Summary
Chronic musculoskeletal pain (CMP) is the most common chronic pain in clinical practice. With the aging population and lifestyle changes, the prevalence of CMP is increasing, and the low treatment and remission rates of CMP significantly reduce patients’ quality of life and cause a waste of medical resources. The accurate distinction between neuropathic and nociceptive CMP can help improve the efficiency of diagnosis and treatment and save medical resources. Our project aims to achieve the classification of CMP through the detection of miRNA biomarkers. To achieve this goal, the feasibility of the pain classification mechanism, biotechnology route, hardware platform for reaction and detection, and final user operation and promotion were all verified from the perspective of the project as a whole. Each step of the biological reaction and the continuous triggering between reactions were verified in the laboratory, and it was proved that the reaction could be carried out on the paper chip and the system could output the required results. Each part shows that our project can achieve the requirements well. For more detailed content, you can see the "Hardware" ,"Modeling", "Results" pages.
Biosensor system
Technical feasibility verification of biosystems
Through the modeling section, we selected two miRNA, miR-98-5p and miR-7d-5p. To realize pain classification based on miRNA detection, we designed an HCR-CRISPR bioassay system that can perform enzyme-free isothermal amplification and quantitative fluorescence detection. The system uses HCR technology to perform isothermal amplification on the miRNA to be tested, and the downstream CRISPR/Cas12a system recognizes and cleaves HCR amplification products, achieving signal detection and transduction to obtain visual results. We did the following experiments to verify that each step of the reaction can occur, the reactions can take place continuously and the final output can meet the demand. We designed corresponding H1 probes for two miRNAs (miR-98-5p and miR-7d-5p) and carried out experiments respectively. Next, we describe the results of miR-98-5p detection.
Verification of nucleic acid isothermal amplification system
HCR is an enzyme-free nucleic acid polymerization reaction, and a cascade reaction of several self-stable DNA stem-loop structures is triggered only by the target molecule to produce a long-strand DNA structure with an incision. The system contains an H1 variable hairpin and H2 and H3 universal hairpins. Each starting target molecule can trigger a hybrid chain reaction to form an ultra-long DNA strand, and the signal of the target molecule can be amplified by the hybrid chain reaction.
To verify the feasibility of HCR detection of miRNA, we carried out an HCR reaction experiment based on miRNA to verify the feasibility of the HCR reaction of miR-98-5p, miR-7d-5p, and probe H1 single trigger chain (CF) respectively. The reaction products of the three reactions were analyzed by gel electrophoresis, and the following experiment results were obtained. It can be seen that Lane 1, Lane 2, Lane 3 and Lane 4 all have tailing phenomenon, which means that there are different sequences of relative molecular mass in the corresponding HCR results of this lane. Lane 5 does not have tailing phenomenon, which means that HCR reaction cannot be triggered in the absence of triggers. The experiment results show that triggers can all trigger HCR reactions, which proves that this system can carry out the HCR amplification of miRNA.
1:D2000；2:miRNA(1μM)；3:H1；
4:miRNA (1μM)+ H1；5:miRNA(1uM)+H1+H2；6:H1+H2+H3；
7:miRNA(600nM)+H1+H2+H3；8:miRNA(400nM)+H1+H2+H3；9:miRNA(200nM)+H1+H2+H3
1:CF(100nM)+H2+H3；2:CF(300nM)+H2+H3；3:CF(600nM)+H2+H3；
4:CF(1μM)+ H2+H3；5:H2+H3；6:D2000
Verification of nucleic acid signal detection and energy conversion
Through the HCR reaction, a DNA duplex strand (miRNA/H1/(H2/H3)n) containing a PAM sequence and the original spacer sequence that can be specifically recognized by Cas12a/crRNA complex is formed to realize the transformation of miRNA into DNA. The CRISPR reaction is to activate the fluorophore quencher (FQ) whose side chain cleavage activity is not specific cleavage nucleic acid marker after recognizing the DNA signal generated by the HCR reaction through the Cas12a/crRNA complex, so as to achieve the specific fluorescence detection of target nucleic acid.
To verify the feasibility of HCR-CRISPR system for detecting miRNAs, we used positive templates to preliminarily prove that. The positive template is a double-stranded DNA synthesized with fragments of H2 and H3 that can be recognized by crRNA. We used different concentrations of positive templates to trigger the CRISPR reaction, and the image of its fluorescence intensity changing with time is shown in the following figure. According to the experiment results, it can be seen that compared with the blank group, the addition of the positive template can significantly increase the fluorescence intensity. And in the concentration range involved in the experiment, the higher the concentration of positive template, the higher the fluorescence intensity, which proves that the HCR-CRISPR system can realize the detection of miRNAs.
To explore the influence of fluorescence probe concentration on experimental results and determine the most appropriate fluorescence probe concentration, we conducted experiments with fluorescence probe concentrations of 2.5μM, 5μM, 7.5μM and 10μM respectively. The experiment results are as follows. According to the results, under different concentrations of the fluorescent probe, the effect of 5uM fluorescent probe concentration and 10uM concentration is better, and the effect of the two is not much different, so the fluorescence probe concentration is set at 5uM.
Optimization of reaction conditions
Optimization of nucleic acid isothermal amplification system
To screen out better reaction conditions, avoid amplifying wrong signals, and facilitate the subsequent CRISPR reaction, we optimized several conditions in the HCR system, including stem-loop length, probe ratio, ion concentration, etc.
Stem loop length optimization
Because different stem-loop lengths will change the stability of H1, the stem-loop ratio will affect the difficulty of opening the H1 structure, and an inappropriate stem-loop ratio will easily produce false positive results or cause difficulties in H1 structure change and affect the subsequent amplification intensity. Therefore, we explored the experimental effect under different stem-loop lengths through experiments. When the end length of H1 is 8, 11, and 13 (which referred to as H1-8, H1-11, H1-13), the reaction results of the HCR system with miRNA were tested, and the experiment results are as follows. According to the results, the HCR response was different under different stem-loop lengths, and the best response was achieved when the length of the H1 end was 11.
3：H1-11+H2+H3+miR-98-5p(300nM); 4：H1-11+H2+H3;
5：H1-8+H2+H3+miR-98-5p(300nM); 6：H1-8+H2+H3; 7：D2000
1:miRNA(300nM)+H1(15)+H2+H3；2:H1(15)+H2+H3；
3:miRNA(300nM)+H1(13)+H2+H3；4:H1(13)+H2+H3；
5:miRNA(300nM)+H1(11)+H2+H3；6:H1(11)+H2+H3；
7:miRNA(300nM)+H1(9)+H2+H3；8:H1(9)+H2+H3；
9:miRNA(300nM)+H1(7)+H2+H3；10:H1(7)+H2+H3；11:D2000
Probe ratio optimization
To screen out the probe ratio with better reaction effect, we tested the reaction results of the HCR system with miRNA when the concentration of H1 was 0.1μM, 0.3μM, 0.6μM and 0.9μM. The gel electrophoresis results are shown as follows. According to the analysis of the experiment results, the probe ratio has a certain degree of influence on the HCR reaction, among which the H1 concentration of 0.6μM has the best reaction effect.
3: miR-98-5p+H1-98-5p（0.6）+H2+H3+Na+；4: miR-98-5p +H1-98-5p（0.4）+H2 +H3+Na+
5: miR-98-5p +H1-98-5p（0.2）+H2+H3+Na+；6: miR-98-5p +H1-98-5p（0.1）+ H2+H3+Na+
7: H1-98-5p+H2+H3+Na+；8：D2000
1: D2000；2: H1（0.6μM）+H2+H3；
3: miRNA(300nM)+H1（0.1μM)+H2+H3；4: miRNA(300nM)+H1（0.3μM）+H2+H3；
5: miRNA(300nM)+H1（0.6μM）+H2+H3；6: miRNA(300nM)+H1（0.9μM）+H2+H3
Ion concentration optimization
To obtain better experiment results, we optimized the concentration of Na+ and tested the reaction results of the HCR system (H1 is H1-98-5p) with miRNA when the concentration of Na+ is 95nM, 140nM, 350nM, 550nM, and 750nM, respectively. The experiment results are shown as follows. According to the analysis of the experiment results, The reaction degree of HCR was different with different Na+ concentrations, and it was found that the reaction effect was best when the Na+ concentration was 350nM.
3:miRNA(400nM)+H1(0.6μM)+H2+H3+Na+(350nM)；4:miRNA(400nM)+H1(0.6μM)+H2+H3+Na+(140nM)；
5:miRNA(400nM)+H1(0.6μM)+H2+H3+Na+(95nM)；6:H1(0.6μM)+H2+H3+Na+(350nM); 7:D2000
Optimization of nucleic acid signal detection and energy conversion system
Probe stem-loop length optimization
We explored the experiment effects of different probe stem loop lengths to find the appropriate probe stem lengths for the HCR-CRISPR system, and the experiment results were as follows. According to the experiment results, under the H1-2 probe reaction, the experimental blank had low false positives and the control reaction effect was good, so the H1-2 probe was selected for the reaction.
Determination of the detection limit of reaction
In order to determine the detection limit of the reaction, we tested the reaction of HCR-CRISPR system under different miRNA concentrations through experiments. The experiment results are shown in the following figure.
Specificity
In order to explore the specificity of the reaction system for the detection of the target miRNA, we detected the participation of three different miRNAs (miR-98-5p, miR-320A, and miR-7d-5p) in the reaction of the HCR-CRISPR system. To test the influence of non-target miRNAs on target miRNA detection, the experiment results are shown in the following figure. According to the experiment results, the HCR-CRISPR reaction can be triggered in the presence of miR-98-5p, and the fluorescence effect is obvious, while other miRNAs will not trigger the HCR-CRISPR reaction.
After a series of optimization, we completed the feasibility verification of the reaction principle of each mechanism of the biological detection system, realized the quantitative detection of miRNA, and found the more appropriate reaction conditions, which is convenient for the subsequent testing and verification of the hardware system, in order to finally realize the miRNA-based pain classification system.
Hardware platform feasibility verification
Our hardware consists of a paper-based chip module, a fluorescence detection module, a temperature control module, and a mobile App. For the paper chip module, we validated its performance and its ability to integrate biological reactions and output fluorescence. For the fluorescence detection module, we mainly validated the linear relationship between the processed data results and the fluorescence intensity. For the temperature control module, we verified that the device can continuously and uniformly provide temperature for the reaction. Finally, we validated the overall functional implementation of the system.
Paper-based chip module
The paper-based chip module consists of three parts: the transfer pad, the fluid pathway pad, and the reaction pad in a total of five layers. The transfer pad and fluid channel pad use asymmetric structures and hydrophobicity to control the correct and uniform flow of liquid, and the reaction pad is used to integrate the biological reaction. We verified the performance of the paper chip itself, its ability to integrate for biological reactions, and the normal output of fluorescence after the overall integrated reaction respectively.
Paper-based chip performance verification
We first verified the effect of transfer pads and fluid channel pads on fluid flow control through leakage testing and uniformity testing. The sample volume of the chip was also tested to ensure that the sample volume is more appropriate. The test results and process diagrams are shown below. It is proven that the liquid can flow evenly into the reaction pad without leakage after passing through the paper chip. The sample volume of the paper-based chip is about 60 μL.
Biological reaction validation
We verify the normal realization of biological reactions on the reaction pad. The electrophoresis results after the HCR reaction in the laboratory, on the dry paper base and the wet paper base, are shown below. It can be seen that the reaction results on the wetted paper substrate are not much different from the laboratory, while the bands on the dry paper substrate are clearer, and the bands have a linear relationship with the sample concentration. This successfully verified the compatibility of the paper-based chip with the reaction and the tolerance of the dry paper-based chip to the reaction.
Final fluorescence output verification
In order to achieve the output of the final result and the cooperation with the fluorescence and other modules, we performed overall validation for the two-step HCR+CRISPR reaction ensuring normal fluorescence output after the reaction. It is verified that the paper-based chip can generate the two-step reaction normally and output the fluorescence normally.
Fluorescence detection module
Our device uses a smartphone to take fluorescence shots and then achieves matching with fluorescence intensity by certain image data processing. The schematic diagram of the device is shown below. The linear relationship between the processed data and the fluorescence intensity is used to verify the proper operation of our equipment and the correctness of the data processing method.
We tested the fluorescent device using sodium fluorescein solutions with different concentration gradients. The result plots and data fitting plots are as follows. The results show good linearity between data results and fluorescence intensity, successfully verifying the normal operation of the device.
(A) Three sets of fluorescent original film shooting pictures
(B) Relationship between grayness of images and sodium fluorescein solution
Temperature control module
The temperature control module consists of four main parts as follows. We first tested whether the equipment could maintain a stable temperature of 37℃. Then the uniformity, accuracy, and precision of the temperature control of the equipment were tested separately to ensure that the equipment can provide more accurate temperature control for the reaction.
The temperature data collected by the sensors were monitored in real-time, and the temperature changes per second were derived to create the images shown below. The results show that our incubation device can maintain a stable temperature of about 37℃ and the heating process heats up at a uniform rate.
The uniformity, accuracy, and precision of the module temperature control were tested experimentally. The results show that the temperature at six positions is basically the same, which proves that the uniformity of temperature control of the equipment is good. The difference between the average value and the set temperature is no more than 0.1℃ in absolute value, which proves that the accuracy of temperature control of the equipment is good. Half of the difference between the highest temperature and the lowest temperature is within 0.2℃ from the preset temperature, which proves that the precision of temperature control of the equipment is good.
Overall system validation
The overall schematic diagram after the assembly of each module is as follows. Through the reasonable design of relative position and convenient pulling and one-time feeding, the quantitative output of final results and reasonable user interaction can be realized.
We went through overall user testing to ensure that the entire unit was working properly and could meet our needs. A demonstration of the final system in use is shown below.
Model
miRNA model
Although pain has been characterized as a disease, we lack an exact theory that can explain how miRNA affects pain symptoms from a biological perspective due to various mechanisms that induce pain. However, many studies have shown that when a specific kind of pain occurs, there will be sharp fluctuations in the corresponding miRNA concentrations in the patient's body, and many times when the corresponding miRNA is studied, it can be promoted as a means of pain treatment. Based on the above research background, we try to make some push in the cross direction of pain & miRNA, that is, whether miRNA can be used as a marker to preliminarily determine the type of pain in patients. We divided the types of pain into Np (neuropathic pain) and No (nociceptive pain).
Our project goal is to determine the type of pain in patients by detecting the concentration of two miRNAs. The first problem is how to select the most appropriate miRNAs from a wide range of miRNA populations as biomarkers. For our screening and classification model, the innovation lies in how to find the miRNAs with the highest correlation with pain, due to the limitations of actual detection and sampling conditions, We decided to screen the most appropriate miRNAs from 119 miRNAs with the highest concentration in the blood. Finally, we decided to screen the most appropriate miRNAs based on the 14 kinds of pain-related miRNAs (as follows) obtained by Dayer et al [1]. The clinical data of patients' pain type miRNA concentration in the literature.
After screening with multiple models, the best two miRNAs are hsa-miR-98-5p and hsa-miR-7d-5p. We hope to determine the pain type of patients through classification model analysis after detecting the concentration of the two types of miRNAs. Combining the clinical data provided by literature [1], we have trained a decision tree as a classification model for the project. The classification accuracy of the model is about 70%, which matches the accuracy of the binary regression model adopted by Dayer et al. Because the decision tree model has fewer features to learn, that is, the data volume of the training set needs to be supplemented more, which is partly due to the lack of experimental research in this field, We need more clinical data to promote progress in the field of miRNA & pain, from which we can also develop more useful classification models.
NUPACK model
When performing an HCR reaction using the probe sequence given in the references, the results were not ideal under our laboratory conditions. For better results, we use NUPACK for modeling to achieve optimization of the probe sequence.
NUPACK mainly has three functions: design, analysis, and application, and we mainly use the analysis function.
Take probe sequence optimization as an example: we made changes to the probe sequence based on the probe already in the lab:
The initial sequence and the optimized sequence are:
5-AGTCTAGGAAACTGCGTGGGTTAAAACAATACAACTTACTACCTCATTAACCCA
5-ATCAGACTGAAAGATGTTGACAAAGTAACAATACAACTTACTACCTCAACTTTGTCAAC
The comparison of its structure with NUPACK simulation is as follows:
There is no significant change in its structure, but there is a significant change in the structure free energy, which also leads to easier binding of miRNA to H1, triggering the subsequent HCR reaction, so the probe HCR reaction after changing the sequence is carried out more.
The mismatch between the H1 stem length and the probe also affects the free energy of the probe, which in turn affects the result of the HCR reaction, and we also used NUPACK to optimize it and select the probe structure with the best effect.
Using NUPACK to simulate the HCR reaction, the optimization of the HCR reaction probe was realized, and at the same time, the number of actual experiments we conducted was greatly reduced, which promoted the progress of the project and played an important role in guiding and promoting our experiments.
CRISPR model
We hope to fit the kinetic indexes
In the process of building the model, in addition to the equation derivation method used to establish the Michaelis-Menten equation, we make reasonable assumptions and use the method to understand the differential equation to get the desired analytical solution. As for the calculation of the parameters
The ordinary differential equation is solved to obtain:
where
Clearly, τ decreases with increasing
Our modeling analysis has guiding significance for the CRISPR experimental optimization. Based on this analytical solution and a series of CRISPR experiments, we proposed three indexes to evaluate the self-consistency and correctness of the experiments. By analyzing the influence of the parameters in the equation on the solution, the improvement direction of the experiment is put forward. The experiment time is predicted by plotting the normalized variables with t/τ as the horizontal axis and P/
[1] Dayer, Camille Florine, et al. Differences in the miRNA signatures of chronic musculoskeletal pain patients from neuropathic or nociceptive origins[J]. PloS one 14.7 (2019): e0219311.