Multiple RNA analysis for blood-based disease diagnosis and typing
By utilizing the EZ-READ platform and RNA-responsive sensors and microfluidic devices, the limitations of existing technologies that require invasive brain biopsies and RNA detection for GBM genotyping have been overcome, enabling highly sensitive GBM diagnosis and genotyping.
Patent Information
- Application Number
- CN202480034767.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-03-27
- Filing Date
- 2024-03-27
- Publication Date
- 2026-03-17
AI Technical Summary
Existing GBM typing methods require invasive brain biopsies, and current technologies have limitations in measuring low concentrations and diversity of circulating RNA targets, making it difficult to achieve reliable RNA detection and multiplexing.
The EZ-READ platform was developed, which utilizes an RNA-responsive sensor to connect a catalytic nanoparticle complex to a specific single-stranded DNA probe of the target RNA. Digital quantification and diagnosis are achieved using a microfluidic device. After the sensor recognizes the target RNA, it releases the catalytic nanoparticle complex to enhance the signal.
It enables reliable detection of different RNA subtypes of various sizes in unprocessed samples with high sensitivity, and can complete GBM diagnosis and subtyping within 30 minutes, accurately diagnosing patients and distinguishing different tumor subtypes.
Smart Images

Figure CN121693662A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present invention relates to a RNA-responsive sensor comprising a catalytic nanoparticle complex attached to a magnetic bead by a specific single-stranded DNA probe to a target RNA, wherein the single-stranded DNA probe will form a duplex with the target RNA, and wherein in the presence of a duplex-specific nuclease recognizing the RNA-DNA duplex, the DNA linker is cleaved, thereby: i) releasing the catalytic nanoparticle complex, and ii) releasing the target RNA to form a DNA-RNA duplex with another RNA-responsive sensor. The present invention further relates to a method for detecting RNA using the sensor and catalytic enzyme, wherein digital quantification, diagnosis and typing of a disease is achieved with a microfluidic device. BACKGROUND
[0002] Identification of disease subtypes has broad applications in patient stratification and treatment selection. For example, the most common malignant brain tumor type, glioblastoma (GBM), is characterized at the transcriptional level into three glioma intrinsic (GI) subtypes, namely, proneural, classical, and mesenchymal, which exhibit different prognoses and treatment responses. Most importantly, this tumor is susceptible to subtype conversion; GBM subtypes have been found to shift during relapse, particularly to the mesenchymal subtype, which is a highly aggressive form associated with the worst prognosis. Given the dynamic nature of GBM subtypes and their differential responsiveness to treatment, a minimally invasive method that can characterize and monitor disease subtypes can bring new clinical opportunities for personalized treatment. However, current GBM typing requires an invasive brain biopsy for RNA analysis; for continuous monitoring, the complexity and morbidity of repeated biopsies make this approach clinically unacceptable.
[0003] Extracellular vesicles (EVs) have recently emerged as an attractive circulating biomarker for GBM tumors (see Skog, J. et al. (2008), Shao, H. et al. (2012), Batool, S. M. et al. (2022)). EVs are nanoscale membrane vesicles actively secreted by a variety of cells, especially rapidly dividing cancer cells (see Théry, C. et al. (2009), van Niel, G., et al. (2018), Yekula, A. et al. (2020)). As promising blood biomarkers, they easily cross the blood-brain barrier and carry a variety of nucleic acids (e.g., short miRNAs and long mRNAs) from their parent cells. Despite the RNA composition being rich for blood-based GBM typing, the clinical translation of EVs is challenging, mainly due to the limitations of the existing technologies in assaying these circulating RNA targets at low concentrations and diversity (see Shao, H. et al. (2018), Théry, C. et al. (2018)).
[0004] For example, to detect rare RNA biomarkers, conventional assays (e.g., polymerase chain reaction, PCR) mainly rely on target amplification, in which target nucleic acid sequences need to be replicated in large quantities before detection (see Shao, H. et al. (2015), Figueroa, J. M. et al. (2017), Reátegui, E. et al. (2018)). This approach requires extensive processing (e.g., reverse transcription and heat cycle-based sequence amplification) and precise sequence design (e.g., tailored primers and specialized probe pairs), thus poses challenges to reliable assays and multiplexed detection, especially for different RNA targets of varying lengths and sequences. To improve detection adaptability, several new hybridization-based methods (see Tavallaie, R. et al. (2018), Labib, M. et al. (2018)) have been developed to reduce the dependence on nucleic acid amplification. However, these methods generally have limited sensitivity and cannot be easily applied to the analysis of rare RNA targets.
[0005] There is an urgent need to provide a platform to at least improve one or more of the above-mentioned problems in the current RNA detection system. SUMMARY
[0006] The inventors developed a specialized assay technique to directly sense and catalytically enhance signal reactions for RNA targets, enabling reliable quantification of different EV RNA for GBM subtyping. The technique was named Enzyme ZIF-8 Complex for Regeneration and Catalytic Digital Detection of RNA (EZ-READ), which utilizes RNA-responsive sensors for direct activation and catalytic digital quantification. Each sensor comprises a hybrid nanoparticle complex, i.e., encapsulating a protease within a metal-organic framework, to configure strong catalytic activity and robust protection, even against potential inhibition in complex biological environments. The sensor recognizes the target RNA and regenerates activation; when the RNA hybridizes with the sensor, the catalytic complex is released in a target-recoverable manner, thereby enhancing signal sensing. The released complex can catalyze a strong chemical fluorescence reaction on its own; when individually allocated and reacted in a fractal-branching microfluidic chip, the released complex can enable digital quantification of different RNA targets.
[0007] The EZ-READ platform utilizes its regenerative sensing and catalytic detection to achieve superior performance without target amplification. It enables programmable and reliable RNA detection of RNA subtypes (miRNAs and mRNAs) of varying sizes directly in minimally processed sample lysates. The technology establishes a sensitive detection limit (<10 RNA copies) and can be completed within 30 minutes. Using the EZ-READ platform, the inventors determined miRNA and mRNA targets in blood samples from GBM patients and control subjects. Based on these clinical assays, the inventors constructed a multilevel decision model to establish circulating RNA signatures for GBM diagnosis and subtyping. This technology not only accurately diagnoses patients but also effectively differentiates between different tumor subtypes.
[0008] The embodiments of the present invention are described in the following statements:
[0009] 1. An RNA-responsive sensor comprising a catalytic nanoparticle complex attached to a magnetic bead via a target RNA-specific single-stranded DNA probe, wherein the single-stranded DNA probe forms a double strand with the target RNA, and wherein, in the presence of a double-stranded-specific nuclease that recognizes the RNA-DNA double strand, a DNA linker is cleaved, thereby:
[0010] i) Release of catalytic nanoparticle complex, and
[0011] ii) Release the target RNA to form a DNA-RNA duplex with another RNA-responsive sensor.
[0012] 2. The RNA-responsive sensor according to Statement 1, wherein the catalytic nanoparticle complex comprises one or more enzymes encapsulated in a metal-organic framework.
[0013] 3. The RNA-responsive sensor according to Statement 2, wherein the metal-organic framework is selected from the group consisting of: zeolite imidazolate backbone-8 or 2-methylimidazolate zinc salt (ZIF-8), zeolite imidazolate backbone-67 or 2-methylimidazolate cobalt salt (ZIF-67), 1,4-phthalic acid (BDC), and 1,3,5-benzenetricarboxylic acid (BTC).
[0014] 4. The RNA-responsive sensor according to any one of statements 2 or 3, wherein the one or more enzymes are selected from the group comprising horseradish peroxidase (HRP), acetylcholinesterase, dehydratase, catalase, cellulase, cytochrome c, dehydrogenase, epoxide hydrolase, glucoamylase, glucose dehydrogenase, glucose oxidase, hemin, hydrolase, invertase, laccase, lipase, microperoxidase-11, myoglobin, organophosphate dehydratase, pectinase, protease, soybean epoxide hydrolase, trypsin, tyrosinase, urease, xylanase, a-amylase, b-galactosidase, b-glucosidase.
[0015] 5. The RNA-responsive sensor according to any one of statements 1 to 4, wherein the target RNA is selected from the group comprising mRNA, miRNA, IncRNA, rRNA and tRNA.
[0016] 6. The RNA-responsive sensor according to any one of statements 1 to 5, wherein the target RNA is present in a body fluid sample, preferably a blood sample.
[0017] 7. The RNA-responsive sensor according to statement 6, wherein the body fluid sample is from a diseased subject.
[0018] 8. The RNA-responsive sensor according to statement 7, wherein the body fluid sample is from a subject afflicted with glioblastoma.
[0019] 9. The RNA-responsive sensor according to statement 8, wherein the specific single-stranded DNA probe targets an RNA selected from the group comprising miR-17-5p, lysyl oxidase (LOX) mRNA and solute carrier organic anion transporter family member 3A1 (SLCO3A1) mRNA or a combination thereof.
[0020] 10. The RNA-responsive sensor according to statement 8, wherein the specific single-stranded DNA probe targets an RNA selected from the group comprising miR-21-5p, miR-27a-3p and aldehyde dehydrogenase 1A3 (ALDH1A3) mRNA or a combination thereof.
[0021] The person skilled in the art will understand that catalytic enzymes can be selected to generate a signal from colorimetric substrates and other types of substrates.
[0022] 11. A method of determining the abundance of a biomarker in a sample, the method comprising:
[0023] a) combining a biomarker-responsive sensor with the sample, the biomarker-responsive sensor being configured to release a catalytic nanoparticle complex in response to the biomarker;
[0024] b) introducing the supernatant of the sample and biomarker-responsive sensor combination into a microfluidic device having a plurality of reaction wells;
[0025] c) introducing a colorimetric substrate or other signaling substrate into the reaction wells, wherein the released catalytic nanoparticle complex catalyzes a reaction with the colorimetric substrate or other signaling substrate, thereby generating a detectable reaction signal;
[0026] d) acquiring an image of the microfluidic device, the image capable of showing signal-positive and / or signal-negative reaction wells;
[0027] e) determining from the image a count of reaction wells that generated a detectable signal; and
[0028] f) determining from the count of reaction wells that generated a detectable reaction signal the abundance of biomarker in the sample.
[0029] 12. The method of statement 11, wherein the substrate is a chemiluminescent substrate and the detectable reaction signal is fluorescence.
[0030] 13. The method of statement 11 or 12, wherein the biomarker-responsive sensor is an RNA-responsive sensor according to any one of statements 1 to 10.
[0031] 14. The method of any one of statements 11 to 13, wherein the microfluidic device comprises a plurality of fractal branching paths extending from the inlet to respective reaction wells in the plurality of reaction wells.
[0032] 15. The method of statement 14, wherein the plurality of fractal branching paths are configured to have substantially equal fluidic resistance.
[0033] 16. The method of statement 14 or statement 15, wherein the plurality of fractal branching paths are configured to have substantially equal length between the inlet and each respective reaction well.
[0034] 17. The method of any one of statements 14 to 16, wherein the branching angle of each fractal branching path ranges from 85 degrees to 95 degrees, or preferably ranges from 89 degrees to 91 degrees.
[0035] 18. The method of any one of statements 11 to 17, further comprising sealing each reaction well after introducing the chemiluminescent fluorescent substrate to isolate the fluorescent reaction in each respective reaction well.
[0036] 19. The method of any one of statements 11 to 18, wherein determining from the fluorescent image a count of reaction wells that generated a fluorescent signal comprises comparing the fluorescent intensity of each respective reaction well to a threshold value and determining the count as the number of reaction wells having a fluorescent intensity above the threshold value.
[0037] 20. An in vitro method of detecting and / or classifying a disease suffered by a subject, comprising performing the method of any one of statements 11 to 19 on a body fluid sample from the subject and using the abundance of at least one biomarker associated with the disease to detect the disease and / or to classify the disease.
[0038] 21. The method of statement 20, wherein the disease is glioblastoma, and
[0039] i) the RNA-responsive sensor is defined according to statement 9 for detecting the disease, or
[0040] ii) the RNA-responsive sensor is defined according to statement 10 for classifying the glioblastoma as mesenchymal subtype or pre-neural / classic subtype.
[0041] 22. A microfluidic device comprising a plurality of fractal branching paths extending from an inlet to respective ones of a plurality of reaction wells.
[0042] 23. The microfluidic device of statement 22, wherein the plurality of fractal branching paths are configured to have substantially equal fluidic resistance.
[0043] 24. The microfluidic device of statement 22 or 23, wherein the plurality of fractal branching paths are configured to have substantially equal length between the inlet and each respective reaction well.
[0044] 25. The microfluidic device of any one of statements 22 to 24, wherein the branching angle of each fractal branching path ranges from 85 degrees to 95 degrees, or preferably ranges from 89 degrees to 91 degrees.
[0045] 26. A kit for screening a body fluid sample for the presence of RNA indicative of a disease, wherein the RNA is differentially expressed in the diseased state compared to a reference sample representative of a subject not suffering from the disease, the kit comprising the RNA-responsive sensor of any one of statements 1 to 10.
[0046] 27. The kit of statement 26, wherein the RNA-responsive sensor comprises a specific single-stranded DNA probe for a target RNA that will form a duplex with the target RNA, wherein the target RNA is selected from:
[0047] a) the group comprising miR-17-5p, LOX mRNA and SLCO3A1 mRNA, or
[0048] b) the group comprising miR-21-5p, miR-27a-3p and ALDH1A3 mRNA.
[0049] 28. A glioblastoma screening kit for screening a body fluid sample for the presence of RNA indicative of a glioblastoma, which is clinically pathologically of mesenchymal, proneural or classical subtype, wherein the RNA is differentially expressed in the diseased state compared to a reference sample representing a subject not suffering from the disease, wherein the kit comprises one or more RNA-responsive sensors as described in statement 10.
[0050] 29. The kit according to any one of statements 26 to 28, further comprising a microfluidic device according to any one of statements 22 to 25.
[0051] 30. The kit according to any one of statements 26 to 29, further comprising:
[0052] i) a duplex-specific nuclease, and / or
[0053] ii) a substrate reactive with the catalytic nanoparticle complex released from the RNA-responsive sensor. BRIEF DESCRIPTION OF DRAWINGS
[0054] Figure 1The platform of the present application for reliable analysis of circulating RNA is shown. (a) EZ-READ regenerates, converts and catalytically enhances rare RNA targets with RNA-responsive sensors. Each sensor is loaded with catalytic nanoparticles, i.e. horseradish peroxidase (HRP) encapsulated within metal-organic frameworks (ZIF-8) catalytic nanoparticles, via specific DNA probes. Therefore, the scaffold HRP@ZIF-8 nanoparticles have strong catalytic activity and confer robust protection against the environment. During the regenerative sensing process, when RNA targets (e.g. miRNA and mRNA) hybridize specifically, the DNA linker bound to RNA is selectively cleaved by duplex-specific nuclease (DSN), and the sensor is directly activated to release HRP@ZIF-8 nanoparticles in a target-recoverable manner. When partitioned within fractal-branching microfluidic chips, the released nanoparticles can individually catalyze strong chemical fluorescence reactions to achieve digital quantification. (b) RNA-responsive sensors. Scanning electron micrographs of assembled sensors confirm the high-density coverage of HRP@ZIF-8 nanoparticles. Experiments were repeated independently three times with similar results. (c) Catalytic digital quantification. To take advantage of the strong activity of HRP@ZIF-8 nanoparticles for digital counting, the inventors developed fractal-branching microfluidic chips with resistance-matched microwells to achieve uniform nanoparticle partitioning and independent catalysis. Left: exploded view of the microfluidic chip. Middle: schematic of catalytic digital quantification. HRP@ZIF-8 nanoparticles, reaction substrates and air were sequentially introduced into the chip by vacuum loading to achieve independent catalysis. Right: fluorescence images of the reaction chip for digital counting. Experiments were repeated independently three times with similar results. (d) Direct and reliable RNA detection by EZ-READ. Compared with conventional methods (e.g. PCR) that require target amplification and have unstable amplification efficiency, EZ-READ bypasses all steps of conventional PCR detection and achieves direct sensing and catalytic signal enhancement. (e) Clinical applications. Taking advantage of its programmable detection, the inventors applied EZ-READ to assay various RNA subtypes in clinical blood samples and developed a multi-step classification model to accurately diagnose GBM patients and molecularly subtype tumors.
[0055] Figure 2Catalytic digital quantification of RNA with EZ-READ is shown. (a) Enhanced performance of HRP@ZIF-8. HRP@ZIF-8 shows more efficient catalysis, higher loading capacity, and improved enzyme activity compared to HRP-functionalized polystyrene beads with equal amount of protease. HRP@ZIF-8 shows more robust activity recovery after rehydration when treated with lyophilization. (b) Temperature stability of HRP@ZIF-8. HRP@ZIF-8 can tolerate high temperature and shows negligible loss of enzyme activity compared to free HRP. (c) Regeneration sensing. The inventors determined the release of FITC-labeled HRP@ZIF-8 nanoparticles from the sensor over time when these nanoparticles were incubated with different concentrations of RNA targets (high concentration: 10 pmol, and low concentration: 3 pmol). The fluorescence signal of the supernatant was determined. EZ-READ shows a continuous release of nanoparticles, reflecting the amount of RNA input. (d) Correlation with RNA target concentration. The number of released HRP@ZIF-8 nanoparticles is linearly related to the target concentration. (e) Schematic of catalytic digital quantification. The inventors performed a series of operations, namely vacuum loading, HRP@ZIF-8 dispensing, substrate application, air sealing, and fluorescence reading. The sample RNA concentration can be determined by counting the number of fluorescent microwells. All assays were performed in triplicate (n = 3 independent experiments), and the data are presented as mean ± standard deviation in panels b-f and as mean in panel g.
[0056] Figure 3 Comparison with conventional target amplification analysis is shown. Conventional methods (e.g., PCR) require multiple processing steps to achieve exponential target amplification. Due to the influence of various factors (e.g., inhibitors in biological samples) on amplification efficiency, a small difference in amplification efficiency grows exponentially, leading to a huge difference in the final assay value. This is especially true for PCR detection of miRNAs, as short miRNA molecules need to be elongated and pre-amplified before PCR detection. On the other hand, EZ-READ utilizes linear regeneration sensing to directly release HRP@ZIF-8 nanoparticles after target RNA hybridization. These released particles have strong enzyme activity, thus achieving catalytic digital quantification. Therefore, this method has robust resistance to efficiency fluctuations, can be applied to different RNA subtypes (e.g., miRNA and mRNA), and achieves direct and reflective quantification of biomarker signatures.
[0057] Figure 4A schematic diagram of the EZ-READ clinical analysis using regression methods is shown. (a) Evaluation of EZ-READ clinical assays and regression models. The EZ-READ platform was used to measure circulating RNA biomarkers in GBM and control plasma samples. Single-step multinomial regression and multi-step multiple linear regression scoring models were developed using a training cohort of clinical samples for disease diagnosis and disease subtype classification. Model performance was independently validated in separate validation patient cohorts using the area under the curve (AUG) in receiver operating characteristic (ROC) analysis. (b) Schematic diagram of single-step multinomial regression and multi-step multiple linear regression scoring models. MS, mesenchymal subtype; PN, preneurial subtype; CL, classical subtype.
[0058] Figure 5 Characterization of the HRP@ZIF-8 hybrid complex is shown. (a) Hydrodynamic diameter of the synthesized HRP@ZIF-8 nanoparticles. Dynamic light scattering analysis was used to determine the hydrodynamic diameter. The inset shows a transmission electron micrograph of the prepared HRP@ZIF-8 nanoparticles. HRP is embedded in ZIF-8 and immunolabeled with 10 nm gold nanoparticles. Scale bar: 100 nm. The experiment was independently repeated three times, and the results were similar. (b) Powder X-ray diffraction (PXRD) analysis confirmed the high crystallinity of the 3D epitaxial ZIF-8. (c) Fourier transform infrared (FT-IR) spectra of ZIF-8, HRP, and HRP@ZIF-8. The HRP@ZIF-8 nanoparticles showed multiple characteristic peaks, thus confirming their hybrid nature. (d) Enhanced enzyme activity. HRP@ZIF-8 and HRP of equal concentration were treated with increased concentrations of hydrogen peroxide. HRP@ZIF-8 exhibited improved enzyme kinetics at different substrate concentrations. (e) Retention of enzyme activity. When incubated in DSN buffer, HRP@ZIF-8 maintained strong enzymatic activity, while pure HRP exhibited minimal activity. (f) Regulation of zeta potential. To reduce the nonspecific binding of negatively charged DNA probes to HRP@ZIF-8, the synthesized particles were treated with polyacrylic acid (PAA) to reduce their zeta potential from 11.7 mV to -44 mV. (g) Adjustment of PAA and PEG linkers. To enhance sensing, PEG linkers were added to reduce steric hindrance and improve target accessibility. When treated with target RNA, the sensor prepared with both PAA coating and PEG linkers showed the highest sensing and release of HRP@ZIF-8 nanoparticles. All assays were performed triplicate (n=3 independent experiments), and data are presented as mean ± standard deviation in Figures de and g.
[0059] Figure 6RNA target recovery is illustrated. (a) Schematic diagram of the differences in enzyme mechanisms (left). DSN recognizes RNA-DNA heteroduplexes and cleaves the DNA strand in a random, repetitive manner to produce short DNA products (≤6 nt). BstNI, as an example of an enzyme mediating sequence-specific cleavage, recognizes specific cleavage sites (CC / WGG) in RNA-DNA heteroduplexes to cleave DNA into different products. Experimental verification (right). The inventors incubated the same RNA-DNA heteroduplexes with DSN or BstNI and characterized the DNA products by polyacrylamide gel electrophoresis (10% PAGE). DSN did not produce different DNA products, while BstNI produced different bands (17 nt DNA products) corresponding to the specific BstNI cleavage sequence. Note that DNA products ≤6 nt are too short for PAGE visualization. The experiment was independently repeated three times with similar results. (b) Two-step sequential reaction verifies target recovery. In the first step, an RNA-DNA heteroduplex (composed of a 1:1 mixture of RNA target and DNA probe) was reacted with excess DSN. The cumulative signal was measured over time, showing a rapid plateau with low saturation signal. In the second step, after signal saturation, the inventors stopped the reaction with excess specific DNA probe, mixed DNA probe, or buffer. Only the addition of specific DNA probe restored signal generation and yielded a high signal, while the mixed control and buffer control produced minimal signal increases. Therefore, the observed signal increase is primarily due to RNA target cycling, as the signal output is determined by the amount of DNA probe available to form a heteroduplex with a fixed RNA target input. All measurements were performed three times (n=3 independent experiments), and the data are presented as mean ± standard deviation in Figure b.
[0060] Figure 7An RNA-responsive regeneration sensor is shown. (a) Enzyme conformational changes. Compared to the original HRP, the ZIF-8 scaffold (HRP@ZIF-8) provides a more accessible active site for the enzyme, as evidenced by a characteristic red shift in the fluorescence spectrum of HRP@ZIF-8. (b) Enhanced enzyme activity. Michaelis-Menten kinetics of the original HRP and HRP@ZIF-8, when treated with increased substrate concentrations, indicate that HRP@ZIF-8 exhibits improved substrate affinity and catalytic activity. (c) Robust protection against environmental influences. Both HRP@ZIF-8 and the original HRP were incubated in double-stranded nuclease-specific (DSN) buffer. ZIF-8 encapsulation protected the embedded enzyme and maintained its activity in complex environments. However, the original HRP showed negligible residual activity in DSN buffer. The robustness of HRP@ZIF-8 allows for the merging of DSN and its associated buffer to achieve regeneration sensing in a single reaction. (d) Regeneration sensing of RNA targets. EZ-READ exhibits enhanced responsiveness to RNA targets but no responsiveness to DNA targets. When incubated with increased amounts of RNA targets, EZ-READ only shows a reactive signal in the presence of DSN. (e) Programmability for different RNA isoforms. The EZ-READ sensor is designed to carry DNA probes of different sizes. When incubated with specific RNA targets of different lengths (20 nt to 200 nt), the sensor produces a strong signal and shows minimal background signal compared to confounding controls. (f) Mismatch specificity of EZ-READ detection. When incubated with synthetic RNA targets with different numbers of mismatches, the system exhibits good performance in distinguishing rare mismatches. (g) Specificity of EZ-READ detection. Developed for the detection of both GBM-related mRNAs (e.g., LOX, SLCO3A1, COL1A2) and miRNAs (e.g., miRNA-17-5p, miRNA-21-5p, miRNA-27a-3p). Heatmap signals were analyzed and normalized (rows). All assays were performed triplicate (n=3 independent experiments), and data are presented as mean ± standard deviation in Figure af and as mean in Figure g. au is in arbitrary units.
[0061] Figure 8The catalytic digital quantification of the RNA target is illustrated. (a) Schematic diagram of catalytic digital quantification. The inventors developed a microfluidic platform to digitally detect the strong catalytic activity of HRP@ZIF-8 nanoparticles. Specifically, the platform comprises a resistance-matched micropore array of fractal branches to enable stepwise loading of HRP@ZIF-8 particles, reaction substrate, and air sealant, thereby achieving uniform particle compartmentation and independent catalytic readouts. The left inset shows a photograph of the microfluidic chip loaded with fluorescently labeled HRP@ZIF-8 particles. (b) Optimization of stepwise reaction. The inventors monitored the liquid occupancy in the microfluidic channels and micropores to ensure loading efficiency and reaction integrity. (c) EZ-READ detection of RNA. The RNA target was incubated directly with an EZ-READ sensor to regenerate and release HRP@ZIF-8 nanoparticles. The released particles were uniformly distributed within the microfluidic micropores and reacted to generate a chemiluminescent signal. Using fluorescently labeled HRP@ZIF-8 nanoparticles, the inventors demonstrated an independent chemiluminescence reaction closely matched to the nanoparticle occupancy rate. (d) Performance validation for RNA quantification. The inventors treated EZ-READ with varying amounts of input RNA. The number of fluorescent wells (i.e., positive wells) obtained was not only closely correlated with the nanoparticle occupancy rate but also accurately determined the amount of RNA input. (e) Sensitivity of EZ-READ detection. The limit of detection (LOD) was determined by titrating a known amount of the short miRNA target (miR-222-3p) and counting the number of fluorescent wells. EZ-READ showed a LOD of 9 copies for the miRNA target. The LOD was defined as 3 × standard deviation for the target-free control. (f) Reactive quantification of RNA characterization. When treated with a series of samples containing a known ratio of input RNA, EZ-READ accurately measured the input RNA characterization, while RT-qPCR showed a poorer correlation. All assays were performed three times (n=3 independent experiments), and the data are presented as mean ± standard deviation in Figures b and df.
[0062] Figure 9The fractal branched microfluidic design and characterization are shown. (a) Two-dimensional fluid simulation and particle tracking analysis of the fractal branched design (left) and the conventional sequential layout (right). The fractal branched design exhibits more consistent flow rates and uniform particle compartmentalization. (b) Optimization of the branch angle. Different branch angles (α = 15°, 60°, and 90°) yield different micropore densities. Theoretical calculations show that α = 90° achieves a high pattern density. (c) Schematic diagram of the optimized microfluidic. The chip has six layers of fractal branches, in which the main control channel is repeatedly divided and subdivided into thousands of branches. Each branch ends in four micropores. Therefore, this design achieves consistent fluid length and resistance to individual micropores, which is beneficial for uniform micropore loading and enables scalable expansion. (d) Large-scale device image. Magnified views of different spatial regions demonstrate consistent reagent loading and uniform distribution in each micropore. The experiment was independently repeated three times with similar results.
[0063] Figure 10 Liquid occupancy characteristics are shown. (a) Schematic diagram of the imaging device. The PDMS chamber is covered with an impermeable optical adhesive membrane to reduce liquid evaporation. (b) Experimental characterization. The inventors monitored the liquid occupancy in the micropores. High liquid occupancy could be maintained for at least several hours, which is significantly longer than the detection duration (approximately 30 min). (c) Sequential solution loading. The microfluidic platform supports simultaneous micropore filling and effective air sealing of each reaction chamber. The experiment was independently repeated three times with similar results. (d) Quantification of reactive nanoparticles. Different numbers of HRP@ZIF-8 nanoparticles were introduced into the microfluidic chip. After loading a chemiluminescent substrate and effective air sealing, endpoint fluorescence images were recorded. The experiment was independently repeated three times with similar results. All measurements were performed three times (n=3 independent experiments), and the data in Figure b are presented as mean ± standard deviation.
[0064] Figure 11 The quantification of HRP@ZIF-8 nanoparticles is shown. The limit of detection (LOD, dashed line) was determined to be approximately 10 particles, defined by 3 × standard deviation of a matched control (no particle control). The maximum limit of detection was estimated to be approximately 4000 particles based on the dynamically distinguishable signal range. All assays were performed triplicate (n=3 independent experiments), and data are presented as mean ± standard deviation. Source data are provided in the source data file.
[0065] Figure 12EZ-READ performance evaluation is shown. (a) Threshold determination for positive wells. The fluorescence distribution of the microwells showed two distinguishable peaks. Positive wells were determined by setting a threshold to eliminate background caused by substrate autofluorescence. (b) Correlation between positive microwells and target RNA molecules. The number of fluorescent wells showed a good correlation with target abundance. (c) Calibration curve for absolute quantification of RNA levels. Different amounts of Arabidopsis thaliana ath-miR159a, a target not present in human samples, were added to human plasma. Spiked plasma was incubated with a sensor targeting ath-miR159a, and the number of fluorescent wells was determined to establish a calibration curve for target quantification. (d) Sensitivity of RT-qPCR for miRNA detection. The limit of detection (LOD) was determined by titrating a known amount of miR-222-3p target. miRNA samples were reverse transcribed and qPCR was performed directly (without pre-amplification) or pre-amplified and then followed by PCR amplification (pre-amplification before qPCR analysis). LOD was defined as 3 × standard deviation of the target-free control. The n=3 independent experiments are shown in Figure bd. All measurements were performed three times (n=3 independent experiments), and the data are presented in Figure bd as mean ± standard deviation.
[0066] Figure 13The analytical evaluation of EX-READ is shown. (a) EZ-READ quantifies RNA targets in lysed biological samples. GLI36vIII cell lysates were prepared by chemical lysis and sonication, respectively, without RNA purification. Using a designed target sensor, EZ-READ detection was able to detect endogenous miR-222-3p in these samples prepared by chemical lysis and sonication. EZ-READ produced a signal comparable to that measured in RNA extracted with the gold standard. The control sensor for Arabidopsis ath-miR159a produced a negligible signal. Measurements were performed three times (n=3 independent experiments), and data are presented as mean ± standard deviation. (b) EZ-READ analytical validation. Measurements of the same samples were performed on different EZ-READ chips by different users. The technique showed robust performance with a small coefficient of variation (within-user = 2.04%, and between-user = 3.83%). n=6 technical replicates. Data are presented as mean ± standard deviation. Multiple comparisons were adjusted using Bonferroni correction for a two-tailed Student's t-test. (c) EZ-READ analysis of IDH1 mutation status. Tumor tissue and paired plasma samples (n=10 patients, 7 of whom were IDH1 wild-type and 3 of whom were IDH1 R132H mutated) were analyzed using the EZ-READ platform. Patient classification was performed using the gold standard of immunohistochemistry and sequencing of tumor tissue. EZ-READ analysis of tumor tissue and plasma achieved overall accuracy of 100% and 90%, respectively. The assays were performed three times (n=3 independent experiments).
[0067] Figure 14 Molecular characterization of EVs is shown. (a) Transmission electron micrograph of EVs from GLI36vIII cells. Scale bar: 100 nm. (b) Single-peak size distributions of EVs from cell lines (left) and plasma (right), determined by nanoparticle tracking analysis, show similar average diameters (approximately 130 nm). 6.69 × 10⁻⁶ EVs were measured. 8 The sum is 3.22 × 10⁻⁶ 9 Total vesicles were analyzed using bar graphs for cellular and plasma-derived EVs. (c) Western blot analysis of GLI36vIII EV lysates. EVs were lysed and immunoblotted for enriched biomarkers, particularly exosome biomarkers. (d) Bioanalyzer electrophoresis images show that RNA extracted from GLI36vIII cells (left) and EVs (right) exhibited good structural integrity and mass spectra. RNA from EVs showed a higher miRNA:mRNA ratio than cellular RNA. All experiments were independently repeated three times with similar results.
[0068] Figure 15Molecular profiles and biomarker selection for GBM samples are shown. (a) RNA analysis of GBM EVs and their parental cell lines. The inventors obtained cell pellets from six GBM lines and isolated EVs. EVs showed good correlation with cellular mRNA and miRNA levels, indicating that EVs reflect their parental cells and have the potential to be used as alternative biomarkers for cells. (b) RNA analysis of EVs and tumor tissues derived from matched plasma obtained from GBM patients. Plasma and tumor miRNA levels correlated, while their mRNA levels did not. mRNA and miRNA expression levels were determined by RT-qPCR. mRNA levels were normalized relative to the internal control GAPDH, and miRNA levels were normalized relative to total RNA input. (c) Selection of biomarkers to be included in subsequent clinical studies. mRNA biomarkers with plasma expression and all miRNA biomarkers were included in the clinical studies. All assays were performed three times (n=3 independent experiments), and data are presented as mean values.
[0069] Figure 16 The selection of mRNA and miRNA biomarkers is shown. (a) Quantification of mRNA and miRNA in cell lines and derived EVs. Cell lines and derived EVs showed good expression correlations for the measured mRNA and miRNA biomarkers (p = 0.800 and 0.701, respectively, with two-sided Spearman). (b) Quantification of mRNA and miRNA in tumor tissue and EVs derived from matched plasma. Poor correlations were observed between mRNA and miRNA biomarkers in tumor tissue and matched plasma (p = 0.009 and 0.584, respectively, with two-sided Spearman). (c) Comparison of biomarkers eliminated in diseased and control plasma samples. No significant differences were observed in biomarkers of these mRNA targets between diseased plasma samples (n=5) and control plasma samples (n=5). All assays were performed three times (n=3 independent experiments). mRNA levels were normalized relative to the internal control GAPDH expression, and miRNA levels were normalized relative to total RNA amount. Data are presented as mean ± standard deviation (c) in Figure c. ns = not significant, two-tailed Student's t-test.
[0070] Figure 17The validation of biomarkers in clinical tumor tissues is shown. (a) Representative images of classic, mesenchymal, and preneural transcriptomic subtypes of GBM tumors show similar morphology. Scale bar: 100 μm. The experiment was independently repeated three times with similar results. (b) Weighted total score of glioma intrinsic (GI) classification by RNA sequencing analysis. Single-sample gene set enrichment analysis (ssGSEA) was performed using GI features of GBM tumor samples to assign tumor subtypes. (c) Hierarchical clustering of 22 GBM tumors, where the expression levels of cross-validation threshold-filtered genes and EZ-READ genes showed three distinct clusters corresponding to GI subtypes.
[0071] Figure 18 The clinical sample workflow is illustrated. First, clinical samples are collected and processed to extract plasma. The plasma samples are then lysed and subjected to EZ-READ assays.
[0072] Figure 19 Characterization of clinical plasma samples is shown. Neither (a) mean diameter nor (b) plasma EV concentration showed any statistical significance between different clinical groups (n = 60 subjects; training cohort included 12 patients and 8 controls, validation cohort included 22 patients and 18 controls). Clinical information is shown in Table 3. All measurements were performed three times, and data are presented as mean ± standard deviation. ns = not significant, two-tailed Student's t-test.
[0073] Figure 20Clinical validation of EZ-READ is shown. (a) EZ-READ analysis of mRNA and miRNA biomarkers in the training clinical cohort. The training set consisted of n=20 subjects: 12 disease patients (6 mesenchymal and 6 preneural / classical) and 8 controls. Measurements were performed three times (n=3 independent experiments). (b) Multilevel decision model for stratifying plasma samples. Disease indicators were first obtained by multiple linear regression to classify samples into disease or control groups. For samples classified as disease, mesenchymal indicators were generated using multiple linear regression to distinguish tumor subtypes from other subtypes. (c) Disease classification in the training cohort. The selected three-biomarker model outperformed the component-based two-biomarker model when classifying disease and control samples. n=20 subjects. Measurements were performed three times (n=3 independent experiments), and data are presented as mean ± standard deviation. ns = not significant, *P<0.05, **P<0.01, and ***P<0.001, two-tailed Student's t-test. For LOX & SLCO3A1 & miR-17-5p, P = 0.0010. (d) Mesenchymal classification of the training cohort. The selected three-marker model was superior to the component two-marker model when classifying mesenchymal samples. n = 12 subjects. Measurements were performed three times (n = 3 independent experiments), and data are presented as mean ± standard deviation. ns = not significant, *P<0.05, **P<0.01, and ***P<0.001, two-tailed Student's t-test. For miR-21-5p & miR-27a-3p & ALDH1A3, P = 0.0071. (ef) Receiver operating characteristic (ROC) curves for the disease group regression model (e) and the mesenchymal regression model (f) in the training, validation, and combined patient cohorts, respectively. The independent validation cohort consisted of n=40 subjects; 22 patients (8 mesenchymal and 14 preneurotic / classical) and 18 controls. All measurements were performed three times, and data are presented as mean ± standard deviation in Figure cd. ns = not significant, *P<0.05, **P<0.005, and ***P<0.0005, Student's t-test.
[0074] Figure 21The multi-step decision model is shown. (a) Performance of multi-step linear regression with single biomarkers. Receiver operating characteristic (ROC) curves for single biomarkers for the disease group (left) and mesenchymal type (right) in the training set. Note that the area under the curve (AUC) for the single biomarker is lower than that for the composite biomarker group. (b) Comparison with single-step multivariate regression. Using six derived biomarkers (miR-17-5p, miR-21-5p, miR-27a-3p, LOX, ALDH1A3, and SLCO3A1), the multi-step EZ-READ model accurately classified the validation clinical samples into control, preneurotic / classical, and mesenchymal types (78% accuracy). The single-step multivariate regression model failed to accurately classify the validation samples into control, preneurotic / classical, and mesenchymal types (48% accuracy). (c) RT-qPCR quantification and model classification for the disease group (left) and mesenchymal type (right). Multiple linear regression analysis using biomarkers determined by PCR failed to classify samples from the disease and control groups, and also failed to distinguish mesenchymal stem cell (MSC) samples from other samples. In Figure c, n=20 subjects were in the disease classification and n=12 subjects were in the MSC classification. All measurements were performed three times, and data are presented in Figure c as mean ± standard deviation. ns= not significant, two-tailed Student's t-test. For the disease classification, P=0.1170. For the MSC classification, P=0.0518. Detailed Implementation
[0075] For convenience, the references mentioned in this specification are listed in the form of a reference list, which is appended after the embodiments. The entire contents of these references are incorporated herein by reference.
[0076] Definitions
[0077] For convenience, certain terms used in the specification, embodiments and appended claims are compiled herein.
[0078] As used herein, the term "comprising" or "including" should be understood to specify the presence of the feature, whole, step, or component mentioned therein, but does not preclude the presence or addition of one or more other features, wholes, steps, components, or combinations thereof. However, in the context of this disclosure, the term "comprising" or "including" also covers "consisting of". Variations of the word "comprising" such as "comprise" and "comprises" and variations of "including" such as "include" and "includes" have correspondingly varied meanings.
[0079] The terms “nucleic acid” and “nucleotide” are intended to be consistent with their use in the art and include naturally occurring species or their functional analogs. Particularly useful nucleic acid functional analogs are capable of hybridizing with nucleic acids in a sequence-specific manner (e.g., hybridizing with two nucleic acids such that a link can occur between the two hybridized nucleic acids) or can be used as templates for replicating specific nucleotide sequences. Naturally occurring nucleic acids typically have a backbone containing phosphodiester bonds. Analog structures may have alternative backbone linkages, including any of the various backbone linkages known in the art. Naturally occurring nucleic acids typically contain deoxyribose (e.g., found in deoxyribonucleic acid (DNA)) or ribose (e.g., found in ribonucleic acid (RNA)).
[0080] Nucleic acids may contain nucleotides having various analogs of these sugar moieties known in the art. Nucleic acids may include natural or non-natural nucleotides. In this regard, natural deoxyribonucleic acid may have one or more bases selected from the group consisting of adenine (A), thymine (T), cytosine (C), or guanine (G), and ribonucleic acid may have one or more bases selected from the group consisting of uracil (U), adenine (A), cytosine (C), or guanine (G). Useful non-natural bases that may be included in nucleic acids or nucleotides are known in the art.
[0081] The terms "oligonucleotide" and "polynucleotide" are used interchangeably to refer to single-stranded polymers of nucleotides ranging from about 2 nucleotides to about 500 nucleotides in length. Oligonucleotides can be synthetic, prepared by enzymatic reactions (e.g., by polymerization), or prepared using a "split-mix" method. Oligonucleotides can include ribonucleotide monomers (i.e., oligoribonucleotides) and / or deoxyribonucleotide monomers (i.e., oligodeoxyribonucleotides). In some embodiments, oligonucleotides can include combinations of deoxyribonucleotide monomers and ribonucleotide monomers (e.g., random or ordered combinations of deoxyribonucleotide monomers and ribonucleotide monomers). For example, the length of an oligonucleotide can be 4 to 10, 10 to 20, 21 to 30, 31 to 40, 41 to 50, 51 to 60, 61 to 70, 71 to 80, 80 to 100, 100 to 150, 150 to 200, 200 to 250, 250 to 300, 300 to 350, 350 to 400, or 400 to 500 nucleotides. An oligonucleotide may include one or more functional parts linked (e.g., covalently or non-covalently) to a multimeric structure. For example, an oligonucleotide may include one or more detectable tags (e.g., radioisotopes or fluorophores).
[0082] When used to refer to nucleic acids or nucleic acid sequences, "probe" or "target" is intended as a semantic identifier of the nucleic acid or sequence in the context of a method or composition and does not limit the structure or function of the nucleic acid or sequence unless explicitly stated otherwise. In this invention, a probe is a single-stranded DNA containing a region complementary to a target RNA sequence to form a double strand with said target RNA. The target RNA is a selected RNA that, if present in a sample, can provide diagnostic and / or prognostic information.
[0083] As used herein, the terms "catalytic nanoparticle complex" or "hybrid nanoparticle complex" refer to one or more proteases encapsulated within a metal-organic framework. The enzyme may be selected from, for example, the group comprising: horseradish peroxidase (HRP), acetylcholinesterase, dehydratase, catalase, cellulase, cytochrome c, dehydrogenase, epoxide hydrolase, glucosylamylase, glucose dehydrogenase, glucose oxidase, ferric heme, hydrolase, invertase, laccase, lipase, microperoxidase-11, myoglobin, organophosphate dehydratase, pectinase, protease, soybean epoxide hydrolase, trypsin, tyrosinase, urease, xylanase, α-amylase, β-galactosidase, β-glucosidase. The metal-organic framework may be selected from, for example, the group comprising: ZIF-8, ZIF-67, BDC, and BTC.
[0084] The invention has now been generally described, and will be more readily understood by referring to the following embodiments, which are provided by way of illustration only and are not intended to limit the invention.
[0085] Example
[0086] Standard molecular biology techniques known in the field but not specifically described are generally performed as described in Sambrook and Russel, Molecular Cloning: A Laboratory Manual, Cold Springs Harbor Laboratory, New York (2001).
[0087] Example 1
[0088] Materials and methods
[0089] Synthesis and characterization of HRP@ZIF-8 nanoparticle composites
[0090] 2-Methylimidazole (2.5 M, Sigma Aldrich) was prepared in water and mixed with streptavidin-HRP (5 μg / ml, Thermo Scientific), polyacrylic acid (0.2 M, Sigma Aldrich), and zinc nitrate hexahydrate (0.5 M, Sigma Aldrich) at 25 °C under stirring. After incubation for 30 min, the suspension was centrifuged at 500 g and washed with phosphate-buffered saline (PBS). The hydrodynamic diameter and zeta potential of the synthesized hybrid HRP@ZIF-8 nanoparticles were characterized by dynamic light scattering analysis (Zetasizer Nano ZSP, Malvern). Three × 14 assay runs were performed at room temperature. The Z-mean diameter, zeta potential, and polydispersity were analyzed. For each assay, the autocorrelation function and polydispersity index were monitored to ensure sample quality for size and surface charge determinations. Powder X-ray diffraction was performed on a Bruker D8 Advanced X-ray diffractometer with a 2θ range of 5° to 50° and a scan rate of 2 ° / min. Cu-Kα X-ray radiation was applied at 40 kV and 40 mA. Fourier transform infrared spectroscopy (Bruker) was performed on an attenuated total reflection unit with a range of 500 cm⁻¹. -1 Up to 3500 cm -1 .
[0091] Enzyme characterization
[0092] To characterize the enzyme conformational change, the inventors measured the fluorescence emission spectrum of tryptophan excited at 280 nm and curve-fitted the peak (Tecan, SparkControl v2.1). A typical redshift in the tryptophan emission spectrum indicates a positive HRP structural change, which helps expose the HRP heme group and increases its substrate affinity (see Duy, C. & Fitter, J. (2006)). Enzyme kinetics were studied using a commercially available kit (QuantaBlu fluorescent peroxidase substrate kit, ThermoScientific). Briefly, working solutions were prepared by mixing the QuantaBlu substrate solution with different concentrations of peroxidase solution at a fixed volume ratio of 9:1. HRP@ZIF-8 was then mixed with the different working solutions. For each reaction, fluorescence intensity (excitation / emission: 325 nm / 420 nm) was monitored for 20 minutes, and time fitting was performed to determine the initial reaction rate (time = 0 s). Finally, the Michaelis-Menten equation was applied to determine K. M and K cat .
[0093] Assembly of RNA-responsive sensors
[0094] To assemble the RNA-responsive sensor, excess NHS / EDC was dissolved in MES buffer (pH 4.7) to form a mixture in which carbodiimide-crosslinked activated carboxyl magnetic beads (3 μm, Spherotech) were placed for 15 min. After rinsing, the activated magnetic beads were incubated with heterobifunctional polyethylene glycol (PEG) (0.1 mM, JenKem) with terminal amino and maleimide groups. Additionally, a DNA probe with 5' thiol and 3' amino groups (30 μM, Integrated DNA Technologies) was activated with reducing tris(2-carboxyethyl)phosphine and recovered by filtration. The activated DNA probe was then added to the prepared maleimide-functionalized magnetic beads. The reaction mixture was washed in excess PBS and then incubated with another heterobifunctionalized PEG (1 mM, Thermo Scientific) with terminal NHS and biotin groups. Finally, DNA-functionalized magnetic beads were added to excess HRP@ZIF-8 nanoparticles, washed in PBS, and stored at 4°C for later use.
[0095] Regenerated signal sensing
[0096] To evaluate the assembled RNA-responsive sensor, the inventors prepared FITC-labeled HRP@ZIF-8 nanoparticles and functionalized them onto magnetic beads. The assembled sensor, along with the RNA target, was incubated in a reaction buffer (50 mM Tris-HCl, pH 8.0, 5 mM MgCl2, 1 mM DTT, 6 U / ml RNase inhibitor) containing 0.5 U of double-stranded nuclease-specific nuclease (DSN, Evrogen). After a magnetic pulldown experiment, the reaction supernatant containing the released HRP@ZIF-8 nanoparticles was collected and analyzed by fluorescence analysis. To investigate the kinetics of HRP@ZIF-8 release, the inventors monitored changes in fluorescence intensity of the supernatant over time and analyzed the signal differences between reactants incubated with high (10 pmol) and low (3 pmol) RNA target amounts. To evaluate the sensor's versatility and specificity, the sensor was mixed with RNA targets of different lengths (20 nt to 200 nt) or different numbers of mismatches.
[0097] Verification of DSN cutting mechanism
[0098] To elucidate the mechanism of DSN cleavage, the inventors prepared RNA-DNA heteroduplexes by mixing equimolar amounts of RNA target and DNA probe in a reaction buffer (50 mM Tris-HCl, pH 8.0, 5 mM MgCl2, 1 mM DTT), and incubated the mixture at 80°C for 15 min, followed by cooling to 4°C. DSN cleavage was then performed by incubating the annealed RNA-DNA heteroduplexes at 60°C for 7 min, followed by cooling to 4°C. For comparison, the inventors performed BstNI digestion 49 (cleavage site CC / WGG, New England Biolabs) on the annealed RNA-DNA heteroduplexes after incubation at 60°C for 1 h, followed by cooling to 4°C. DNA samples before and after digestion (Qiaquick, Qiagen) were extracted and analyzed by non-denaturing polyacrylamide gel electrophoresis (PAGE). Dilute the DNA sample with an appropriate amount of 6X loading dye (Promega) and perform electrophoresis at 150 V in a 10% polyacrylamide gel containing TAE buffer (Vivantis). Stain the gel with 1X SYBR Gold (Invitrogen) TAE buffer for 30 min and then image it using an iBright FL1500 imaging system (ThermoScientific).
[0099] To evaluate the DSN efficiency of the EZ-READ platform, the inventors prepared free-floating RNA-DNA heteroduplexes using FRET-based DNA probes as described above. The inventors treated this gold standard mixture with DSN and determined the amount of fluorescent probe cleaved to determine the maximum amplification efficiency (100%). For comparison, to characterize the EZ-READ platform, RNA-DNA heteroduplexes were immobilized with magnetic beads of equal concentration for DSN cleavage. The EZ-READ amplification efficiency was normalized to the free-floating gold standard and measured to be 93.4%. The decrease in efficiency may be due to reduced enzyme accessibility caused by probe immobilization on the beads.
[0100] Design of digital microfluidics
[0101] To develop a microfluidic platform for the digital characterization of the released HRP@ZIF-8 catalytic reactor, the inventors first performed computational fluid dynamics analysis (COMSOL Multiphysics) to establish flow profiles and optimize the channel design. During the simulation, the fluid material within the channel was set to water, and an incompressible flow model was applied. The numerical simulation considered no-slip wall conditions and zero pressure at the outlet. The inlet velocity was 0.1 m / s, and a laminar flow module with steady-state studies was used. The ambient temperature was set to T = 293.15 K. Furthermore, a particle tracking module considering particle resistance was used to simulate particle trajectories within the micropores.
[0102] Manufacturing and fabrication of digital chips
[0103] Standard soft lithography is used to fabricate microfluidic chips. The inventors used SU-8 negative photoresist (SU8-50 and SU8-100, Microchem) to prepare the mold. In short, to form the connecting fluid channels, SU8-50 photoresist was spin-coated onto a Si wafer at 3000 rpm for 30 s, and baked at 65°C and 95°C for 5 min and 15 min, respectively. After UV line exposure, the photoresist was baked again before development. To fabricate the second layer containing digital microvias, SU8-100 photoresist was spin-coated onto a Si wafer at 1000 rpm for 30 s, degassed for 30 min, and baked at 65°C and 95°C for 30 min and 90 min, respectively. Before exposure and photoresist development, the photomask was precisely aligned with the markings of the first layer. The fabricated fluid channels and digital vias have heights of 40 μm and 250 μm, respectively. The developed mold was chemically treated with trichlorosilane vapor in a desiccator for 15 min and baked at 150°C for 10 min to evaporate excess silane before subsequent use. Polydimethylsiloxane polymer (PDMS) and a crosslinking agent were mixed at a 10:1 ratio and cast onto an SU-8 mold. After curing at 65°C for 4 h, the PDMS layer was cut from the mold and plasma-treated before being assembled onto a glass slide. The inlet was fabricated using a 4 mm biopsy puncture device for sample loading. Post-manufacturing, to achieve automated sample loading, the inventors first vacuum-loaded the manufactured digital chip. The chip was sealed using MicroAmp optical adhesive film (Applied Biosystems) and incubated at 10 mbar for 2 h before sample loading.
[0104] EZ-READ testing workflow
[0105] The inventors employed the prepared EZ-READ microfluidic platform to achieve catalytic digital quantification of RNA targets. Specifically, 3 μl of a sample containing the RNA target was first mixed with 8 μl of an RNA-responsive sensor to trigger the release of HRP@ZIF-8 nanoparticles. After incubation at 60°C for 7 min, 5 μl of HRP@ZIF-8 supernatant was added to a vacuum-loaded microfluidic chip. At an average flow rate of 10 μl / min, the released HRP@ZIF-8 particles were automatically and uniformly loaded into each microwell within 2.5 min. To catalytically quantify individual HRP@ZIF-8 nanoparticles, the inventors introduced a commercially available kit (QuantaRed enhanced chemiluminescence HRP substrate, Thermo Scientific), which generates an HRP-dependent fluorescence signal by producing a halogenated molecule (a soluble fluorescent product with an excitation / emission maximum of 570 / 585 nm). In short, 20 μl of the QuantaRed substrate was introduced into the microfluidic chip within 5.5 min. Subsequently, air was introduced into the microfluidic channels to seal each micropore, thereby enabling independent catalytic reactions and reducing cross-contamination between micropores. The catalytic process was allowed to continue for 1 min prior to fluorescence data acquisition. All experiments were performed with a sample-matched control, where each sample was incubated with a sensor targeting Arabidopsis thaliana ath-miR159a, a target not present in human samples.
[0106] EZ-READ signal processing
[0107] The inventors captured fluorescence and bright-field images of the EZ-READ micropores using an inverted fluorescence microscope (DMi8 Leica). Image acquisition was performed using LAS X software (v.3.6.123246), and image analysis was performed using ImageJ software (v.1.53k). The final image was composed of 81 individual images, each acquired at 5x zoom and 16-bit depth to achieve a total pixel count of 16377 × 16377, with a final area of 21.28 × 21.28 mm. 2 The original images were processed in grayscale, and the fluorescence signal was pseudo-color red. Image contrast and brightness were adjusted and consistently applied to both the control and experimental groups. To eliminate background caused by substrate autofluorescence, a positive / negative decision threshold was determined by analyzing the fluorescence intensity distribution of known positive and negative wells. This threshold was consistently applied to all images in both the sample and control groups. The final signal for each sample was determined by subtracting the number of positive wells measured in its sample-matched control.
[0108] Cell culture
[0109] Human glioblastoma (GBM) cell lines GLI36vIII and SKMG3 were provided by Dr. Timothy Chan of Memorial Sloan-Kettering Cancer Center and grown in Dulbecco modified essential medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS). Human glioma proliferating cell lines NNI-11, NNI-22, NNI-32, and NNI-24 were generated from primary GBM samples provided by the National Institute of Neuroscience (NNI) and cultured into neurospheres in a serum-free 3:1 mixture of Dulbecco-modified Eagle medium (DMEM, Sigma-Aldrich) and Ham's F-12 nutrient mixture (F12, Gibco) (see Chong, Y. K et al., (2009)). This medium was supplemented with basal fibroblast growth factor (bFGF, Peprotech Inc., 20 ng / ml), epidermal growth factor (EGF, Peprotech Inc., 20 ng / ml), heparin (Sigma-Aldrich, 5 μg / ml), and a serum-free supplement (B27, 1x, Gibco). All media were supplemented with penicillin-streptomycin. Cultures were incubated at 37°C in a water-saturated environment containing 5% CO2. All cell lines were tested and free of mycoplasma contamination (MycoAlert Mycoplasma Detection Kit, Lonza, LT07-418).
[0110] Isolation and quantification of extracellular vesicles
[0111] Prior to vesicle collection, cells were cultured for 48 h in vesicle-depleted medium (containing 5% depleted FBS). All vesicle-containing media were filtered through a 0.2 μm membrane filter (regenerated cellulose, Millipore), separated by differential centrifugation (first at 10,000 g, then at 100,000 g), and used for molecular analysis. To independently quantify vesicle concentration, the inventors used a nanoparticle tracking analysis (NTA) system (NS300, Nanosight). Vesicle concentration and diameter were adjusted to obtain approximately 50 vesicles per field of view for optimal counting. To maintain consistency, all NTA assays were performed using the same system settings.
[0112] Clinical samples
[0113] Delabeled GBM tumor specimens were provided by the National Neuroscience Institute Tissue Bank, Singapore (application number SBRSA2019 / 002) with informed consent. Fresh tumor tissue was obtained during surgical resection and processed immediately. Under aseptic conditions, the tumor tissue was subdivided into small pieces and subjected to RNA sequencing and immunohistochemical staining separately. Specifically, for RNA analysis, tumor pieces were rapidly frozen in liquid nitrogen-immersed cryovials and stored at -150°C. For immunohistochemical studies, tumor pieces were fixed in 4% paraformaldehyde at 4°C, dehydrated, and embedded in paraffin, or incubated in 30% sucrose at 4°C and then rapidly frozen in a plastic mold in an embedding medium at the optimal cutting temperature on dry ice. The tissue pieces were stored at -80°C for subsequent analysis. For plasma samples, venous blood was drawn into EDTA tubes prior to tumor resection and processed immediately. Blood samples were centrifuged at 400 g (4°C) for 10 min. The recovered plasma was centrifuged again at 1100g (4°C) for 10 min. Plasma samples were stored at -80°C before use. All EZ-READ assays were performed without knowing the clinical diagnosis.
[0114] Direct EZ-READ detection in biolysates
[0115] To evaluate the performance of EZ-READ in lysed biological samples, the inventors prepared GLI36vIII lysates using various lysis protocols (see Sundah, NR et al. (2021), incorporated herein by reference) and applied EZ-READ for direct assays without RNA extraction. For chemical lysis, the inventors prepared lysis buffers containing varying amounts of Triton X-100 and sodium dodecyl sulfate (SDS) and incubated these buffers with the cell pellet at room temperature. For sonication, the cell pellet was resuspended in PBS and sonicated at 80°C for 5 min (Elma). For plasma lysis, plasma samples were incubated with Triton X-100 and SDS for 5 min to a final concentration of 1% Triton X-100 and 0.1% SDS. Using the above assay procedure, these lysates were directly measured using EZ-READ.
[0116] RNA extraction and quantification
[0117] As the gold standard, biological samples were harvested for RNA extraction using a commercially available kit (miRNeasy, Qiagen) in TRIzol reagents. RNA was extracted from the lysates according to the manufacturer's protocol and treated with DNase I (Qiagen) to remove DNA contamination. The extracted RNA was quantified using a NanoDrop spectrophotometer (Thermo Scientific). To assess the size distribution of the extracted RNA, the purified RNA samples were evaluated using an RNA Pico Chip with a 2100 Bioanalyzer (Agilent).
[0118] RT-qPCR analysis
[0119] For RT-qPCR, extracted mRNA and miRNA were first reverse transcribed to generate first-strand cDNA (High-Capacity cDNA Reverse Transcription Kit and TaqMan MicroRNA Reverse Transcription Kit, Applied Biosystems). The cDNA was pre-amplified as needed (TaqMan PreAmp Master Mix, Applied Biosystems) before qPCR. All qPCR reactions were performed using the QuantStudio 5 Real-Time PCR System (Applied Biosystems) following the manufacturer's instructions using the TaqMan FastAdvanced Master Mix (Applied Biosystems). For mRNA, TaqMan Gene Expression Detection (Applied Biosystems) was used. For miRNA, TaqMan MicroRNA Detection (Applied Biosystems) was used. Amplification was performed with one cycle at 95°C for 10 min, followed by 50 cycles of 95°C for 15 s and 60°C for 60 s. All experiments were performed three times. For mRNA, relative quantification was performed by normalization relative to the internal control GAPDH expression. For miRNAs, relative quantification is achieved by normalizing them relative to the total RNA amount.
[0120] Intrinsic classification of gliomas in GBM patients
[0121] To determine molecular subtypes and validate the selected RNA biomarkers, the inventors performed RNA sequencing analysis on patient tissue samples. Briefly, 1 μg of total RNA was extracted from rapidly frozen patient tumor tissue for RNA-seq library preparation, using the Illumina TruSeq Stranded Total RNA Library PrepGold kit according to the manufacturer's instructions. Library fragment sizes were determined using a DNA 1000 Bioanalyzer (Agilent), and the libraries were quantified by qPCR using the KAPA Library Quantification Kit (KAPA Biosystems). The libraries were pooled at equimolar concentrations and clustered in an Illumina cBot system. Sequencing (150 bp paired ends) was performed using an Illumina HiSeq 3000 system according to the manufacturer's protocol. RNA sequencing reads were aligned to the GRCh38 human reference genome using STAR v2.7.9a (see Dobin, A. et al. (2013)). Gene expression levels, measured in kilobase million transcripts (TPM), were then calculated using RSEM v1.3.3 (see Li, B. & Dewey, CN (2011)). The TPM values were further log2 transformed for subsequent molecular classification. Using the intrinsic glioma (GI) features described in Wang et al. (2017), molecular subtypes of GBM tumors were queried and identified through single-sample gene set enrichment analysis (ssGSEA).
[0122] Characteristic testing of combined gene traits
[0123] To evaluate the ability of the selected mRNA targets to classify GI subtypes, 22 EZ-READ genes sharing a common gene set of LOX, TGFBI, PLAUR, COL1A2, and CYP1B1 were combined with 150 GI characteristic genes for further validation in separate clinical studies. A 5-fold cross-validation analysis was performed on gene expression data determined by tumor RNA sequencing using PAMR package v1.56.1 software (see Hastie, T. et al. (2019)) to calculate the threshold for minimizing misclassification error. A threshold of 0.7 was determined, and genes that did not contribute to inter-subjectivity variability were removed, retaining 144 out of 167 genes that correctly predicted the GI subtype of the samples. All EZ-READ targets except ARC contributed to inter-subjectivity variability. The 144 cross-validation thresholds were used to filter genes for hierarchical clustering of classical (n=6), mesenchymal (n=10), and preneural (n=6) samples.
[0124] Development and validation of EZ-READ plasma genetic signatures
[0125] To identify plasma genetic signatures for GBM diagnosis and subtyping, the inventors used the EZ-READ platform to determine mRNA and miRNA expression levels in clinical plasma samples and developed a multistep linear regression scoring model to classify disease states (affected = 1, control = 0) and mesenchymal subtype states (mesenchymal type = 1, anterior neural / classical type = 0). For both classifications, the models were developed using a training patient cohort, where biomarkers were selected based on their classification performance (i.e., the area under the curve (AUC) in receiver operating characteristic (ROC) analysis). For comparison, the inventors also developed a single-step multinomial regression model using the same set of biomarkers identified through the multistep model. The inventors further evaluated the performance of both models (multistep and single-step regression models) in an independent validation patient cohort and demonstrated that the multistep model outperformed the single-step model.
[0126] The gold standard classification of a patient's tumor IDH1 mutation status
[0127] To determine the IDH1 mutation status, the inventors performed immunohistochemical (IHC) staining and independently verified it by tissue DNA sequencing. IDH1 mutation status was defined based on IDH1 R132H immunopositivity and IDH1 gene mutation. For IHC, paraffin-embedded tumor sections were stained with the monoclonal antibody IDH1 R132H (Master Diagnostica, clone H09, diluted 1:1000), as described previously.54 For sequencing, genomic DNA was extracted from frozen primary tumors using the DNeasy Blood and Tissue Kit (Qiagen), according to the manufacturer's protocol. As described previously,6, after standard PCR amplification, the IDH1 mutation alteration at the hotspot codon R132 was sequenced using the Sanger method. The primers used are as follows:
[0128]
[0129] EZ-READ analysis of IDH1 mutation status
[0130] To evaluate the EZ-READ platform's ability to determine patients' IDH1 mutation status, the inventors analyzed tissue and paired plasma samples from IDH1 wild-type and IDH1 R132H patients. As described above, patient classification was performed using tissue IHC as the gold standard, validated by tissue DNA sequencing. The inventors used the EZ-READ sensor (Table 1) to determine the expression levels of IDH1 wild-type and IDH1 R132H mRNA. Samples identifying wild-type but not R132H mRNA were classified as wild-type; samples identifying both wild-type and R132H mRNA were classified as mutant.
[0131]
[0132] Western blot
[0133] Extracellular vesicles separated by ultracentrifugation were lysed in radioimmunoprecipitation assay (RIPA) buffer containing a protease inhibitor (Thermo Scientific) and quantified using a quinolinic acid assay (BCA assay, Thermo Scientific). Protein lysates were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to polyvinylidene fluoride membranes (PVDF, Invitrogen), and immunoblotted with antibodies against the following protein markers: HSP90 (Cell Signaling, 1:1000 dilution), Flotillin 1 (BD Biosciences, 1:1000 dilution), CD63 (Santa Cruz, 1:250 dilution), ALIX (Cell Signaling, 1:1000 dilution), TSG101 (BD Biosciences, 1:1000 dilution), and LAMP-1 (R&D Systems, 1:1000 dilution). As needed, after incubation with anti-mouse IgG, HRP-linked antibody (Cell Signaling, diluted 1:2000) or anti-rabbit IgG, HRP-linked antibody (Cell Signaling, diluted 1:2000), immunoassay was performed using enhanced chemiluminescence (ThermoScientific).
[0134] Hematoxylin and eosin tissue staining
[0135] 5 μm thick paraffin-embedded tissue sections were dewaxed, rehydrated, and then stained with hematoxylin. After rinsing with tap water, the tissue sections were stained with eosin, dehydrated, and mounted with DPX. Micrographs of the stained tissue sections were obtained using a Nikon Eclipse TE2000-S microscope.
[0136] electron microscope
[0137] For scanning electron microscopy, samples were fixed with half-strength Karnovsky's fixative and washed twice with PBS. After dehydration in a series of increasing ethanol concentrations, samples were transferred for critical drying (Leica), followed by gold sputtering (Leica), and then imaged using a scanning electron microscope (JEOL 6701). For transmission electron microscopy, dried samples were imaged using a transmission electron microscope (JEOL 2200FS).
[0138] Statistical analysis
[0139] Unless otherwise stated, all measurements were performed triplicate, and data are presented as mean ± standard deviation. Two-tailed Student's t-tests were used for significance testing. For between-sample comparisons, multiple pairs were tested, and the resulting p-values were adjusted using Bonferroni correction for multiple hypothesis testing. Adjusted p-values < 0.05 were considered significant. Linear regression was used for correlation analysis to determine goodness of fit (R²) unless otherwise stated. Statistical analysis was performed using R (v.4.0.3) and GraphPad Prism (v.9.0.0) software. For clinical analyses of GBM diagnosis and subtyping, sample size was limited by patient sample availability. No data were excluded from the analysis. Plasma samples were randomly assigned to two representative cohorts. All experiments were performed without prior knowledge of the clinical diagnosis.
[0140] Example 2
[0141] EZ-READ platform
[0142] The EZ-READ platform utilizes an RNA-responsive sensor to achieve two functional steps: regeneration signal sensing and catalytic digital quantification. Figure 1 (a)). The sensor comprises a hybrid nanoparticle complex covalently linked to magnetic beads via a specific DNA probe. Figure 1 (b) In each nanoparticle complex, horseradish peroxidase (HRP) is encapsulated in a metal-organic framework (ZIF-8); this encapsulation (HRP@ZIF-8) not only enhances the enzyme's intrinsic catalytic activity but also protects it from potential inhibition by the complex biological environment. Figure 2 (ab)). The EZ-READ platform employs this powerful catalytic sensor to achieve direct and sensitive RNA quantification. During regeneration signal sensing, the RNA target binds in a target-recoverable manner and triggers enhanced release of HRP@ZIF-8 nanoparticles. Specifically, when RNA hybridizes with the sensor (i.e., via a single-stranded DNA probe), a duplex-specific nuclease (DSN) (see Shagin, DA et al. (2002)) recognizes the RNA-DNA duplex and cleaves only the DNA adapter; this not only releases the bound HRP@ZIF-8 nanoparticles but also regenerates the RNA target, preparing it for the next reaction. Figure 1 (a) and Figure 2 (cd)). To achieve sensitivity measurement of target sensing, the inventors developed a fractal branched microfluidic chip for digital quantification, utilizing the strong catalytic activity of HRP@ZIF-8 nanoparticles. Figure 1(c) HRP@ZIF-8 nanoparticles released from the target are distributed and dispensed into individual microfluidic micropores, where they individually catalyze the enhancement of strong chemiluminescence signals for digital quantification. All fluid-driven processes (i.e., nanoparticle dispensing, substrate introduction, and micropore sealing) are powered by vacuum loading. Figure 2 (e)).
[0143] Compared to conventional RNA detection methods (such as PCR), this method requires multiple processing steps (such as reverse transcription and exponential target amplification) and is susceptible to variable amplification efficiency. Figure 3 EZ-READ utilizes a highly catalytic sensor to achieve direct and reliable RNA detection. Figure 1 (d) Specifically, through regenerative sensing, EZ-READ bypasses all steps of conventional PCR detection and achieves robust assays; through the potent catalysis of released HRP@ZIF-8 nanoparticles, the platform enables digital counting and reliable detection of RNA signatures. Inspired by its direct and reactive detection capabilities, the inventors used the EZ-READ platform to determine different RNA subtypes (e.g., long mRNA and short miRNA) found in clinical blood samples from GBM patients. Figure 1 (e) and further developed a multi-step classification model to accurately diagnose patients and perform molecular subtyping of GBM tumors. Figure 4 ).
[0144] Example 3
[0145] Regenerative RNA Sensing
[0146] To achieve regenerative signal sensing, the inventors first prepared and characterized HRP@ZIF-8 nanoparticles ( Figure 5 (ac)). Compared to the original HRP, HRP@ZIF-8 exhibited enhanced enzyme activity and improved stability in complex environments. Specifically, HRP@ZIF-8 nanoparticles exhibited a characteristic red shift in their fluorescence spectrum (ac). Figure 7 (a) indicates a conformational change in the HRP active pocket (see Duy, C. & Fitter, J. (2006)). Importantly, HRP@ZIF-8 exhibited improved enzyme kinetics when assayed at different substrate concentrations. Figure 7 (b) and Figure 5 (d) When fitted according to the Michaelis-Menten equation, the HRP@ZIF-8 reaction exhibited a lower KM, thus confirming the enhanced affinity of the nanoparticles for the substrate. All comparisons were performed under conditions that matched the enzyme concentration. In addition to exhibiting strong catalytic performance, the HRP@ZIF-8 nanoparticles also protect the embedded enzyme from environmental disturbances ( Figure 7 (c) andFigure 5 (e) enables the regeneration of signal sensing in a one-step reaction (i.e., the merging of DSN and associated buffer). Subsequently, the inventors covalently linked catalytic HRP@ZIF-8 nanoparticles to magnetic beads using specific DNA probes, assembling an EZ-READ RNA-responsive sensor. In fabricating this sensor, the inventors coated the HRP@ZIF-8 nanoparticles with polyacrylic acid (PAA) and polyethylene glycol (PEG) spacers to improve DNA probe binding and RNA-induced nanoparticle release. Figure 5 (fg)). When treated with DSN to mediate the cleavage of HRP@ZIF-8 nanoparticles, the assembled sensor exhibited specific and amplifying responses to RNA targets, but no response to DNA targets. Figure 3 (d) It is worth noting that, compared to sequence-specific DNA cleavage that recognizes specific cleavage sites and produces different DNA products, the DSN system can enhance DNA cleavage and has broad applicability to different target sequences. The DSN system recognizes RNA-DNA heteroduplexes and cleaves only the DNA strand in a random, repetitive manner, thereby producing short DNA products (≤6 nt). Figure 6 (a) and Table 2); therefore, these short DNA products can be readily separated from the RNA target. In fact, in the EZ-READ system, this efficient DNA cleavage not only separates the linked HRP@ZIF-8 nanoparticles to release the signal, but also regenerates the RNA target for subsequent reactions, thereby enabling regeneration signal enhancement ( Figure 6 (b) Inspired by this effective sensing, the inventors then evaluated the detection programmability and specificity of EZ-READ. To evaluate its ability to detect RNA targets of different lengths (e.g., miRNA and mRNA), the inventors designed sensors with DNA probes of different sizes. When incubated with RNA targets of different lengths (20 nt to 200 nt), the EZ-READ technology produced a strong specific signal and had minimal background signal in the confounding controls. Figure 2 (e) When incubated with RNA targets that have mismatches, the system further demonstrates good performance in distinguishing rare mismatches. Figure 7 (f) and Table 1). Based on this, the inventors developed EZ-READ sensors for various GBM-related RNA biomarkers (e.g., mRNA: LOX, SLCO3A1, COL1A2, miRNA: 17-5p, 21-5p, 27a-3p) (see Lottaz, C. et al. (2010); Huang, SW et al. (2018)) and confirmed their detection specificity ( Figure 7 (g) and Table 1).
[0147]
[0148] Example 4
[0149] Catalytic digital EZ-READ for reactive target quantification
[0150] To accurately determine the regeneration sensing (i.e., the release of HRP@ZIF-8 particles), the inventors developed a microfluidic platform capable of digital quantification of catalysis. This platform comprises a fractal branched micropore array to distribute and partition individual HRP@ZIF-8 particles, thereby leveraging their strong catalytic activity for digital chemiluminescence detection. Figure 8 (a) Conventional microfluidic designs fill micropores sequentially, leading to uneven distribution. Figure 9 (a) Compared to conventional microfluidic designs, the fractal branching configuration ensures uniform fluid resistance to each micropore, enabling uniform and controllable filling of the micropores (e.g., using HRP@ZIF-8 particles). The inventors further optimized the fractal branching angle to maximize the micropore density in the chip. Figure 9 (b)). Through six layers of fractal branching, the main control channel is repeatedly divided and subdivided into thousands of branches ( Figure 9 (c)). By connecting each branch terminal to four micropores, the current prototype accommodates a total of 4096 micropores within a 2.5cm × 2.5cm area. Figure 9 (d)). To achieve digital quantification of catalysis, the inventors sequentially loaded HRP@ZIF-8 particles, a chemiluminescent substrate, and air into a microfluidic platform to achieve a stepwise reaction: 1) distributing HRP@ZIF-8 into each micropore, 2) introducing the substrate into each micropore, and 3) sealing each reaction chamber for independent catalysis and digital counting. Vacuum loading powers all fluid drives and minimizes micropore evaporation during detection. Figure 10 (ab)). To optimize these stepwise reactions, the inventors monitored the liquid occupancy in individual micropores to ensure loading efficiency and reaction integrity. Figure 8 (b) Figure 10 (c) Experiments confirmed uniform particle distribution and independent chemiluminescence reactions, which accurately reflected the input particle concentration ( Figure 10 (d)), thereby accurately quantifying the concentration of HRP@ZIF-8 nanoparticles ( Figure 11 ).
[0151] The inventors then evaluated the performance of the integrated EZ-READ platform for direct RNA quantification. Using the developed platform, they performed a series of titration experiments on known amounts of RNA targets, such as mRNA and miRNA. RNA strands were incubated with the EZ-READ sensor to release HRP@ZIF-8 particles (Table 2). The released particles were uniformly distributed within the microfluidic pores, producing a strong and representative chemiluminescent signal. Figure 8 (c) The number of fluorescent microwells (i.e., positive wells) Figure 12 (a) is not only closely related to the occupancy rate of nanoparticles, but also accurately measures the amount of RNA input. Figure 8 (d) When used to determine RNA target copy number, EZ-READ can be used to determine a wide range of RNA target concentrations and establish linear calibration for target quantification. Figure 12 (bc)). Especially for short miRNA detection, although the gold standard RT-qPCR requires numerous preparation steps to extend and detect short miRNAs (i.e., RNA extraction, reverse transcription, and pre-amplification) Figure 12 (d)), but the EZ-READ platform enables direct and sensitive miRNA detection, and establishes detection limits for 9 miRNA copies ( Figure 8 (e) Specifically, the EZ-READ platform bypasses all the preparation steps for RT-qPCR used for short miRNA analysis and enables direct assays in minimally processed sample lysates. Figure 13 (a)). Through analysis and evaluation, the inventors then confirmed the technical robustness when measured by different users on different chips. Figure 13 (b) The inventors ultimately evaluated the accuracy of the EZ-READ platform in detecting input RNA features. Using a sensor targeting clinically relevant mutations (i.e., the IDH1 R132H mutation in brain tumors) (Table 1), the inventors performed EZ-READ assays in patient tumor tissue and paired plasma samples with known IDH1 mutation status and established accurate tumor classification (overall classification accuracy ≥90%). Figure 13 (c) Finally, regarding the determination of complex ratio RNA characteristics, although RT-qPCR showed a poor correlation with the input RNA ratio, possibly due to its extensive sample processing and sensitivity to variable amplification efficiency, the EZ-READ platform directly and accurately reflected the input RNA characteristics. Figure 8 (f)).
[0152] Example 5
[0153] Selection of RNA biomarkers for GBM genotyping
[0154] To clinically apply the EZ-READ platform for GBM subtyping, the inventors first selected RNA biomarkers for GBM diagnosis and subtype classification. Based on previous research (referencing Lottaz, C. et al. (2010); Huang, SW et al. (2018)), the inventors selected 22 mRNA biomarkers and 10 miRNA biomarkers, and evaluated these biomarkers using cell lines and clinical specimens. For cell line characterization, the inventors cultured six primary GBM cell lines (NNI-22, NNI-24, NNI-32, NNI-11, GLI36vIII, and SKMG3) and isolated their vesicles (EVs). Through molecular characterization, the inventors demonstrated that the average diameter of the vesicles was approximately 130 nm and that they expressed characteristic EV protein biomarkers and RNA composition (…). Figure 14 The inventors further determined the RNA content of the cells and the EVs derived from them. Figure 15 (a)). The mRNA and miRNA profiles of EVs and parental cells showed good correlation ( Figure 16 (a) indicates that EVs reflect their parental cells and have the potential to be used as alternative biomarkers for cell use. The inventors further evaluated RNA markers in clinical plasma samples and primary tumor tissues from GBM patients to assess the expression of these markers in plasma and their correlation with tissue subtypes. First, the inventors measured these RNA markers in paired plasma and tumor tissues. Interestingly, clinical plasma samples showed low expression of many mRNA markers ( Figure 15 (b) This is supported by the decrease in mRNA abundance in EVs. Figure 14 (d)), and showed an RNA profile different from that of the primary tissue ( Figure 16 (b) The inventors attribute this to the heterogeneous composition of EVs in plasma; EVs are released into circulation by multiple cell types (see van Niel, G. et al., (2018)). Therefore, the inventors selected mRNA targets expressed in plasma and all miRNA targets for our subsequent clinical studies. Figure 15 (c) and Figure 1 (c)). Next, the inventors independently validated these selected RNA targets using other GBM tumor tissues of different GI subtypes. Figure 17 (ab)). Specifically, by testing biomarkers with established GI characteristic genes including LOX, TGFBI, PLAUR, COL1A2, and CYP1B (see Wang, Q. et al. (2017)), the inventors performed a 5-fold cross-validation analysis and demonstrated that the selected mRNA targets contribute to the variability analysis of GI subtype classification. Figure 17(c)). These targets are not only expressed in plasma, but are also closely associated with different GI subtypes, thus supporting their potential application in blood-based disease typing.
[0155] Example 6
[0156] Clinical Analysis of GBM Diagnosis and Classification
[0157] Using selected RNA markers, the inventors conducted a clinical feasibility study to evaluate the ability of the EZ-READ platform to analyze blood-based RNA in GBM. GBM is a grade 4 glioma and can be classified into three subtypes at the transcriptional level with varying prognoses and treatment responses. In this study, the inventors aimed to determine (1) whether the EZ-READ platform could be applied to clinical plasma samples for diversified analysis of different RNA molecules (e.g., miRNA and mRNA), and (2) the accuracy of EZ-READ RNA signatures in differentiating and subtyping GBM patients. Blood samples were obtained from GBM patients (n=34) and control subjects (n=26) and randomly assigned to two representative cohorts (i.e., the training cohort and the validation cohort, Table 3) to independently assess the robustness of the EZ-READ assay. Figure 18 All GBM subjects were clinically characterized by disease typing through RNA sequencing of primary tumor tissue, and indistinguishable plasma EV characteristics were observed between groups. Figure 19 In a training cohort comprising 12 diseased plasma samples and 8 control plasma samples, the inventors first used the EZ-READ platform to determine the selected RNA biomarkers. Figure 20 (a) ). Using the measured RNA abundance, the inventors constructed a multilevel decision model that can (1) distinguish between diseased samples and control samples, and (2) distinguish between mesenchymal GI subtypes, namely the most aggressive and worst-prognostic GBM subtype ( Figure 20 (b)). In developing this sequential model, the inventors characterized the performance of each marker through receiver operating characteristic (ROC) curve analysis. Figure 21 (a)). To improve classification accuracy, the inventors further performed stepwise linear regression to identify a combination of multiple biomarkers: disease markers distinguishing patients from controls included miR-17-5p, LOX mRNA, and SLCO3A1 mRNA ( Figure 20 (c)), and the mesenchymal subtype markers that distinguish mesenchymal subtypes from other disease samples include miR-21-5p, miR-27a-3p, and ALDH1A3 mRNA ( Figure 20(d) To evaluate the performance of the EZ-READ classification model, the inventors further evaluated the technique in an independent validation cohort comprising 11 diseased patients and 8 control patients. Compared to the training cohort, the developed model metrics demonstrated comparable performance in the validation cohort, achieving accurate classification. Figure 20 (ef), the AUGs of the merged queues are 0.897 and 0.854, respectively. The multi-step decision model outperforms the single-step multinomial regression and exhibits higher classification accuracy. Figure 21 (b)). In all patients tested, the EZ-READ classification model demonstrated better performance than conventional RT-qPCR. Figure 21 (c) This is due to its improved robustness and reliable measurement.
[0158]
[0159] Summarize
[0160] Recent studies have demonstrated that EVs are attractive circulating biomarkers for assessing GBM in blood (see Skog, J. et al. (2008); Shao, H. et al. (2012); Batool, SM et al. (2022)). EVs readily cross the blood-brain barrier and carry abundant nucleic acids (e.g., short miRNAs and long mRNAs). However, current analytical techniques face limitations in reliably measuring these scarce and diverse circulating RNAs, particularly due to the complexity of sample processing and susceptibility to amplification variability. To address these challenges, the inventors have developed an analytical platform that directly transduces and catalyzes enhanced RNA signaling, enabling reliable and digital quantification of a wide range of circulating RNAs for disease analysis.
[0161] The EZ-READ platform achieves regenerative sensing and digital quantification through: 1) a catalytic sensor that releases highly efficient HRP@ZIF-8 nanoparticles for individual digital quantification; 2) efficient RNA target recovery via DSN. This enzyme recognizes RNA-DNA heteroduplexes (rather than sequence-specific cleavage sites) and cleaves only the DNA strand to produce very short fragments; this mechanism enables not only regenerative sensing but also the expansion of programmable analysis; 3) its microfluidic platform features a fractal branching network of microchannels. This design ensures equal fluid resistance to each microwell, enabling uniform reagent loading and independent digital counting reactions without complex liquid processors, and can be easily expanded to increase the number of reaction wells. Compared to conventional RNA detection methods (e.g., hybridization and target amplification methods), the EZ-READ platform is well-suited for the direct and accurate determination of a wide range of RNA signatures (Table 4). Compared to traditional hybridization-based analyses (e.g., molecular beacons based on fluorescence resonance energy transfer / FRET) that rely on target binding to induce fluorescence signal changes and have limited sensitivity, EZ-READ achieves sensitive and accurate digital quantification. Compared to target amplification analysis (such as RT-qPCR) that requires large amounts of sample processing (e.g., reverse transcription and target amplification), EZ-READ bypasses all the steps of conventional PCR to enable direct and robust detection of RNA isoforms (miRNA and mRNA) of different sizes.
[0162]
[0163] The developed technology has broad scientific application potential. It features programmable detection and robust performance in minimally processed samples. By incorporating novel probes (see Li, L. et al. (2021); Sundah, NR et al. (2021)), nanomaterials (see Lian, X. et al. (2017); Wang, Z. et al. (2021)), and reactive cascades (see Chen, Y. et al. (2021); Broto, M. et al. (2022); Zhao, H. et al. (2023)), the technology can be readily extended to determine other biomarkers. Similarly, the integration of advanced microfluidic technologies (see Yeh, EC et al. (2017); Yelesswarapu, V. et al. (2019); Lim, GS et al. (2019)) will not only enhance the analytical performance of this technology but also expand its multiplexing and throughput capabilities for the discovery and evaluation of various biomarkers (e.g., determination of TERT promoter mutations and larger patient cohorts). Clinically, the accurate characterization of disease states and molecular subtypes using minimally invasive methods can bring new advances to the personalized treatment of GBM and other diseases (see Bi, J. et al. (2020); Ricklefs, FL et al. (2018)).
[0164] References
[0165] Any listing or discussion of any seemingly prior art documents in this specification shall not be construed as an admission that such documents are part of the prior art or are common general knowledge.
[0166] Batool, SM et al. Highly sensitive EGFRvIII detection incirculating extracellular vesicle RNA of glioma patients. Clin Cancer Res 28,4070-4082 (2022).
[0167] Bi, J. et al. Altered cellular metabolism in gliomas - an emerginglandscape of actionable co-dependency targets. Nat Rev Cancer 20, 57-70(2020).
[0168] Broto, M. et al. Nanozyme-catalysed CRISPR assay forpreamplification-free detection of non-coding RNAs. Nat Nanotechnol 17, 1120-1126 (2022).
[0169] Chen, Y. et al. Collaborative equilibrium coupling of catalytic DNAnanostructures enables programmable detection of SARS-CoV-2. Adv Sci (Weinh)8, e2101155 (2021).
[0170] Chong, Y. K. et al. Cryopreservation of neurospheres derived fromhuman glioblastoma multiforme. Stem Cells 27, 29-39 (2009).
[0171] Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner.Bioinformatics 29, 15-21 (2013).
[0172] Duy, C. & Fitter, J. How aggregation and conformational scrambling ofunfolded states govern fluorescence emission spectra. Biophys J 90, 3704-3711(2006).
[0173] Figueroa, J. M. et al. Detection of wild-type EGFR amplification andEGFRvIII mutation in CSF-derived extracellular vesicles of glioblastomapatients. Neuro Oncol 19, 1494-1502 (2017).
[0174] Hastie, T., Tibshirani, R., Narasimhan, B. & Chu, G. pamr: Pam:Prediction Analysis for Microarrays. R package version 1.56.1. (2019).
[0175] Huang, S. W., Ali, N. D., Zhong, L. & Shi, J. MicroRNAs as biomarkersfor human glioblastoma: progress and potential. Acta Pharmacol Sin 39, 1405-1413 (2018).
[0176] Labib, M. et al. Single-cell mRNA cytometry via sequence-specificnanoparticle clustering and trapping. Nat Chem 10, 489-495 (2018).
[0177] Li, B. & Dewey, C. N. RSEM: accurate transcript quantification fromRNA-Seq data with or without a reference genome. BMC Bioinform 12, 323(2011).
[0178] Li, L. et al. Nucleic acid aptamers for molecular diagnostics andtherapeutics: advances and perspectives. Angew Chem Int Ed Engl 60, 2221-2231(2021).
[0179] Lian, X. et al. Enzyme-MOF (metal-organic framework) composites. ChemSoc Rev 46, 3386-3401 (2017).
[0180] Lim, G. S. et al. Microhexagon gradient array directs spatialdiversification of spinal motor neurons. Theranostics 9, 311-323 (2019).
[0181] Lottaz, C. et al. Transcriptional profiles of CD133+ and CD133-glioblastoma-derived cancer stem cell lines suggest different cells oforigin. Cancer Res 70, 2030-2040 (2010).
[0182] Reátegui, E. et al. Engineered nanointerfaces for microfluidicisolation and molecular profiling of tumor-specific extracellular vesicles.Nat Commun 9, 175 (2018).
[0183] Ricklefs, F. L. et al. Immune evasion mediated by PD-L1 onglioblastoma-derived extracellular vesicles. Sci Adv 4, eaar2766 (2018).
[0184] Shagin, D. A. et al. A novel method for SNP detection using a newduplex-specific nuclease from crab hepatopancreas. Genome Res 12, 1935-1942(2002).
[0185] Shao, H. et al. Protein typing of circulating microvesicles allowsreal-time monitoring of glioblastoma therapy. Nat Med 18, 1835-1840 (2012).
[0186] Shao, H. et al. Chip-based analysis of exosomal mRNA mediating drugresistance in glioblastoma. Nat Commun 6, 6999 (2015).
[0187] Shao, H. et al. New technologies for analysis of extracellularvesicles. Chem Rev 118, 1917-1950 (2018).
[0188] Skog, J. et al. Glioblastoma microvesicles transport RNA and proteinsthat promote tumour growth and provide diagnostic biomarkers. Nat Cell Biol10, 1470-1476 (2008).
[0189] Sundah, N. R. et al. Catalytic amplification by transition-statemolecular switches for direct and sensitive detection of SARS-CoV-2. Sci Adv7, eabe5940 (2021).
[0190] Tavallaie, R. et al. Nucleic acid hybridization on an electricallyreconfigurable network of gold-coated magnetic nanoparticles enables microRNAdetection in blood. Nat Nanotechnol 13, 1066-1071 (2018).
[0191] Théry, C., Ostrowski, M. & Segura, E. Membrane vesicles as conveyorsof immune responses. Nat Rev Immunol 9, 581-593 (2009).
[0192] Théry, C. et al. Minimal information for studies of extracellularvesicles 2018 (MISEV2018): a position statement of the International Societyfor Extracellular Vesicles and update of the MISEV2014 guidelines. JExtracell Vesicles 7, 1535750 (2018).
[0193] van Niel, G., D’Angelo, G. & Raposo, G. Shedding light on the cellbiology of extracellular vesicles. Nat Rev Mol Cell Biol 19, 213-228 (2018).
[0194] Wang, Q. et al. Tumor evolution of glioma-intrinsic gene expressionsubtypes associates with immunological changes in the microenvironment.Cancer Cell 32, 42-56.e6 (2017).
[0195] Wang, Z. et al. Surfactant-guided spatial assembly of nano-architectures for molecular profiling of extracellular vesicles. Nat Commun12, 4039 (2021).
[0196] Yeh, E. C. et al. Self-powered integrated microfluidic point-of-carelow-cost enabling (SIMPLE) chip. Sci Adv 3, e1501645 (2017).
[0197] Yekula, A. et al. Large and small extracellular vesicles released byglioma cells in vitro and in vivo. J Extracell Vesicles 9, 1689784 (2020).
[0198] Yelleswarapu, V. et al. Mobile platform for rapid sub-picogram-per-milliliter, multiplexed, digital droplet detection of proteins. Proc NatlAcad Sci U S A 116, 4489-4495 (2019).
[0199] Zhao, H. et al. A hydrogel-based mechanical metamaterial for theinterferometric profiling of extracellular vesicles in patient samples. NatBiomed Eng 7, 135-148 (2023).
Claims
1. An RNA-responsive sensor comprising a catalytic nanoparticle complex linked to a magnetic bead by a specific single-stranded DNA probe to a target RNA, wherein the single-stranded DNA probe will form a duplex with the target RNA, and wherein in the presence of a duplex-specific nuclease that recognizes the RNA-DNA duplex, the DNA linker is cleaved, thereby: i) releasing the catalytic nanoparticle complex, and ii) releasing the target RNA to form a DNA-RNA duplex with another RNA-responsive sensor.
2. The RNA-responsive sensor of claim 1, wherein the catalytic nanoparticle complex comprises one or more enzymes encapsulated in a metal-organic framework.
3. The RNA-responsive sensor of claim 2, wherein the metal-organic framework is selected from the group comprising zeolitic imidazolate framework-8 or 2-methylimidazolate zinc salt (ZIF-8), zeolitic imidazolate framework-67 or 2-methylimidazolate cobalt salt (ZIF-67), 1,4-benzenedicarboxylic acid (BDC), and 1,3,5-benzenetricarboxylic acid (BTC).
4. The RNA-responsive sensor of claim 2 or 3, wherein the one or more enzymes are selected from the group comprising horseradish peroxidase (HRP), acetylcholinesterase, dehydratase, catalase, cellulase, cytochrome c, dehydrogenase, epoxide hydrolase, glucoamylase, glucose dehydrogenase, glucose oxidase, hemin, hydrolase, invertase, laccase, lipase, microperoxidase-11, myoglobin, organophosphate dehydratase, pectinase, protease, soybean epoxide hydrolase, trypsin, tyrosinase, urease, xylanase, a-amylase, b-galactosidase, b-glucosidase.
5. The RNA-responsive sensor of any one of claims 1 to 4, wherein the target RNA is selected from the group comprising mRNA, miRNA, IncRNA, rRNA, and tRNA.
6. The RNA-responsive sensor of any one of claims 1 to 5, wherein the target RNA is present in a body fluid sample, preferably a blood sample.
7. The RNA-responsive sensor of claim 6, wherein the body fluid sample is from a diseased subject.
8. The RNA-responsive sensor of claim 7, wherein the body fluid sample is from a subject afflicted with glioblastoma.
9. The RNA-responsive sensor of claim 8, wherein the specific single-stranded DNA probe targets an RNA selected from the group comprising miR-17-5p, lysyl oxidase (LOX) mRNA, and solute carrier organic anion transporter family member 3A1 (SLCO3A1) mRNA, or a combination thereof.
10. The RNA-responsive sensor of claim 8, wherein the specific single-stranded DNA probe targets an RNA selected from the group comprising miR-21-5p, miR-27a-3p, and aldehyde dehydrogenase 1A3 (ALDH1A3) mRNA, or a combination thereof. 11. A method of determining the abundance of a biomarker in a sample, the method comprising: a) binding a biomarker-responsive sensor to the sample, the biomarker-responsive sensor configured to release catalytic nanoparticle complexes in response to the biomarker; b) introducing supernatant of the sample and the biomarker-responsive sensor combined into a microfluidic device having a plurality of reaction wells; c) introducing a colorimetric or other signaling substrate into the reaction wells, wherein the released catalytic nanoparticle complexes catalyze a reaction with the colorimetric or other signaling substrate, thereby generating a detectable reaction signal; d) acquiring an image of the microfluidic device, the image capable of showing signal-positive and / or signal-negative reaction wells; e) determining from the image a count of reaction wells that generated a detectable signal; and f) determining from the count of reaction wells that generated a detectable reaction signal the abundance of the biomarker in the sample.
12. The method of claim 11, wherein the substrate is a chemiluminescent substrate and the detectable reaction signal is fluorescence.
13. The method of claim 11 or 12, wherein the biomarker-responsive sensor is an RNA-responsive sensor according to any one of claims 1 to 10.
14. The method of any one of claims 11 to 13, wherein the microfluidic device comprises a plurality of fractal branching paths extending from an inlet to respective reaction wells in the plurality of reaction wells.
15. The method of claim 14, wherein the plurality of fractal branching paths are configured to have substantially equal fluidic resistance.
16. The method of claim 14 or claim 15, wherein the plurality of fractal branching paths are configured to have substantially equal length between the inlet and each respective reaction well.
17. The method of any one of claims 14 to 16, wherein a branching angle of each fractal branching path ranges from 85 degrees to 95 degrees, or preferably ranges from 89 degrees to 91 degrees.
18. The method of any one of claims 11 to 17, further comprising sealing each reaction well after introducing the chemiluminescent substrate to isolate the fluorescent reaction in each respective reaction well.
19. The method of any one of claims 11 to 18, wherein determining from the fluorescent image a count of reaction wells that generated a fluorescent signal comprises comparing a fluorescent intensity of each respective reaction well to a threshold value, and determining the count as a number of reaction wells having a fluorescent intensity above the threshold value.
20. An in vitro method of detecting and / or classifying a disease suffered by a subject, comprising performing the method of any one of claims 11 to 19 on a sample of bodily fluid from the subject, and using the abundance of at least one biomarker associated with the disease to detect the disease and / or to classify the disease.
21. The method of claim 20, wherein the disease is glioblastoma, and i) the RNA-responsive sensor is defined according to claim 9 for detecting the disease, or ii) the RNA-responsive sensor is defined according to claim 10 for classifying the disease. ii) the RNA-responsive sensor is defined according to claim 10 for classifying the glioblastoma as mesenchymal subtype or pre-neural / classic subtype.
22. A microfluidic device comprising a plurality of fractal branching paths extending from an inlet to respective ones of a plurality of reaction wells.
23. The microfluidic device of claim 22, wherein the plurality of fractal branching paths are configured to have substantially equal fluidic resistance.
24. The microfluidic device of claim 22 or 23, wherein the plurality of fractal branching paths are configured to have substantially equal length between the inlet and each respective reaction well.
25. The microfluidic device of any one of claims 22 to 24, wherein a branching angle of each fractal branching path ranges from 85 degrees to 95 degrees, or preferably ranges from 89 degrees to 91 degrees.
26. A kit for screening a body fluid sample for the presence of an RNA indicative of a disease, wherein the RNA is differentially expressed in a diseased state compared to a reference sample representative of a subject not suffering from the disease, the kit comprising the RNA-responsive sensor of any one of claims 1 to 10.
27. The kit of claim 26, wherein the RNA-responsive sensor comprises a specific single-stranded DNA probe for a target RNA that will form a duplex with the target RNA, wherein the target RNA is selected from: a) the group comprising miR-17-5p, LOX mRNA and SLCO3A1 mRNA, or b) the group comprising miR-21-5p, miR-27a-3p and ALDH1A3 mRNA.
28. A glioblastoma screening kit for screening a body fluid sample for the presence of an RNA indicative of a glioblastoma that is clinically pathologically of mesenchymal, pre-neural or classic subtype, wherein the RNA is differentially expressed in a diseased state compared to a reference sample representative of a subject not suffering from the disease, wherein the kit comprises one or more RNA-responsive sensors of claim 10.
29. The kit of any one of claims 26 to 28, further comprising the microfluidic device of any one of claims 22 to 25.
30. The kit of any one of claims 26 to 29, further comprising: i) a duplex-specific nuclease, and / or ii) a substrate that reacts with the catalytic nanoparticle complex released from the RNA-responsive sensor.
