μl−1. Alpha B. 1.1.7 variant was studied with a concentration of 104 RNA copies. μl−1. For selectivity studies, RNA from MERS CoV, HCoV, and Influenza A H1N1 viruses were tested.
QolorEXLAMP assay for H1N1 studied spiked solutions (RNase-free water and healthy saliva) of 8×105 RNA copies. μl−1 —5 RNA copies. μl−1 of Influenza A H1N1 RNA. For selectivity studies, RNA from multiple viruses (SARS-COV-2, MERS CoV, and HCoV 229E) were tested at concentrations of 8×105 RNA copies. μl−1.
QolorEXLAMP assay For E. coli studied spiked solutions (RNase-free water) of 70 ng. μl−1—0.2 ng·μl−1 of E. coli DNA. For selectivity studies, DNA from multiple bacteria (E. coli, MRSA, and Pseudomonas Aeruginosa) were tested at the concentration of 50 ng·μl−1.
QolorEXLAMP assay For MRSA studied spiked solutions (RNase-free water) of 50 ng·μl−1—0.2 ng·μl−1 of MRSA DNA. For selectivity studies, DNA from multiple bacteria (MRSA, E. coli, and Pseudomonas Aeruginosa) were tested at the concentration of 50 ng·μl−1.
QolorEXRCA assay was done in spiked solutions (RNase-free water and healthy saliva) of 106 cDNA copies·μL−1—5 cDNA copies·μL−1 of synthetic cDNA of P681H, L452R, and WT SARSCOV-2 sequences. For selectivity, the P681H PLP was evaluated in the presence of P681H, WTP681H, and L452R cDNA. The same strategy was employed for selectivity testing of L452R PLP using L452R, WT-L452R, and P681H cDNA targets as well as WT PLP in the presence of WT, P681H, and L452R cDNA sequences. All the cDNA targets were assessed at the concentrations of 105 cDNA copies·μL−1.
Image processing was performed with a dataset comprised of triplicate images from the conditions studied. The processing consisted in cropping the outer 20% of the original RBG image to remove the coffee ring effect. Followed by a blue filter application, where pixels with a hue value between 85 and 140 (blue range) are removed and replaced by the mean value of the rest of the image. The blue-filtered image is then thresholded by replacing the 25% less saturated pixels with the mean value of the rest of the filtered image. Finally, the processed image is cut into 20 sub-images from which several features are extracted consequential to the implementation of formulas. The modification made to the formulas consists of interchanging the greyscale intensity with the intensity of each RGB channel 74. A total of 18 values are extracted from each sub-image corresponding to the mean color value, standard deviation, mode, skew, energy, and entropy for each of the RGB channels.
For the automation of the real human samples, a supervised machine learning algorithm was implemented to classify the images into two classes healthy and patient. An SVM with a rbf kernel was established with its hyperparameters C and gamma assessed by a Bayesian search and an overfitting absence validated via a 5-fold cross-validation. The database for this study is integrated by 33 patients and 15 healthy controls, for every sample, studies were conducted in triplicates, acquiring a total of 9 images per timepoint. The datasets are divided into 2 classes: healthy (negative) and patient (positive). Then they are divided into distinct training and testing sets. The training set consists of 2 thirds of the vectors from patients 1, 7, 9, 11, 15, 19, 21, 23, 25, 29, 31 and negatives 2, 3, 4, 5, 6, 7, 8, 9, 11, 13 and 15; the test set is integrated by the remaining vectors. The SVM produces a prediction for each vector of the test set to be either healthy or patient.
SARS-COV-2 clinical samples including 18 saliva samples and 15 nasopharyngeal swab samples were collected from adult patients with COVID-19 symptoms such as fever, fatigue, and dry cough through the University Health Network's PRESERVE-Pandemic Response Biobank (REB #20-5364). All samples tested positive for SARS-COV-2 using RT-PCR. Moreover, the viral load in the saliva samples was evaluated by qPCR (QuantStudio 12K Flex, ThermoFisher™). The samples were assessed at a Level 2+ facility situated in the Lady Davis Institute at the Jewish General Hospital, Montreal Canada.
Electrochemical measurements were performed in a conventional three-electrode cell utilizing Autolab PGSTAT204 potentiostat/galvanostat. The plasmonic platforms were used as the working electrode, while Ag/AgCl and platinum wire served as the reference and counter electrodes, respectively. The potential of cyclic voltammetry tests ranged from −1 to 1 V compared to the reference electrode with a scan rate of 50 mV·s−1. Measurement of the photoresponse was done employing the chronoamperometry technique under chopped ambient visible light (light on/off cycles: 5 s) at a bias potential of 1 V vs. Ag/AgCl in an aqueous nucleic acid amplification assay solution (10 parts LAMP master mix, 2 parts 10× primer stock, 7 parts RNAse free water, and 1 part target RNA sample).
Results are conferred as the mean value±the standard error for triplicate measurements. OriginPro (OriginLab, 2021) software package is used for statistical analysis. Limits of detection and linear ranges are calculated using linear regression methods including the line slope and the standard error of the intercept. Statistical significance is evaluated using a oneway analysis of variance (ANOVA) with post hoc Tukey's test for mean comparison. Datasets difference is considered statistically significant for p<0.05. Paired Comparison Plot (version 3.60, OriginLab) graphing application is used to generate the figures using conservative p values.
Clinical throat swab samples were collected from the West China Hospital of Sichuan University (Ethical Approval no. 2020(100)) and frozen extracted RNAs from these samples were provided by the hospital for MARVE and in-house RT-qPCR testing. RNA samples were thawed on ice, divided into 5 μL aliquots and stored at −80° C. until use. Negative throat swab specimens were acquired from healthy donors in the Deng laboratory. Influenza viruses were kindly provided by a collaborator at the Institute of Microbiology, Chinese Academy of Sciences, Beijing, China.
The QolorEX sensing chamber was designed to display plasmonic effects tuned by geometric and material parameters. To express plasmon resonance in large areas, fabless polystyrene nanoparticle self-assembly was used to construct the assemblies on biocompatible layered materials (
To investigate electron engagement with the amplification assay, the kinetics of the reaction in the presence and absence of the emancipated plasmonic electrons via electrochemistry was studied. By employing real-time chronoamperometry in the presence/absence of the target nucleic acid, it was demonstrate that the electron injection under illumination directly affects the amplification reaction, and does not involve the side reactions, like phenol red oxidation (
Next, the sensing platform was optimized to maximize the electromagnetic field enhancement and consequently surface electron emancipation. A series of platforms with different diameters of polystyrene nanoparticles (ranging from 200 to 1000 nm) were fabricated, demonstrating a repeatable array of curved surface topography with distinguished pitch (
Where Green, Red, and blue represent the intensity of each image channel, respectively.
QolorEX enables multiplex analysis of RNA/DNA targets when coupled with RT-LAMP assay to establish the diagnostics capability of QolorEX. The assay incorporates a thermal lysis step to release nucleic acids (95° C., 3 minutes), followed by incubation at 65° C. for the isothermal amplification reaction. As discussed in the previous section, the plasmonically boosted amplification accelerates the H+ release, leading to quantitative colorimetric signal change in the presence of phenol red. The diagnostic capability of QolorEX LAMP was demonstrated using a paradigm of viral respiratory infections such as Influenza and SARS-COV-2.
First, the sensitivity of the assay with SARS-COV-2 viral particles (wild-type strain) was establish in a series of dilutions (0.01 to 90 viral particles/μL) within the physiological range. Sequential images of the color-sensing chamber were acquired over 60 min to create a mosaic color matrix (
QolorEX LAMP for SARS-COV-2 generated a differentiable colorimetric signal for all tested concentrations of SARS-COV-2 samples. The signal has a positive correlation with the concentration of the heat-inactivated viral particles within the physiological relevant range of 0.01 to 90 viral particles/μL in both buffer and human saliva media
The serial dilutions of SARS-Cov-2 samples (heat-inactivated and RNA) were tested with a qPCR assay
To establish the selectivity of the QolorEX LAMP for SARS-COV-2, the positive signals of different SARS-COV-2 variants (Wild type, Omicron B.1.1.529, Delta B.1.617.2, Alpha B.1.1.7, Eta B.1.525, and Gamma P.1) were compared in buffer and human saliva to the signals of other viral infections which has the potential to interfere with SARS-COV-2 detection either for similar molecular composition or for similar patient symptoms. Thus, Influenza A H1N1 virus, human coronavirus 229E (HCoV-229E), and Middle East respiratory syndrome coronavirus (MERSCOV) are tested along with negative control (no RNA). The colorimetric signal for SARS-COV-2
RNA (wild type and variants) demonstrated a higher value than all the other tested samples in both buffer and human saliva media (
To further investigate the universal application of QolorEX, a highly selective RT-LAMP primer set was incorporated for sensitive detection of Influenza A H1N1 RNA (QolorEX LAMP for H1N1) through targeting the Segment 4 Hemagglutinin gene (
In addition to viral respiratory infections, the QolorEXLAMP was employed for DNA profiling of antimicrobial resistant E. coli and MRSA, as a proof of concept for DNA amplification detection. The QolorEXLAMP for E. coli detected color changes for DNA concentrations ranging from 0.2 to 70 ng/L (
The QolorEXLAMP for MRSA targeted the mecA gene (
The versatility of the QolorEX platform can be extended to variants and subtypes when coupled with a highly sensitive rolling circle amplification assay (RCA). The RCA assay utilizes a highly efficient HiFi Taq DNA Ligase, which shows surged discrimination between mismatched and accurate base pairs at either 3′ or 5′-side of ligation. It was demonstrated that by utilizing three probes, the QolorEX system can differentiate wild-type (WT) SARS-COV-2 from Delta (B.1.617.2) and Omicron (B.1.1.529) variants (
The RCA-based raw colorimetric readout for different concentrations of SARS-COV-2 cDNA shows a gradual color change during the 0 minutes to 60 minutes time lapse (
A PLP targeting the L452R mutation (
Omicron SARS-COV-2 variant cDNA was selectively detected using an RCA assay (QolorEX RCAOmicron assay) using a PLP targeting the P681H mutation (
Real human samples are highly complex, so the evaluation of the QolorEX device with patient samples can determine its efficacy when challenged with untreated human biofluids. Due to the availability of testing samples in the past two years, QolorEX was validated with COVID-19 patient samples. With a two-step operation, the patients simply introduce their saliva into the QolorEX microfluidic cartridge and place the cartridge inside the imaging box. Inside the box, the sample will be processed and imaged in an automated fashion, and diagnostic results will be displayed on a cell phone or a computer. A set of raw human samples obtained from 33 patients clinically diagnosed with the wild-type SARS-COV-2 strain were analysed (
As demonstrated herein, the QolorEX system combines rapidity, clinically relevant sensitivity, specificity, and versatility for detection of a variety of pathogens (viruses and bacteria) in a multiplex automated fashion at the point of care on par with current FDA-approved molecular diagnostic tests. The first key feature of the QolorEX is the plasmonic color-sensing substrate confined in microfluidics that allows for ultra-rapid colorimetric signal transduction (10 min) by accelerating the amplification rate in RT-LAMP and RCA, and allows for ultrasensitive and specific detection of a few copies of genomic RNA/DNA in one-step amplification. The plasmonic color-sensing substrate oscillates upon illumination with white light, leading to the injection of electrons into the media that is consumed by the amplification assay, accelerating the reaction rate and enabling a 75% decrease (three-fold) in detection time compared to the original RT-LAMP assay. Unlike fluorescent read-out, which is commonly used in reading the amplification signals, QolorEX is a label-free colorimetric approach with a quantitatively linear scaling response as a function of the pathogen load that only requires tuning amplification reagents. By testing different respiratory viruses (SARS-COV-2, Influenza A H1N1) and antibiotic-resistant bacteria (E coli, MRSA), the versatility of the QolorEX was demonstrated for the specific detection of different pathogens (viruses and bacteria) as well as the differentiation between different target variants and subvariants (e.g. WT, Delta B.1.617.2, and Omicron B.1.1.529 for Covid-19) through detection of single nucleotide polymorphism. In the case of viral respiratory infections, the ability of QolorEX to differentiate between variants and subvariants would assist with the management of the virus spread and facilitate decision-making and planning globally. QolorEX is thus a versatile tool deployable for the detection of new pathogens and emerging viral variants as it only requires adjustment in amplification primers and reagents. The second feature of the QolorEX is the integrated microfluidic cartridge and the imaging box that allows for automation of sample collection, RNA/DNA extraction, amplification, and multiplex detection in a two-steps action by untrained users in 13 minutes (3 min lysis and 10 min detection). The automated fluid actuation system in the microfluidic cartridge incorporates the use of mechanically driven sub-millimeter-sized thin suction cups for generating negative pressure to drive and mix the sample and assay reagents in parallel fluidic compartments. This eliminates the need for auxiliary components such as pumps or complex chip design as required by capillary-driven systems. In addition, the imaging box custom-designed x-y translation stage with micron resolution holds the microfluidic cartridge and operates in tandem with controllers and linear actuators to achieve full process automation. Moreover, the imaging box reflected light microscopy module with controlled illumination, gives QolorEX similar performance as commercial microscopes using Koehler illumination in a much smaller dimension and a fraction of the cost. The final colorimetric event is captured by imaging the detection chamber using a CMOS sensor housed in the portable setup. Using a simple supervised machine learning data interpretation (SVM) connected to the signal transmission system of QolorEX, the results are displayed on a cellphone or a computer. All the key features of QolorEX allow for an automated user-friendly operation for lay users and testing patients simultaneously for multiple targets at the point of care.
The innovative design aspects of the QolorEX address limitations in the current point-of-care diagnostic tests in the areas of assay rapidity, versatility, multiplex detection of pathogenic strains and mutations, process automation, and ease of use for lay users. The QolorEX testing device was established in a clinical setting by successful validation with untreated samples (saliva) obtained from 33 patients clinically diagnosed with the wild-type SARS-COV-2 strain against 15 SARSCOV-2 negative samples. Accordingly, QolorEX achieved clinically relevant sensitivity and specificity with 95% accuracy on par with qPCR as required by WHO for point-of-care tests. The multiplexity aspect was tested by running QolorEX for every sample against Influenza A H1N1 and all known SARS-COV-2 variants to date (WT, Alpha B.1.1.7, Delta B.1.617.2, Gama P.1, Eta B.1.525, and Omicron B.1.1.529. With a two-step operation, the patients simply introduce their saliva into the microfluidic cartridge and place the cartridge inside the imaging box for further sample processing and imaging in a fully automated fashion. The cell phone interface has the potential for simple interpretation and communication of the test results for harmonized data collection and rapid evaluation in national and international large-scale trials. These types of systems are paramount for the spread management of respiratory infections.
Antimicrobial resistance (AMR) poses a grave threat to global health and remains a research priority. Overuse of antibiotics significantly contributes to the rise of AMR in human infections. Addressing this critical issue, the present example provides QolorAST, an ultrasensitive and fully automated platform designed for rapid bacterial identification and phenotypic antibiotic susceptibility testing.
QolorAST integrates a high sensitivity colorimetric detection based on plasmonic color-printing, microfluidic design, automated microfluidic device for confinement and reagent handling, custom illumination and image acquisition, as well as a supervised machine learning, support vector machine (SVM) method for rapid and accurate bacterial identification and phenotypic AST readouts. Central to its functionality is a plasmonic nanostructured material enabling ultra-sensitive optical readouts during the conversion of resazurin to resorufin-a pH-sensitive dye serving as a vital indicator of bacterial viability and growth during antibiotic exposure.
This versatile system adeptly determines minimum antibiotic inhibitory concentrations for 12 bacterial species against 7 antibiotics, showcasing its adaptability. By harnessing plasmonic color platforms, QolorAST achieves early detection, reducing genotypic identification and phenotypic profiling times to just 15 and 30 minutes, respectively. This is a monumental advancement, compared to the conventional 3-4 days with existing methods, that is achieved in a portable, automated, and multiplex fashion.
In rigorous validation through a double-blinded clinical study involving 47 clinical specimens from patients suspected of urinary tract infections, QolorAST demonstrated a 90% essential agreement and a 96% category agreement. QolorAST is therefore an improvement over the use of 96 well plate assays. By replacing such assays, QolorAST significantly slashes the time-to-result for antibiotic susceptibility testing. Its cost-effectiveness and efficiency mark a paradigm shift in global healthcare and controlling the pathogenic bacterial drug efficacy.
Conventional medical systems used in health care facilities are important keys to fight against pathogenic infections. Yet resource-limited settings, pose unique challenges due to interruption of electrical power, shortage of skilled professionals, and the lack of reliable automatic data processing systems. The conventional routine for infectious disease treatments involves empirical antibiotic prescription or long (2-3 days) man power intensive antibiotic susceptibility tests (AST) based on phenotypic bacterial growth. Therefore, a low-cost, rapid, automated, and easy-to-use AST operation can offer a user-friendly point-of-need-testing setting while remarkably reducing the miss prescription of antibiotics.
Colorimetric based assays are well established in clinical settings. They provide ease of operation, ability to be implemented in low resource settings, and no need for expensive reagents. Resazurin, is a pH sensitive and weakly fluorescent dye characterized by its dark blue color in its oxidative state while rendering fluorescent pink color upon reduced to resorufin in the presence of intercellular NADH. Traditionally, resazurin was used for testing cell viability indicators but due to its pH sensitive properties it can be utilized in genotypic detection based on pH sensing. Yet, traditional colorimetric assays lack analytical sensitivity and rapidity to output a quantifiable result and therefore suit unchaperoned point of need applications. Sensitivity in colorimetric sensing largely relies on the resolution of pigmented color that is captured, while the rapidity of the sensor depends on how fast the color change occurs according to the chemical reactions.
Non-pigment color systems, operating based on various physical resonances such as thin-film interference, Mei resonances, and surface/gap plasmons attracted huge interest in different applications including color displays, optical anti-counterfeiting, and sensing. According to Abbe's classical diffraction limit theory, an optical microscope ideally can resolve juxtaposed color elements down to a pitch of λ/2NA.
To explore these possibilities, QolorAST was developed, an automated point-of-need system coupled with a genotypic loop mediated isothermal assay (LAMP) for bacterial identification (ID) and a phenotypic resazurin reduction assay for antibiotic susceptibility profiling (AST). QolorAST uniquely utilizes nanoplasmonic colorimetric structures to enable rapid detection of the onset of color change and identify AST in less than 30 mins. The core principle of QolorAST signal generation is based on the reduction of the dark blue resazurin to light pink resorufin on top a plasmonic color printing platform where the plasmonic color is initially inhibited by the dark blue resazurin but is fully detectable at the early stages' reduction process. The resazurin reduction can be initiated through pH change (genotypic QolorAST assay) or due to the metabolic activity of viable bacterial cells (QolorAST phenotypic AST assay). It is demonstrated in the present example that the rapid plasmonic colorimetric detection leads to rapid identification (<15 mins) and phenotypic AST profiling (<30 mins). To fully automate the process, an autonomous imaging box was developed to allow integrated incubation, fluid actuation and imaging. The imaging box utilizes a unique cyclic filtration/actuation process to allow repeated and precise control over the temporal sample release. Finally, a supervised machine learning code was utilized for autonomous data analysis. Overall, QolorAST provides a complete solution towards bacterial identification (ID) and phenotypic AST profiling in less than 30 mins in comparison with 48 hours for the current conventional methods.
QolorAST combines unique technological innovations of high sensitivity colorimetric detector and econonomic 3D-printed microfluidic design, with well-validated genotypic ID and phenotypic AST readouts (LAMP and bacterial viability-based assays) and custom image acquisition and machine learning (support vector machine). The QolorAST device features two components that together present an automated, integrated and cost-effective approach to sample preparation, detection, and drug susceptibility test at the point-of-care or point-of-need. The microfluidic cartridge is a user-friendly platform with attached accessories for urine reservoir, preloaded reagent storage and flow actuation that automates the assay process when placed inside the colour reading module. The microfluidic cartridge includes two chambers; (i) a “mixing chamber” including urine reservoir, and preloaded reagent storage which is used to enrich bacteria, mixing, and the infusion of antibiotics; (ii) a “detection chamber” that is equipped with a plasmonic colour printed microchip for sensing. These two chambers are separated by paper filters. The QolorAST colour reading module that is a portable battery-operated platform that consists of linear actuators and a motorized stage for automated sample processing, and an illumination-coupled imaging system that operates with a Bluetooth-interfaced smartphone application. The antimicrobial resistance test involves four key stages using a microfluidic device which is discussed in greater detail below. Briefly, the first step, urine sample loading involves introducing the sample and ensuring even distribution within the inlet by covering the orifice, agitating, and using the inlet filling button, moving the sample towards the mixing junction with a suction membrane. Second, the imaging system adjusts to 37 degrees, autonomously mixing the sample with growth media and advancing it towards filters by raising the suction membrane to a first level (Level 1 (L1)). Third, a filtration process eliminates bacteria from the mixed solution, raising the suction membrane to a higher Level (L2) to bring the solution to the colorimetric window. In the fourth stage, imaging is performed for all 24 chambers by moving the stage beneath the objective, and the suction membrane is returned to L1 to reintroduce the filtered solution and combine it with bacteria. Steps 2 to 4 can be repeated sequentially as needed thanks to the bidirectional flow rendered possible by the suction membrane. Accordingly, it is possible to determine the fluorescence at different timepoints during the same assay by flowing the sample into the detection chamber and then back to the mixing chamber to continue reacting with the colour or other detection agents.
In this example, resazurin was used as a well-established bacterial viability-based assay. Resazurin is a well-known colorimetric assay that is widely used in clinical colorimetric read-out systems to identify the viability of the cells. Resazurin assay functions in two steps, first, the assay is absorbed by the membrane of the cell, and second, over the metabolic reactions in the live cells the Resazurin molecule is reduced to Resorufin. Consequently, the color of the cell medium will change from navy blue to light pink after a few hours evidencing a YES/NO metabolic sensor. This phenomenon is detected for more than four hours using the conventional method. However, the partially reduced resazurin can be sensitively detected via the QolorAST device integrated with plasmonic platform. This is thanks to the plasmonic features embedded in the detection chamber, allowing for rapid and sensitive colorimetric response based on the refractive index changes in the media. Unlike the conventional Resazurin assay paragons, QolorAST detection chamber executes a quantifiable color gamut providing information concerning the pathogen concentration, and antibiotic efficacy dose. The plasmonic parameters of the QolorAST were investigated via different physical characterization methods. The light interaction with the plasmonic platform is the key parameter in determining the superlative color gamut that in turn enables higher resolution in colorimetric AST. Therefore, a comprehensive theoretical study was performed via finite-difference time-domain (FDTD) to optimize the geometrical parameters with a deeper understanding. Unlike the conventional use of resazurin assay which indicates the color transition from blue to pale pink upon absorption within the cell membrane and gets reduced by the metabolic reactions within the cells, the detection chamber ratifies a wide gamut of high-resolution colors, from dark blue to green, which generates a quantifiable and sensitive measure of the live cells. Sensitivity in colorimetric sensing largely relies on the resolution of pigmented color that is captured, while the rapidity of the sensor depends on how fast the color change occurs according to the chemical reactions. The light interacting with resonances associated with the discrete harmonic energy states are structurally engineered, and could offer new opportunities to enhance the colorimetric sensing. The plasmonic enhanced color-generation strategy involves the patterning of various geometrical metallic nanostructures and investigates the hue and gamut of colors experimentally. The optimization of the plasmonic platform included the selection of material, nanoparticle diameter, and thickness of the metallic layer. Moreover, the morphology was characterized via SEM and AFM microscopy. The reflectance spectra of plasmonic platforms with different nanoparticle diameter was experimentally investigated. The 400 nm diameters particle platform showed the widest gamut change during the reduction of resazurin to resorufin. This was further confirmed through a theoretical FDTD study to visualize the local EM-field enhancement and identify the modes of excitation for platforms with different nanoparticle diameters. Finally, the microfluidic device parameters were optimized to enable endurance of the plasmonic detection chamber.
For QolorAST, a microfluidic device was developed with three primary objectives: First, to facilitate the mixing of urine samples with antibiotics and detection reagents. Second, to filter the sample, producing a bacteria-free solution suitable for imaging. Finally, to position the bacteria-free solution over the colorimetric detection chamber. The design of this microfluidic device has the capacity to conduct twenty-four independent tests concurrently within separate channels. The microfluidic device 100 as shown in
The first 3D-printed layer 243 serves as the reservoir for the urine sample and detection reagent, as well as the location for the filters and colorimetric platforms (
The 3D-printed components are designed using CAD software (SolidWorks 2022 SP5) and fabricated using a Stereolithography-SLA 3D printer (FormLabs-FormLab). The PSA patterns are designed using CAD software (Graphtec Studio) and transferred onto biocompatible PSA material (Adhesive Research-ARcare 90445Q) using a cutting plotter (Graphtec-CE7000-40). The filters 104 with a 0.45-micron pore size hydrophilic polyethersulfone membrane (Pall) are punched and securely bonded to the microfluidic device using the plotted PSA. The paramagnetic layer 249 is made from tight tolerance air hardening A2 tool steel bar (McMaster-Carr) with 1.6 mm thickness.
The antimicrobial resistance test was conducted using the microfluidic device of
The automated QolorAST setup is depicted in
A custom PCB was designed to minimize the electronics' size footprint and improve the system's rigidity. An Arduino Uno R3 was used to run most of the components operating the microscopy platform for the imaging apparatus.
Additionally, a smaller incubator set-up was connected separately which ran the same Yosoo heater and fan combined device but was connected to the Inkbird ITC-100 PID module for discrete temperature control. Both this incubator and the microscopy platform PCB were connected to a two-plug power bar with a built-in switch and fuse protection. Therefore, the entire system is turned on and off via this switch which allows for it to remain plugged into a wall, but it does not draw power until the user decides to do so.
The outer enclosure of the setup was built using 0.25 inch thick Clear Scratch and UV-Resistant Cast Acrylic Sheets (#8505K755, Mcmaster carr). The epi-illumination imaging set-up was employed for imaging the media over the nanostructured platform. The setup features an illumination column with a LED placed at the focal point of a diffuser convex lens. The collimated beam then reaches a beam splitter that projects the light onto the back aperture of the objective (TU Plan Fluor EPI ×20, Nikon). The captured image is focused using a tube lens (f=200 mm) onto a CMOS sensor (Sony IMX477R, 12.3 MP, Raspberry Pi.) and processed further by Raspberry Pi 4 (Raspberry Pi).
The electronic system of the apparatus consists of three main blocks: the incubation block, which includes a small space constant temperature fan heater (PTC heater, Yosoo), a solid state relay (SSR-25 DA, Jekewin), an AC/DC power supply (S-60 W-12, YXDY), and a PID temperature controller (ITC-100VH, Inkbird); the imaging block, which includes a single-board computer (Raspberry Pi 4 Model B, Raspberry Pi), and a 12.3 MP high quality camera (697-1, Raspberry Pi); and the axis movement block, which includes a microcontroller board (UNO R3, Arduino), two stepper motor drivers (A4988, Pololu), two stepper motors (11HS12-0674S, NEMA), a servo motor (DS3225, DSSERVO), a HC-05 Bluetooth module (1258, Canada Robotix), and a 12/5V power supply (TOL-15664, SparkFun). The system's workflow involves the Bluetooth module receiving commands to move the sample via the stepper motors and servo to pre-set positions while the camera module displays the sample through the microscope to a monitor. Independently, the incubator block maintains a temperature of 37° C. which can be adjusted via the PID temperature controller.
The fabrication of the QolorAST microfluidic was done as follows. The detection chamber (3 mm×3 mm×1 mm), microfluidic channels and inlet/outlet ports (01.5 mm) were designed using CAD software (SolidWorks 2022 SP5, licensed by CMC) and fabricated using a Stereolithography-SLA 3D printer (FormLabs-FormLab3). The PSA patterns are designed using CAD software (Graphtec Studio 2) and transferred onto biocompatible PSA material (Adhesive Research-ARcare 90445Q) using a cutting plotter (Graphtec-CE7000-40). The filters with a 0.45-micron pore size hydrophilic polyethersulfone membrane (Pall) are punched and securely bonded to the microfluidic device using the plotted PSA.
Plasmonic platforms were fabricated using a fabless nano-patterning technique. A generic approach is used to develop a colloidal self-assembly monolayer (SAM) of nanoparticles at a water/air interface 48, followed by transferring the resulted honey-comb patterns to the silicon (Si) substrate. Next, a ZnO thin film (120 nm) is deposited, followed by a thin aluminum layer (10 nm) to provide a tunable localized surface plasmon resonance with a white background. Last, the plasmonic platform is integrated with the microfluidic cartridge.
The fast colorimetry strategy based on the plasmonic substrate involves patterning of metallic nanostructures with sizes smaller than of diffraction limit of light to support plasmonic oscillations as described in greater details above. Unlike the organic-dye color filters, the plasmonic color filters offer advantages such as high color tunability, sensitive color changing based on medium permittivity, and low color degradation rate. Amongst the plasmonic nanostructures used for plasmonic color production are nanodisks, ellipses, nanocubes, and multimers made of plasmon-supported materials such as using gold, silver, and aluminum. The geometrical nanostructures and materials can be designed to resonate at a specific optical frequency leading to the production of different colors across the visible spectrum. Direct-write technologies such as e-beam lithography and focused ion-beam lithography are the most commonly used patterning techniques allowing for high resolution and placement accuracy, however, they are time-consuming, expensive, and do not allow large scale production. Fabless approaches such as sacrificial polymeric nanoparticle templates, controlled phase separation, mesoporous particles assembly, and colloidal particles self-assembly have been found promising alternatives for large-area patterning.
The imaging studies of the present Example were all performed by collecting images in triplicates for each timepoint (from 0-60 min) and condition combination, integrating a database for each bacteria studied. Each image in the datasets is cropped to 80% of its original dimensions to avoid the coffee-ring effect, followed by a brightness normalization. Additionally, defects on the images are removed by implementing a threshold that removes 25% lower saturated pixels and replaces them with the mean value of the rest of the image, having as an output standardized images.
In the manual color signal evaluation, for each image, the code selected 30 areas at random, read, and averaged their RGB values. The G/B color feature was calculated by dividing the Green channel value by the blue channel value (Eq. 1) and depending on a pre-established threshold the Green channel was modified by doubling or dividing by half its value, and so the color feature was recalculated as Gmod/B (Eq. 2) prior to statistical analysis. For the CIE plots, the RGB values were converted into XYZ, followed by the x and y values calculation which were scattered on the CIE1931 color space to analyze the color change through time. All images were processed and analyzed through python and MATLAB™ scripts.
One of the characteristic parts of QolorAST is the detection chamber embedded with plasmonic nanostructures to determine the viability of bacteria infused with antibiotics in the presence of Resazurin assay. It was herein methodically investigated the optimized characteristics of the PC platform experimentally and theoretically. It is essential to remember that the color of the plasmonic platform other than the refractive index of the media depends on the choice of the metallic layer and the geometrical features of the plasmonic nanostructures i.e. size and pitch of nanoparticles. Silver, gold, and aluminum are the most commonly used. First, the choice of the metallic layer was studied. Silver was excluded due to its antibacterial properties which would interfere with the AST resulting in false negatives. Next, the color change from resazurin to resorufin was studied using glass platform, 750 nm Au platform, and 750 nm Al platform. To investigate the wideness of color gamut between the resazurin and resorufin, RGB values were extracted from the microscopy images and converted to x-y coordinates on standard CIE 1931 chromaticity diagram using standard conversion models. The Al platform showed a wider color gamut compared to Au and glass platforms demonstrated by the greater change in the y-value of 0.17 for the Al platform compared to 0.075 and 0.01 for the Au and glass platforms respectively (
For characterizing the surface morphology of the nanoparticle monolayers tapping mode atomic force microscopy (AFM) and scanning electron microscopy (SEM) techniques were performed. Optical characterization was carried out with a Lambda 750 UV/Vis/NIR Spectrophotometer (PerkinElmer). The incident and collected light beams had normal incidence to the platform. Also, we performed optical imaging under a highly controlled environment using Nikon Eclipse LV150 (Nikon) with a ×100, 0.9 NA air objective and Nikon digital sight ds-fi1 CCD camera. The fundamental fluid flow characteristics of the microfluidic cartridge were studied using COMSOL Multiphysics (V5.6).
AFM micrographs for 200 nm, 400 nm, 750 nm, and 1000 nm platforms are shown in
The reflectance spectra was investigated for the 200 nm, 600 nm, 750 nm, and 1000 nm plasmonic platforms performed in water-based media (
A FDTD theoretical simulation was also performed. A strong EM field generated in plasmonic nanopatterns is of importance to control the hue and saturation of the generated color. These effects in the device were investigated via a simulation study based on the finite-difference time-domain (FDTD) module of Lumerical Solutions. The visualization of the local EM-field enhancement is of interest to identify the modes of excitation and therefore help with the design of the optimized colorimetric substrates. The EM-field contour plots were simulated using a plane wave with wavelength in the visible range (400-700 nm) excitation. The simulation resulted top-view and side contour plots of a self-assembly nanoparticle pattern illuminated with a Gaussian beam centered at 532 nm wavelength in transverse electric (TE) and transverse magnetic I modes, revealing the total EM-field enhancement, (|E|/|E0|).
The conventional metric to study the superiority of the structure is the absorption efficiency which is the ratio of the absorption cross-section to the geometric cross-section. The absorption efficiency for a cylindrical particle illuminated along its axis is given by
The experimental and simulated broad-band reflectance spectra (
With noticeable hotspot along the nanocavities formed by the pitch of the SAM, which directly correlates to the diameter of the nanoparticle, the EM-field distribution demonstrates over enhancement factor at the nanocavities of the SAM in TE mode.
A numerical simulation study of the fluid flow inside the microfluidic chip was conducted. The microfluidic system provides the ability to filter the bacteria sample as well as providing accurate control of fluidic flow rate and pressure over the self-assembly monolayer (SAM). Bacteria is captured using the a paper filter; therefore, a pure liquid is passing over SAM which reduces a potential error coming from exposing bacteria to the platform. Additionally, the accurate control over fluid flow helps to keep the inlet pressure in a safe range. High shear stress results in detachment of SAM and subsequently failure of detection.
COMSOL Multiphysics 5.5 was used to simulate the fluid flow inside the microfluidic cartridge. Constant fluid flow is considered as an inlet boundary condition and zero relative pressure (ambient pressure) is set as an outlet boundary condition, respectively. Water was used as the simulation media.
The simulation results show an overall pressure of 8 Pa. The shear stress was examined for different input flow rates on top the SAM layer in the detection chamber. The detachment of SAM starts at 1 Pa. The determination of a safe flow rate was thoroughly studied. The contour of the flow rate was observed for four different input flow rates: (i) 1 μL min−1, (ii) 10 μL min−1, (iii) 17.5 μL min−1, and (iv) 25 μL min−1. The maximum flow velocity inside the PC chamber increases with the increase of the input flow rate (
The QolorAST system utilizes a harmonized automated pipeline with the purpose of overcoming inconsistencies in the images derived from sources of variations such as the quality of the platform and microfluidic fabrication. The workflow starts with the image collection in triplicates for each bacteria at specified conditions. Each image is treated to correct for the coffee-ring effect, brightness, and defects; producing standardized images ready for further processing through our manual QolorAST evaluation or through machine learning. The manual QolorAST signal evaluation was calculated from randomized areas of each image. The color features were then used for the statistical analysis which was as follows. Results were conferred as the mean value±the standard error as mentioned in the image processing section. OriginPro (OriginLab, 2021) software package was used for statistical analysis. Limits of detection and linear ranges were calculated using linear regression methods, including the line slope and the standard error of the intercept. Statistical significance was evaluated using a one-way analysis of variance (ANOVA) with post hoc Tukey's test for mean comparison. Datasets difference was considered statistically significant for p<001. Paired Comparison Plot (version 3.60, OriginLab) graphing application is used to generate the figures using conservative p values.
The incorporation of machine learning in healthcare can be beneficial for diagnostic and surveillance tasks, especially in low-resource settings, where there is no access to high-end medical and high throughput, automation and diagnosis accuracy are essential. The standardized images are each subdivided into 20 mini-images. The modified Sergyan formulas were used to extract a total of 18 color features, corresponding to the mean color value, standard deviation, mode, skew, energy, and entropy for each of the RGB channels of each mini-image21. The 18 values of the 20 mini images integrate the dataset of a sample, the library that contains them all is then divided into 70% for training the machine learning model and 30% for testing it. A support vector machine (SVM) with radial basis function (rbf) kernel was optimized, trained, and validated to classify the images into a positive (color changed) and a negative class (no color change).
The images from the harmonized database are subdivided into 20 mini-images to be analyzed per condition. By implementing the modified Sergyan formulas, where the greyscale intensity was replaced with the color channel intensity, 18 color features were extracted. The 18 color features correspond to the mean color value, standard deviation, mode, skew, energy, and entropy for each of the RGB channels of the mini-image. The collection of values integrates the dataset of a sample, the library containing all datasets is then divided 70% for training the machine learning algorithm and 30% for testing it.
A supervised vector machine (SVM) algorithm was implemented to determine the probability of the image belonging to a color-changed (positive) or non-color-changed (negative) class. The optimization of the C and gamma hyperparameters of the model was determined via Bayesian search, and the kernel radial basis function (rbf) was selected. A 5-fold cross validation was implemented to validate that the algorithm did not overfit. The training set was appropriately labeled to a color-changed (positive) or non-color-changed (negative) and separated into two classes according to the ground truth and clinical annotations.
For the preclinical experiments, the training of the SVM included images taken from various bacteria (E. coli, Enterococcus, MRSA, and PA) for the positive class included the bacteria images were taken from the resistant condition (color change condition) and the negative class included bacteria with antibiotics in a susceptible condition or at 0 min when no color change is present. For the clinical samples phenotypic study, 70% of the urine samples in “No-antibiotic” condition, specifically U05, U10, U14, U28, U32, U35, U37, U49, U53, and U56 integrated the positive class and U01, U04, U16, U20, U25, U26, U29, U30, U38, and U39 integrated the negative samples. The test set was integrated for the remaining vectors. The antibiotic MICs for the clinical samples (Cipro and Nitro) were analyzed with the model trained in the “No-antibiotic” condition, as this model is optimized to predict the probability of color change of the image.
To evaluate the antibiotic MIC in the preclinical or clinical sample studies, the SVM tests each antibiotic concentration for each sample and produces a prediction which is organized in a table, where the MIC can be observed as well.
The SVM outputs a prediction on each of the vectors of the data set belonging to the positive class. The predictions are plotted as the probability of belonging to the positive class or the negative class.
Ampicillin-resistant E. coli strain mm294 was cultured overnight at 37° C. in Luria broth (LB) media supplemented with 100 μg/mL Ampicillin. Methillin-resistant S. aureus (USA 300 SCCmec Type IV), and Ciprofloxacin-resistant P. aeruginosa were streaked on LB miller agar overnight at 37° C. and resuspended in water aliquot. Next, the bacterial concentration was determined by optical density technique using a Spectronic 21D spectrophotometer. Subsequently, aliquots of different concentrations were prepared for the antibiotic susceptibility testing experiments in LB media. Biotyper spectra (MALDI Bruker).
Acinetobacter baumannii, Enterococcus faecalis, Enterococcus faecium, Streptococcus agalactiae, Enterobacter cloacae, Klebsiella pneumoniae, Streptococcus pneumoniae, E coli MG1655, and E coli Trimethoprim resistant were streaked on LB miller agar overnight at 37° C. and resuspended in water aliquot. Next, the bacterial concentration was determined using optical density technique. Subsequently, aliquots of different concentrations were prepared for the antibiotic susceptibility testing experiments in CamHB growth media.
For the LAMP assay, a standard reaction volume of 25 μL was used. It consisted of 2.5 μl 10× primer mix, 12.5 μl 2× master mix, 9 μl Rnase-free water supplemented with 0.25 mg/ml resazurin, and 1 μL RNA sample. The primers are presented in the table 9 below. Bacterial samples were (E. coli spiked solutions of 7.2×106 gDNA copies·ml−1 to 7.2 gDNA copies·ml−1 equivalents to 107 cfu·ml−1-10 cfu·ml−1) first thermally lysed at 60° C. for 5 minutes and then mixed with the assay. This was followed by incubation at 65° C. for different periods to visualize color change versus time.
E. coli LAMP primers sequences
Aliquots of different concentrations were prepared (5×105 CFU ml−1, 105 CFU ml−1, 104 CFU ml−1, 103 CFU ml−1, 102 CFU ml−1, and 50 CFU ml−1) of E coli mm294, MRSA, PA, E coli MG1655, E coli Trimethoprim resistant, A. baumannii, E. faecalis, E. faecium, S. agalactiae, E. cloacae, K. pneumoniae, and S. pneumoniae. Growth media supplemented with 0.02% resazurin were used and aliquots were incubated at 37° C. for different periods starting from 0 minutes incubation till 60 minutes with a time step of 5 minutes. for the standard control study bacteria were inoculated at different concentrations (5×105 CFU ml−1, 105 CFU ml−1, 104 CFU ml−1, 103 CFU ml−1, 102 CFU ml−1, and 50 CFU ml−1) in 96 well plates and incubated 37° C. overnight. Next day resazurin solution was added to the plates, incubated in a shaker incubator for 30 mins, and subsequently the plates were evaluated using a plate reader (TECAN M100).
For resistant bacterial samples, aliquots of 5×105 CFU ml−1 of Amp resistant E. coli mm294, MRSA, and Ciprofloxacin Resistant P. aeruginosa were prepared. A resazurin solution of 0.02% resazurin (R7017, Millipore Sigma, Ontario, Canada) supplemented with 100 μg ml−1 Ampicillin, 4 μg ml−1 Oxacillin, and 2 μg ml−1 Ciprofloxacin for Amp. Resistant E. coli, Oxacillin-resistant S. aureus, and Cipro. Resistant PA respectively was used. The aliquots were incubated at 37° C. different periods starting from 0 minutes incubation till 60 minutes with a time step of 5 minutes.
For susceptible bacterial samples, aliquots were prepared of 5×105 CFU ml−1 of Amp resistant E. coli mm294, MRSA, and Ciprofloxacin Resistant P. aeruginosa. A resazurin solution of 0.02% resazurin (R7017, Millipore Sigma, Ontario, Canada) supplemented with 50 μg/mL Kanamycin, 1 μg/mL Ciprofloxacin, and 4 μg/mL Gentamicin for E. coli mm294, MRSA, and Cipro. Resistant PA respectively was used. The aliquots were incubated at 37° C. different periods starting from 0 minutes incubation until 60 minutes with a time step of 5 minutes.
To evaluate the QolorAST minimum inhibitory concentration (MIC) aliquots of 5×105 CFU ml−1 Amp resistant E. coli with resazurin/kanamycin solution were used. Solutions with different kanamycin concentrations of 1 μg/mL, 2 μg/mL, 4 μg/mL, 8 μg/mL, 16 μg/mL, 32 μg/mL, and 50 μg/mL were used to determine the MIC dose. 5×105 CFU ml−1 Oxacillin-resistant S. aureus aliquots were prepared using resazurin/oxacillin solution using different concentrations of oxacillin antibiotic (16 μg/mL, 32 μg/mL, and 64 μg/mL). 5×105 CFU ml−1 Cipro. Resistant PA aliquots were prepared using resazurin/ciprofloxacin solution using different concentrations of ciprofloxacin antibiotic (8 μg/mL, 16 μg/mL, and 32 μg/mL). In addition, the MIC test was performed for aliquots of 5×105 CFU ml−1 Amp WT E. coli using the same conditions at different concentrations of Ciprofloxacin (0.015625 μg/mL, 0.03125 μg/mL, and 0.0625 μg/mL), Nitrofurantoin (4 μg/mL, 8 μg/mL, and 16 μg/mL), Trimethoprim (2 μg/mL, 4 μg/mL, and 8 μg/mL) and Trimethoprim/Sulfamethoxazole (0.125/2.375 μg/mL, 0.25/4.75 μg/mL, and 0.5/9.5 μg/mL). The same experiment was done for Trimethoprim-resistant E. coli MIC determination at different concentrations of Trimethoprim (2 μg/mL, 4 μg/mL, and 8 μg/mL) as well as for Enterococcus faecalis and faecium using different concentration of Vancomycin (16 μg/mL, 32 μg/mL, and 64 μg/mL) and (32 μg/mL, 64 μg/mL, and 128 μg/mL), respectively.
Through the established analysis pipline QolorAST diagnostic capabilities tested using a colorimetric LAMP assay for genotypic bacterial identification (ID), and a resazurin reduction assay for phenotypic antibiotic susceptibility testing (AST). The QolorAST LAMP assay evaluated gDNA extracted from different physiologically relevant concentrations (10-107 CFU/mL) of E coli as a model organism due to it prevalence in urinary tracts infections. The detection chamber colorimetric readout creates a color matrix that shows a color change from blue to green over a 60 min incubation. The QolorAST signal shown a positive linear correlation with E coli gDNA in the range of 7.2 to 7.2×106 gDNA/ml which covers the physiological range for UTIs (
Phenotypic QolorAST rapid AST performance was established through 105 cfu/mL of E coli, MRSA, and Pseudomonas aeruginosa (bacterial strains confirmed using MALDI/TOF and PCR) with two sets of antibiotics, one set where the bacteria is resistant to the antibiotic and another set where the bacteria is susceptible. The resistant set included ampicillin, oxacillin, and ciprofloxacin while the susceptible set included Kanamycin, ciprofloxacin, and gentamicin for challenging E coli. MRSA and PA respectively. The resistant sets showed a consistent color change from blue to green with wider gamut shown in the CIE 1931 chart while susceptible sets show consistent blue color and limited gamut in the CIE 1931 chart. For the resistant sets the QolorAST signals showed a dose response fit indicative of color change. The inset of color change was dependant on the bacterial species where E coli shows the fastest color change at 15 mins followed by MRSA and PA at 30 mins. Since phenotypic QolorAST AST utilizes a metabolic resazurin assay, the difference in the onset of color change can be attributed to difference in metabolic activity between different bacterial strains. This can be conferred through the faster doubling time of E. coli (20 mins) compared to MRSA and PA (30 mins).
To validate QolorAST applicability to operate at the point-of-need using direct human patient samples were evaluated a set of diverse healthy human samples (urine, serum, and nasal swab in Amies transport buffer) spiked with bacteria (Table 10).
For clinical MIC, bacteria were streaked on LB agar overnight and resuspended in a water aliquot. The aliquot was measured, and the bacteria concentration was adjusted to of 106 CFU ml−1 in LB media. A 96 well plate with an antibiotic gradient from 128 μg/ml to 0.125 μg/ml in a 2-fold concentration dilution step was prepared with a positive control with no antibiotics. Next, the bacteria were introduced to each well of the 96 well plate for a final bacteria concentration of 5×105 CFU ml−1. The well plates were cultured overnight the minimum inhibitory concentration was determined as the antibiotic concentration that didn't show any signs of bacterial growth.
To obtain spiked human samples, first human urine was centrifuged at 5000 rcf for 5 minutes to remove large particles. It was subsequently spiked with 5×105 CFU ml−1 and mixed with a solution of Resazurin to a final concentration of 0.02%. No antibiotics were added for the control aliquots, 32 μg/mL Ampicillin was added for resistant aliquots and 16 μg/mL Kanamycin was added for susceptible aliquots. The aliquots were incubated at 37° C. different periods starting from 0 minutes incubation till 60 minutes with a time step of 5 minutes.
Second, human serum samples obtained from healthy donors (Table 10) were heat-inactivated using the protocol suggested by the supplier. Briefly, serum was heated to 56° C. in a water bath for 30 min. subsequently, the samples were removed and left to cool to room temperature. The heat-inactivated human serum samples were spiked with 102 CFU ml−1 MRSA and mixed with a solution of Resazurin to a final concentration of 0.02%. No antibiotics were added for the control aliquots. Since Oxacillin tends to bind to serum proteins (93% of the drug is bounded to serum proteins) 57 μg/mL Oxacillin was added for resistant aliquots to have an effective concentration of 4 μg/mL. Similarly, 1.38 μg/mL Ciprofloxacin (28% of the drug is bounded to serum proteins) was added for susceptible aliquots to have an effective concentration of 1 μg/mL. The aliquots were incubated at 37° C. different periods starting from 0 minutes incubation until 60 minutes with a time step of 5 minutes.
Third, human nasal swab samples were collected from healthy volunteers and preserved in liquid Amies transport media. The samples were spiked with 102 CFU ml−1 MRSA and mixed with a solution of resazurin to a final concentration of 0.02%. No antibiotics were added for the control aliquots, 4 μg/mL Oxacillin was added for resistant aliquots and 1 μg/mL Ciprofloxacin was added for susceptible aliquots. The aliquots were incubated at 37° C. different periods starting from 0 minutes incubation until 60 minutes with a time step of 5 minutes.
According to the Infectious Diseases Society of America (IDSA), 105 CFU ml−1 and 102 CFU ml−1 were spiked in-vivo in urine, serum, and nasal swab of donors to simulate urinary tract infection, wound infection, and pneumonia by MRSA, respectively. Human urine was spiked with different concentrations of E. coli (
A minimum inhibitory concentration (MIC) was performed to study the QolorAST broad applicability in detecting bacterial resistance profiles and ensuring timely and efficient antimicrobial theraby. 12 bacterial species were tested (A. baumannii, E. cloacae, S. pneumoniae, K. pneumoniae, E. faecalis, E. faecium, MRSA, P. aeruginosa, S. agalactiae, E coli mm294, E coli MG1655, and Trimethoprim resistant E coli) and 7 different antibiotics (Ampicillin, Colistin, Gentamicin, linezolid, Ciprofloxacin, Oxacillin, and Trimethoprim). The antibiotics cover major bacterial inhibition mechanisms as inhibition of protein synthesis, cell wall synthesis, and/or DNA synthesis.
QolorAST exploited two methodologies: Manual MIC (MA-MIC) and support vector machine MIC (SVM-MIC). The MA-MIC is based on the direct QolorAST signal for different concentrations of antibioatic where the first concentration shows a no growth signal is considered the MIC (
A. baumannii
E. cloacae
S. pneumonia
K. pneumonia
E. Faecalis
E. Faecium
S. agalactiae
E. coli mm294
E. coli MG1655
E. coli Tri resis.
indicates data missing or illegible when filed
Both the MA-MIC/SVM-MIC results were benchmarked against the conventional broth microdilution (BMD) MIC results. The essential agreement were established between the standard CLSI based MIC and MA-MIC/SVM-MIC where a variation of two fold concentration in the antibiotic MIC was adeemed acceptable. Moreover, the bacterial susceptibility criteria (category agreement) was investigated using standard MIC cut off values to determine the bacteria profile for being susceptible (s), intermediate (I), or resistant (R). For A. baumannii, data suggests heightened resistance to ampicillin (MIC=>128) relative to colistin (MIC=8), as evidenced both manually and through SVM. A similar trend is discernible for, with distinct resistance patterns against the ampicillin and colistin tested. S. pneumonia showed significant resistance to both ampicillin and colistin with MIC values of larger than 128. K. pneumonia's MIC values showed that the bacteria exhibits enhanced susceptibility to ampicillin over colistin. Concurrent observations are made for E. Faecalis and E. Faecium, highlighting their antibiotic susceptibility profiles. In the case of PA, MIC outcomes for ciprofloxacin and gentamicin are 8 and 0.5, respectively. This indicated the potential efficacy of these antibiotics against PA, though clinical breakpoints classify only gentamicin as susceptible. A general trend is that weaker QolorAST signals in the greater concentrations of antibiotic concentrations suggest a lower MIC value and thus, great susceptibility to the bacteria tested. MRSA showcases resistance to oxacillin but susceptibility to ciprofloxacin, with minor discrepancies observed in MIC values between standard and the two QolorAST methods, potentially attributable to experimental variations. MIC values for S. agalactiae, E. coli mm294, E. coli MG1655, and E. coli Tri resis further delineate their antibiotic susceptibility profiles. More specifically, S. agalactiae was found to be resistant to colistin and susceptible to oxacillin, and the same trend found antibiotics for E. coli mm294, E. coli MG1655, and E. coli Tri resis that were tested against ampicillin, ciprofloxacin, oxacillin, and trimethoprim. Notwithstanding variations in QolorAST signal intensity across different concentrations, it was observed that there is a notable coherence between MIC values and QolorAST signals, signifying robust result precision. This accuracy was further proofed through achieving 100% essential agreement (EA) and 100% categirocal agreement (CA) in comparison to the standard CLSI based BMD MIC by both MC-MIC and SVM-MIC showing a superior performance to current devices. Additionally, the optimized SVM algorithm (SVM-MIC) reached an accuracy of 90.6% and receiver operating characteristic (ROC) showed an area under the curve (AUC) of 0.97. QolorAST provided a rapid and accurate (EA 100% and CA100%) MIC testing solution compatible with a wide range of bacterial species and antibiotics.
To evaluate the QolorAST performance in clinical settings, a double-blinded clinical study at the McGill University Health Center (MUHC) was performed. 47 patients suspected of urinary tract infections, admitted to MUHC emergency clinics between August 2022 to August 2023 were tested. Specimens were consented, collected from the patients and the experiments were done in microbiology labs in MUHC. Each specimen were divided in 3 to be tested in parallel in the central laboratory of the hospital (quantification, identification and MIC), in the research laboratory (quantification, identification and MIC for 2 antibiotics) and with QolorAST genotypic/phenotypic assays in microbiology lab at MUHC. The device has 24 chambers where the media/resazurin/antibiotic mixes as well as the patient samples are injected mixed. In this part the effect of two main antibiotics including Nitrofurantoin at concentrations of 256, 128, 64, 32, 16, 8, 4, 2, 1, 0.5, and 0.25 μg/mL and Ciprofloxacin at concentrations of 32, 16, 8, 4, 2, 1, 0.5, 0.25, 0.125, 0.06, and 0.03 μg/mL and control without antibiotic were investigated. Therefore, reagent loading port was preloaded with media, specific concentration antibiotic and resazurin at final concentration of 0.25 mg/mL and urine was added to urine loading port. Thereafter, by pressing the pushing button the urine was mixed with the rest of solution at a volumetric ratio of 1:4 which would be the starting point for incubation and imaging of the samples. The samples were incubated at 37° C. at different timepoints starting from 0 to 60 minutes with a time interval of 10 minutes and the imaging was performed as described above.
For the standard testing the urine specimens were diluted 1/1000 using CamHB and both the original concentration and a 1/1000 dilution of the specimen were plated on BAP with a 1 μL calibrated loop and incubated for 18-20h at 37 C with 5% CO2. The CFU/L concentration is then calculated for all the isolates coming from each specimen. For each isolate, the MIC is also determined for 2 antibiotics: Ciprofloxacin (0.03 to 128 μg/mL) and Nitrofurantoin (0.125 to 256 μg/mL). The isolates quantification is standardized to 1×106 CFU/mL prior to be incubated with the range of different concentrations of antibiotics diluted in CAMHB for MIC. At the end of the process, the inoculum quantification is checked by counted CFU from serial dilutions of the inoculum. A growth positive control (inoculum with no antibiotic), a wild type strain of E. Coli (MG1655) and E. Faecalis are also included and processed in the same way as the isolates for each experiment. Finally, each isolate is plated on BAP to control their monoculture state and to be identified by MALDI.
MALDI-TOF MS was performed using an Ulttraflextreme TOF/TOF mass spectrometer (Bruker Daltonics, Leipzig, Germany), in accordance with the manufacturer's instructions, and with Flexcontrol software 3.4 Build 169.5 (Bruker Daltonics) for the automatic acquisition of mass spectra in the linear positive mode within the range 2 to 20 kDa. The mass spectrometer was calibrated at each run by using the BTS standard (Bruker Daltonics), as described by the manufacturer. Calibration masses were: RL29, 3636.8, RS22, 5095.8 Da; RS34, 5380.4 Da; RS33meth, 6254.4 Da; RL29, 7273.5 Da; RS19, 10229.1 Da; RNase A, 13682.2 Da; myoglobin, 16952.5 Da.
For optimization studies we employed physical characterization techniques as atomic force microscopy (AFM) and scanning electron microscopy (SEM). For AFM we used Bruker, MultiMode8 system while we used FEI Quanta 450 environmental scanning electron microscope for SEM.
Optical characterization was carried out with a Lambda 750 UV/Vis/NIR Spectrophotometer (PerkinElmer). The incident and collected light beams had normal incidence to the platform. Also, we performed optical imaging under a highly controlled environment using Nikon Eclipse LV150 (Nikon) with a ×100, 0.9 NA air objective and Nikon digital sight ds-fi1 CCD camera. The fundamental fluid flow characteristics of the microfluidic cartridge were studied using COMSOL Multiphysics (V5.6).
Samples were tested in parallel using two methods: conventional (Matrix-assisted laser desorption/ionization-time of flight—MALDI/TOF—for bacterial identification and CLSI BMD method for AST) and QolorAST (genotypic for bacterial ID and phenotypic AST). Thirteen patients suspected of UTI (27.66%) did not show bacterial growth while nine (19.15%) and twenty-five (53.19%) specimens showed growth patterns in consistence with fastidious and no-fastidious (E. coli) bacterial strains respectively (
Sdsd Using QolorAST phenotypic assay the specimens were tested with two antibiotics Ciprofloxacin (Cip) and Nitrofurantoin (Nit) to determine the minimum inhibitory concentration (MIC) and drug resistance profiles. The SVM produced a prediction for each antibiotic at the concentrations tested (Table 12), six out of twenty five (24%) of the E. coli infected samples shown intermediate or high resistance to ciprofloxacin. QolorAST SVM readout shows a statistically significant contrast (P<0.001) between negative, ciprofloxacin susceptible and ciprofloxacin resistant specimens (
indicates data missing or illegible when filed
All the specimens were susceptible to nitrofurantoin (Table 12). QolorAST SVM probabilities shows a statistically contrast (P=0.013) between negative and nitrofurantoin susceptible specimens (
The limit of detection measured for each bacteria is summarized in Table 13.
A. baumannii
E. cloaca
K. pneumoniae
S. pneumoniae
E. Faecalis
E. Faecium
P. aeruginosa
S. agalactiae
E. coli mm294
E. coli MG1655
E. Coli Tri resis
Overall, QoloarAST had an average essential agreement (EA) of 90% and an average categorical agreement (CA) of 96%. Accordingly, QolorAST complies with the International Organization for Standardization (ISO 20776-2:2021) criteria for MIC AST devices of having EA ≥90%.
In conclusion, QolorAST is an automated point-of-need system coupled with a genotypic LAMP for bacterial identification (ID) and a phenotypic resazurin reduction assay for antibiotic susceptibility profiling. QolorAST uniquely utilizes nanoplasmonic colorimetric structures to enable rapid detection of the onset of color change and identify AST in less than 30 mins in comparison with 48 hours for the current conventional methods and on par with FDA approved device. The first key feature of QolorAST system is a microfluidic cartridge that serves as an automated, portable and multiplex testing platform substitute for the traditional 96-well format platform by incorporating the plasmonic nanostructures into a 3D printed microfluidic device with addressable compartments including sample collection, preparation and detection units. This allowed for direct testing of different concentrations of antibiotics per sample without processing or pre-culture.
The core principle of QolorAST signal generation is based on the reduction of the dark blue resazurin to light pink resorufin on top a plasmonic color printing platform where the plasmonic color is initially inhibited by the dark blue resazurin but is fully detectable at the early stages' reduction process. The resazurin reduction can be initiated through pH change (genotypic QolorAST assay) or due to the metabolic activity of viable bacterial cells (QolorAST phenotypic AST assay). The rapid plasmonic colorimetric detection was demonstrated to lead to rapid genotypic ID (<15 mins) and phenotypic AST profiling (<30 mins). Sensitivity in plasmonic color-printing based colorimetric sensing largely relies on the resolution of pigmented color that is captured, while the rapidity of the sensor depends on how fast the color change occurs according to the chemical reactions. The light interacting with resonances associated with the discrete harmonic energy states are structurally engineered, and could offer new opportunities to enhance the colorimetric sensing. The plasmonic enhanced color-generation strategy involves the patterning of various geometrical metallic nanostructures and investigates the hue and gamut of colors experimentally.
The second feature of QolorAST is an autonomous imaging box to allow integrated incubation, fluid actuation and imaging. The imaging box utilizes a unique cyclic filtration/actuation process to allow repeated and precise control over the temporal sample release. Third feature of QolorAST is an automated machine interfaced data collection and interpretation that employs a simple camera to capture the color output of the fluidic device and an image analysis software to easily analyze and interpret the data to automate the color interpretation. This improved the potential of remote applications.
Further, QolorAST was validated in a double-blinded clinical study and established its performance with 47 clinical specimens from patients with suspected UTI achieving 90% essential agreement and 96 category agreement %. It presently demonstrated a rapid test method that is versatile and provides a broad color change for detection of a wide range of metabolically active bacteria in a portable and automated fashion. QolorEX combines rapidity, clinically relevant sensitivity, specificity, and versatility for the detection of a variety of bacteria and minimum antibiotic inhibitory concentrations that is deployable at the point of care in low-resource, remote or congregate settings. Through combining automated genotypic ID and phenotypic AST, QolorAST provides a complete solution to reduce the time-to-result for AST at a low-cost has significant impact on the global health system.
This example provides the use of the suction membrane in a smartphone-operated and additively manufactured multiplexed electrochemical device (AMMED) for the portable detection of biomarkers in blood and saliva. AMMED was developed to employ a suction-based microfluidic system relying on a single-trigger mechanism alongside a filtered-based sample collection approach, a dimeric DNA aptamer-based gold nanostructured (GNS) electrochemical biosensor (GNS aptasensor), and a customized potentiostat. Indeed, the device mainly utilizes a smartphone-controlled portable potentiostat featuring the multiplexed option to run voltammetry and electrochemical impedance spectroscopy (EIS) measurements.
As illustrated in
The AMMED components were fabricated based on additive manufacturing (AM) (3D-printing) techniques, making the whole process almost cleanroom-free and scalable. In particular, the microfluidic chip containing the electrodes and microfluidic parts was fabricated by AM. This approach obviates the necessity for time-consuming, challengeable, and expensive traditional methods such as photoresist coating, photolithography, and solvent-based lift-off.
The biosensing performance of the proposed device was investigated by the detection of a diagnostic representative biomarker in the relevant biofluids, including saliva and diluted whole blood. Spike protein (S-protein) of the SARS-COV-2 omicron variant was selected as the representative biomarker since it has been proven to be a valuable antigen for noninvasive, accurate, and rapid diagnosis of COVID-19. The intermediates obtained during the additive manufacturing process are presented in
In greater details, first, ethanol and dionized water were used for cleaning the glass substrate prior to mask attachment and gold deposition. Then, a 700 μm thick 3D-printed mask (fabricated using stereolithography (SLA) 3D-printer Form 3, Formlabs) featuring the patterns for the three-multiplexed electrodes was attached to the glass substrate using a pressure sensitive adhesive (PSA) film. This 3D-printed mask allowed for the fabless patterning of gold electrodes in a single step. Subsequently, a 20 nm film of Chromium, as the adhesion layer, was deposited onto the masked substrate followed by a 150 nm layer of gold using electron beam evaporator Temescal BJD 1800. The 3D-printed mask was removed after the deposition and the glass slide substrate was diced into individual chips using a Disco DAD 3240 dicing saw. The microfluidic channels and detection chambers of the test chip were fabricated based on SLA 3D printing (at 50 μm resolution on the z-axis). Finally, the 3D-printed microfluidic part was mounted onto the glass chip using PSA to control sample delivery, encapsulate the device, and confine the sensing area of electrodes. The suction membrane was prepared with PDMS as described in Example 1 and then bonded to the microfluidic chip. This was followed by depositing dimeric DNA aptamer-based gold nanostructured (GNS) electrochemical biosensor (GNS aptasensor) on the gold surface.
In this example, the microfluidic chip 100 does not have a filter barrier as the filter embedded within the inlet apparatus is sufficient for providing a sample that can be analyzed by electrical detection.
A multiplexed detection performance of AMMED through biosensing the SARS-COV-2 S-protein in media was performed. The AMMED device was tested for the detection of SARS-CoV-2 S-protein ranging from 1-10 000 pg ml−1 in buffer (
10 human saliva samples (5 samples from adult patients with COVID-19 symptoms, such as fever, fatigue, and dry cough, and tested with RT-qPCR) and 5 samples from healthy controls were supplied by the University Health Network's PRESERVE-Pandemic Response Biobank for testing on the assay (REB #20-5364). Free authorization and consent forms were signed by patients, and their clinical samples were collected according to the laboratory regulation. Accordingly, the AMMED device was further challenged with real saliva samples belonging to patients who had previously tested positive (n=5) or negative (n=5) on the standard polymerase chain reaction (PCR) diagnostic tests and successfully achieving the same diagnostic as that of PCR (
This disclosure claims the priority from U.S. provisional application No. 63/382,904 filed Nov. 9, 2022 and incorporated herein by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 63382904 | Nov 2022 | US |
