Results
Overview
The goal of our project is to develop a breast cancer screening kit, BreFast. Our project integrated the
knowledge of computer science, molecular biology and biochemistry. We used bioinformatics approaches to analyze
breast cancer-associated miRNA (microRNA), and selected miR-21-5p as the best biomarker for detection by
BreFast. Subsequently, we verified the expression of miR-21-5p in breast cancer cells by qRT-PCR, and explored
the function of miR-21-5p by wound healing and transwell assays. Furthermore, we designed a crRNA, an essential
component of CRISPR/Cas13a system, to target miR-21-5p, and made an engineered CRISPR/Cas13a system that could
be activated by miR-21-5p. The cleavage efficiency of the CRISPR/Cas13a system was investigated by fluorescence
reaction and kinetics simulation. Finally, the efficacy of the CRISPR/Cas13a system on detecting high levels of
miR-21-5p was demonstrated with lateral flow strips. In summary, we have combined CRISPR/Cas13a system and
lateral flow strips into a convenient and reliable kit to detect miRNA associated with breast cancer (Figure 1).
Figure 1 The workflow of the Results section.
miR-21-5p
Bioinformatics Analysis
During the progression of breast cancer, the expression levels of miRNAs are changed. Some of these miRNAs are
closely associated with breast cancer and can be used as potential biomarkers for detection of breast cancer. We
focused on utilizing miRNAs that show increased expression in breast cancer as biomarkers for detection by
BreFast.
Differential expression analysis
To identify miRNA with significant changed expression in breast cancer patients, we analyzed TCGA-BRCA dataset
in TCGA database and GSE59247 dataset in GEO database, and selected miRNA with thresholds of |logFC|>1 and
PValue<0.05. 118 up-regulated and 98 down-regulated miRNA were identified from GSE59247 dataset, and 226
up-regulated and 107 down-regulated miRNA were identified from TCGA-BRCA dataset.
| Dataset |
Upregulation |
Downregulation |
| GSE59247 |
118 |
98 |
| TCGA-BRCA |
226 |
107 |
Table 1 Number of miRNA with significantly changed expression in breast cancer.
Venn diagram was plotted to show the total numbers of miRNA with increased expression in breast cancer from
GSE59247 dataset and TCGA-BRCA dataset, and the intersection of the Venn diagram shows that 9 miRNA were
identified from both datasets.
Figure 2 Venn diagram showing the number of up-regulated miRNA from GSE59247 and
TCGA-BRCA datasets.
Volcano plots were generated to show the distribution of the miRNA on fold change and p-value from both
GSE59247 dataset and TCGA-BRCA dataset.
Figure 3 Volcano plot of miRNA from GSE59247 dataset.
Figure 4 Volcano plot of miRNA from TCGA-BRCA dataset.
ROC Curve
To further refine the selection of target miRNAs for BreFast detection, we plotted ROC curves for the 9 miRNAs
identified from differential expression analysis. Since the sample size in the GSE59247 dataset was relatively
small, we used the TCGA-BRCA dataset to generate the ROC curves, as shown in the figure below.
Figure 5 ROC curves generated from TCGA-BRCA data.
The area under the curve is an important index of the ROC curve. The larger the area under the curve, the
higher the sensitivity and specificity of the corresponding miRNA. AUC>0.9 was used as a criterion for
selection, and miR-21-5p, miR-183-5p and miR-96-5p were selected for subsequent analysis.
Feature importance analysis
Feature importance analysis is a machine learning method to analyze how important of a feature to a certain
characteristics (label). We treated each miRNA as a feature, and labeled samples as positive or negative based
on whether they were from cancer patients or normal persons, thus transforming the problem of selecting optimal
cancer-associated miRNA into a machine learning feature importance issue. Datasets were divided into training
data and testing data in an 8:2 ratio. We used grid search and five-fold cross-validation to find the optimal
parameters (achieving an accuracy of 0.98924 in five-fold cross-validation). The binary classification task was
performed using a random forest model, and the accuracy reached 1 on the testing dataset. Feature importance was
calculated based on the Gini index, and the top 10 ranked miRNAs were shown below.
Figure 6 The feature importance of the top ten miRNAs analyzed by random forest, where
the upregulated and downregulated miRNAs have been labeled based on the results of differential expression
analysis of TCGA data.
While tree models, such as random forest, can reflect the classification value of a feature (each miRNA)
concerning sample (each patient) classification, they usually do not consider the interactions between different
features. Considering the possibility of strong interactions among different miRNAs[1], our team has employed
SHAP (SHapley Additive explanations) analysis, a methodology based on cooperative game theory, to conduct
further analysis on the feature importance derived from the Random Forest model. The top 10 miRNA with high SHAP
values were shown below, and miR-21-5p ranked first.
Figure 7 SHAP value analysis of the feature importance derived from Random Forest,
where the upregulated and downregulated miRNAs have been labeled based on the results of differential expression
analysis of TCGA data.
In summary, our bioinformatic analysis demonstrated that miR-21-5p is the optimal breast cancer-associated
miRNA for detection.
Survival Analysis
Differential expression analysis and feature importance analysis have been performed to identify miR-21-5p as
an optimal target for breast cancer screening. Next, we would like to know whether miR-21-5p could truly affect
the progression of breast cancer. Therefore, we did survival analysis of miR-21-5p with Kaplan-Meyer curves. The
data of Metabric and TCGA were used to draw the survival curve of miR-21-5p.
Figure 8 The survival curves for miR-21-5p, where Figure (a) represents data from the
Metabric database, and Figure (b) represents data from the TCGA dataset.
The results of survival analysis demonstrated that the survival time of breast cancer patients with high
expression of miR-21-5p was shorter than that of patients with low expression, and there was a statistically
significant correlation between miR-21-5p and the survival time of breast cancer patients.
Literature Research
In addition to bioinformatics analysis, we also read literatures to confirm that miR-21-5p is the best miRNA
candidate to be used as the optimal breast cancer-associated biomarkers for screening by Brefast.
A comprehensive review on International Journal of Molecular Sciences summarized 75 independent clinical
studies. 20 miRNAs were suggested to be breast cancer associated biomarkers, which were supported by at least
two independent studies in the same biological specimen (serum, plasma, or whole blood). 8 out 20 of them
demonstrated similar tendency of change in expression: miR-21 (12 up versus 1 down), miR-155 (14 up versus 1
down), miR-10b (5 up, only serum), miR-373 (3 up, only serum), miR-652 (3 down, only serum), miR-425 (2 down,
only serum), miR-29a (2 up, only serum), and miR-148b (2 down, only serum). Others miRNA, such as miR145
displayed different tendency of change in expression[2].
As a result, miR-21, miR-155, miR-10b were supported by
multiple studies as highly associated with breast cancer[3-4]. It is worthy to note that the predominant form
of miR-21 is miR-21-5p.
Figure 9 Pyramidal graph of the direction of miRNA expression (microRNA concentration
in breast cancer cases versus controls) by type of specimens (only microRNAs that were analysed in two or more
independent studies). (A) whole blood; (B) plasma; (C) all specimens; (D) serum.
In summary, both our own bioinformatic analysis and studies from multiple research groups demonstrated that
miR-21-5p is the best miRNA biomarker for detection of breast cancer.
Experimental Investigation
Studies have shown that miR-21-5p can promote peritoneal metastasis in gastric cancer[5]. Therefore, we not
only verified the expression of miR-21-5p in breast cancer cells, but also investigated the effects of miR-21-5p
on the migration and invasion of breast cancer cells.
miR-21-5p is highly expressed in breast cancer cells
To determine the expression of miR-21-5p in breast cancer cells, we did qRT-PCR to determine the relative
expression of miR-21-5p in the breast epithelial cell line MCF10A and the breast cancer cell line MCF7, which
showed that the expression of miR-21-5p in breast cancer cells was higher than that in normal breast cells.
(Figure 10).
Figure 10 Expression of miR-21-5p in MCF10A and MCF7 cells (A) qRT-PCR primer sequence
for miR-21-5p ; (B) Raw data of qRT-PCR ;
(C) miR-21-5p was highly expressed in breast cancer cells (p<0.001).
miR-21-5p enhances the migration and invasion abilities of breast cancer cells
The effect of miR-21-5p on the invasion capability of breast cancer cells was investigated by Transwell assay
(Figure 11). miR-21-5p was overexpressed in breast cancer cell line MDA-MB-231 by transfection. The invasion
ability of MDA-MB-231 transfected with miR-21-5p were significantly higher than that of cells transfected with
control RNA, demonstrating that miR-21-5p increased the invasiveness of MDA-MB-231 cells (Figure 12).
Figure 11 Schematic diagram of Transwell invasion assay.
Figure 12 (A) Penetration of cells in the compartment; (B) Cell count results (n=5, p
<0.001) .
To verify the effects of miR-21-5p on the migration of breast cancer cells, wound healing experiment was
conducted. MDA-MB-231 cells were transfected with miR-21-5p or control RNA, and the migration of cells were
recorded at 0h, 24h and 48h. The wound area of the MDA-MB-231 cells tranfected with miR-21-5p group was
significantly smaller than that of MDA-MB-231 cells tranfected with control RNA, indicating that miR-21-5p
enhanced the migration ability of MDA-MB-231 cells (Figure 13).
Figure 13 Wound healing assays of miR-21-5p transfected cells (A) Wound healing
experimental images taken at 0h, 24h and 48h; (B) Relative migration level (p<0.001).
In summary, the expression level of miR-21-5p in breast cancer cell MCF7 is much higher than that in breast
epithelial cell MCF10A, and increased expression of miR-21-5p enhanced the migration and invasion capability of
breast cancer cells. Together, these results demonstrated that increased expression of miR-21-5p had significant
effects on breast cancer cells.
LwaCas13a
Construction of pSB1C3-His-SUMO-LwaCas13a plasmid
Since we decided to use CRISPR/Cas13a system to detect miR-21-5p, we constructed a pSB1C3-His-SUMO-LwaCas13a
plasmid for expression of LwaCas13a. The schematic presentation of pSB1C3-His-SUMO-LwaCas13a plasmid was shown
below (Figure 14C). LacI operon was used for the induced expression of LwaCas13a[6], and a SUMO tag was added to
the N-terminal of LwaCas13a protein to increase the solubility[7]. To purify the LwaCas13a protein, a 6×His tag
was also added to the N-terminal of LwaCas13a protein[8].
The pSB1C3-His-SUMO-LwaCas13a plasmid was constructed using PCR amplification and Golden gate assembly
technology which is an excellent methods of DNA recombination capture the principal advantages of the BioBrick
standard[9]. Fragment containing SUMO-LwaCas13a was
amplified from plasmid pC013-TwinStrep-SUMO-huLwCas13a by
primers Cas13a F and Cas13a R. Fragment containing LacI operon was amplified from the plasmid
pC013-TwinStrep-SUMO-huLwCas13a by primers LacI F and LacI R. Fragment containing pSB1C3 backbone sequence was
amplified from the plasmid Pveg by primers Backbone F and Backbone R (Figure 14A,B).
Figure 14 Cloning strategy for pSB1C3-His-SUMO-LwaCas13a (A) Illustration of Pveg plasmid. (B) Illustration of
pC013-Twinstrep-SUMO-huLwCas13a. (C) Illustration of pSB1C3-His-SUMO-LwaCas13a plasmid.
The PCR products were analyzed by electrophoresis with 1% agarose gel, which showed that the pSB1C3 backbone
fragment (2086 bp), LacI operon fragment (1525 bp) and SUMO-LwaCas13a fragment (4227 bp) were amplified
successfully (Figure 15).
Figure 15 Gel electrophoresis of the PCR product for Golden Gate assembly of pSB1C3-His-SUMO-LwaCas13a plasmid.
Lane1: DNA marker. Lane2-Lane3: LwaCas13a. Lane4-Lane5: LacI operating system. Lane6-Lane7: pSB1C3 Backbone.
These fragments were assembled to generate pSB1C3-His-SUMO-LwaCas13a by Golden Gate assembly, and the assembled
products were transformed into Trans1-T1 competent bacterium (Figure 16).
Figure 16 Colonies that grow after transformation.
To select colonies with successful assembly, colony PCR was performed. PCR Product 1 was amplified by Primer VR
and Primer CHECK F that covered the junction between fragment SUMO-LwaCas13a and fragment LacI operon and the
junction between fragment LacI operon and fragment pSB1C3 backbone (3000 bp). PCR Product 2 amplified by Primer
CHECK R and Primer VF2 covered the junction between fragment pSB1C3 backbone and fragment SUMO-LwaCas13a (3050
bp). The two products were analyzed by agarose gel electrophoresis (Figure 17). Gel electrophoresis results
demonstrated that the Golden Gate assembly was successful.
Figure 17 Gel electrophoresis of the colony PCR product. (A) Illustration of PCR Product 1. (B) Illustration of
PCR Product 2. (C) Lane1: Marker. Lane2-Lane4: The product amplified by Primer VR and Primer CHECK F.
Lane5-Lane7: The product amplified by Primer CHECK R and Primer VF2.
For further verification, we carried out the endonuclease digestion. The restriction endonuclease EcoRI was
used for single cut, and the restriction endonuclease EcoRI and Spel were used for double cut, followed by 1%
agarose gel electrophoresis. Two bands (5712bp and 2068bp) were produced by digestion with both EcoRI and SpeI,
while one band (7780bp) was produced by digestion with EcoRI only. The electrophoresis results confirmed
successful assembly of pSB1C3-His-SUMO-LwaCas13a (Figure 18).
Figure 18 Plasmid pSB1C3-His-SUMO-LwaCas13a was verified by endonuclease digestion (A) Digestion using EcoRI and
SpeI results in two bands. One is 5712 bp, and the other is 2068 bp.; (B) Digestion using EcoRI enzymes results
in a band with a length of 7780bp.
(C) Lane1: Marker. Lane2: Plasmid was digested with EcoRI and SpeI. Lane3: Plasmid was digested with EcoRI.
Expression and purification of LwaCas13a protein
5 mL LB medium and 1 mL of bacteria-glycerol mixture were added into 15 mL tubes, and cultured on a shaker for
4 h until the bacteria culture were cloudy. They were then transferred into 1L flasks, and supplemented with
250mL of medium and 100 uL of CmR solution (50mg/mL), and incubated with shaking at 220rpm for 8 h. 1 mL of
bacterial solution was collected for later use. Subsequently, 200 μL of 0.5M IPTG was added to each vial to
induce protein expression and incubated overnight at 28°C. The cell lysate from bacteria before and after IPTG
induction were analyzed by SDS-PAGE gel electrophoresis, and the results demonstrated that the expression of
LwaCas13a was induced successfully (Figure 19).
Figure 19 Results of SDS-PAGE gel analysis of cell lysate from bacteria before and after IPTG induction. Lane1:
Protein marker. Lane2-Lane6: Uninduced cell lysate from bacteria. Lane7-Lane11: IPTG-induced cell lysate from
bacteria.
Bacteria were lysed by ultrasound, and centrifuged to separate supernatant and pellet. The soluble proteins
were present in the supernatant, and the insoluble proteins were in the pellet. Both supernatant and pellet were
analyzed by SDS-PAGE (Figure 20). SDS-PAGE gel analysis showed that the amount of LwaCas13a protein in the
pellet was very little, and most of LwaCas13a protein was in the supernatant, indicating that SUMO tagged
LwaCas13a was soluble.
Figure 20 SDS-PAGE detection of lysed supernatant of cells with lysis pellet of cells. Lane1: Protein marker.
Lane2-Lane6: Lysed supernatant of cells. Lane7-Lane11: Lysis pellet of cells.
Western blot examined uninduced bacteria lysed by RIPA, induced bacteria lysed by RIPA, and induced bacteria
lysed by ultrasonic, and showed that SUMO-LwaCas13a protein was expressed successfully (Figure 21).
Figure 21 Western Blot with LwaCas13a antibody. Lane1: Protein marker. Lane2-Lane3: Uninduced bacteria for
RIPA-lysed. Lane4-Lane5: Induced bacteria for RIPA-lysis. Lane6-Lane7: Induced bacteria for ultrasonic lysis.
The final step of LwaCas13a production was to use His-Tag Protein Purification Kit with IDA-Ni Magnetic Agarose
Beads for purification[10] and SUMO protease ULP1 to
digest SUMO-LwaCas13a. We used SDS-PAGE gel electrophoresis
to analyze the purification results (Figure 22).
Figure 22 Results of purification of LwaCas13a protein. Lane1:Protein marker. Lane2-Lane3: Uninduced bacteria
for RIPA-lysed. Lane4-Lane5: Induced bacteria for RIPA-lysis. Lane6-Lane7: Induced bacteria for ultrasonic
lysis. Lane8-Lane9: Purified LwaCas13a protein.
crRNA
Construction of pSB1C3-crRNA plasmid
crRNA consists of two parts, a repeat part that contains a loop region and a detection part that contains a
linear guide sequence to recognize target RNA[11].
Studies have shown that the length of crRNA guide sequence
affects its activity. When the length of crRNA guide sequence is less than 20nt, its activity is significantly
decreased[12]. We designed four crRNA guide sequences
that were complementary to miR-21-5p, the length of which
ranges from 20nt to 22nt. The secondary structures of these four crRNA were as below (Figure 23).
Figure 23 The secondary structure of crRNA1 (A), crRNA2 (B), crRNA3 (C), crRNA full length (D).
crRNA sequences were cloned into Pveg plasmid with PCR. Most loop region of crRNA was synthesized as an
overhang part of primer loop R, which was introduced into the fragment 1 by PCR with primer loop R and primer
loop F. The rest of loop region and the guide sequence of crRNA was synthesized as an overhang part of primer
spacer F, which was introduced into the fragment 2 by PCR with primer spacer F and primer spacer R. Fragment 1
and 2 were assembled into a plasmid with full length crRNA by Golden Gate assembly. (Figure 24).
Figure 24 cloning of Pveg plasmid with full-length crRNA (A) Schematic presentation of cloning strategy of Pveg
plasmid with full-length crRNA sequences. (B) Primer sequences for cloning of crRNA into Pveg.
PCR was carried out, followed by gel electrophoresis, which demonstrated that fragments containing loop region
and guide sequences were amplified successfully (Figure 25).
Figure 25 Gel electrophoresis of PCR products for Pveg plasmid with full-length crRNA. Lane1: DNA marker. Lane2:
Guide sequence of crRNA1. Lane3: Guide sequence of crRNA2. Lane4: Guide sequence of crRNA3. Lane5: Guide
sequence of crRNA full length. Lane6: Loop region.
The assembled product was transformed into Trans1-T1 competent cells, and then colony PCR was performed to
select the assembled product. PCR product amplified by primer VR and VF2 covered the junction between the
fragment 1 and the fragment 2 (403bp/401bp). Agarose gel electrophoresis showed that the assembly of the two
fragments was successful (Figure 26).
Figure 26 PCR verified that the plasmid has been successfully assembled. （A）Schematic representation of PCR
product.（B） Colony gel electrophoresis after PCR. Lane1: DNA Marker. Lane2-Lane4: crRNA full length.
Lane5-Lane7: crRNA1. Lane8-Lane10: crRNA2. Lane11-Lane13: crRNA3.
In vitro transcription and purification of crRNA
Prior to in vitro transcription of crRNA, Pveg plasmids with full crRNA sequences were linearized by SapI
digestion. The linearized plasmids were used for in vitro transcription using the T7 High Yield RNA Synthesis
Kit from YEASEN Biotechnology Co., Ltd. The reaction system was shown in Table 2, and the mixture was incubate
at 37 °C for 5h to allow the transcription reaction to proceed sufficiently.
| Reagent |
Quantily μL(for 20μL mix) |
| T7 RNA Polymerase Mix |
2μL |
| NTP Mix (100mM) |
8μL |
| T7 Reaction Buffer (10×) |
2μL |
| RNase Inhibitor |
0.5μL |
| Template DNA |
1μg |
| RNase-free water |
to 20μL |
| Total |
20μL |
Table 2 In vitro transcription reaction system.
Transcription products were recovered using the RNA Purification Kit from Tiangen. RK solution and absolute
ethanol were added to the product, and mixed well. The mixture was added to the adsorption column and
centrifuged. The column was washed with deproteinization solution and rinse solution, and elution buffer was
added to obtain purified crRNA. Agarose gel electrophoresis showed that all four crRNAs were successfully
transcripted and purified (Figure 27).
Figure 27 Gel electrophoresis plot of in vitro transcription of crRNA. Lane1: crRNA full length. Lane2:
crRNA1.Lane3: crRNA2. Lane4: crRNA3.
In summary, we have made two essential components of CRISPR/Cas13a system: crRNA and LwaCas13a protein.
Cleavage activity
Since fluorescence reaction has the characteristics of high sensitivity and quantification[13], the cleavage
activity of our CRISPR/Cas13a system was first analyzed by fluorescence detection. We used a reporter RNA
designed based on the principle of fluorescence resonance energy transfer (FRET)[14], with a FAM group at one
end and a black hole quencher group (BHQ1) at the other end, connected by 6 U in between. Activated
CRISPR/Cas13a system could cleave the reporter at 6 U, and release FAM group from the quencher group. As a
result, fluorescence from the released FAM group could be detected. If the crRNA that we designed can form a
functional complex with LwaCas13a, the cleavage activity of this complex could be activated by miR-21-5p, which
could cleave the reporter RNA to release FAM group for fluorescence detection.
Four crRNA named crRNA1, crRNA2, crRNA3 and crRNA full length, were used in the experiment. The results showed
that the CRISPR reaction system assembled by crRNA full length and Cas13a protein had the strongest cleavage
activity, followed by crRNA3, while crRNA2 and crRNA3 had the weakest activity (Figure 28).
Figure 28 Fluorescence intensity measured at different time with different crRNAs.
Next, for the fluorescence reaction, the concentrations of crRNA, LwaCas13a, and reporter RNA were fixed at 20
nM, 100 nM and 100 nM respectively, and the concentration of miR-21-5p was gradually increased from 0 nM to 100
nM. The experimental results showed that the fluorescence intensity increased as the concentration of the miRNA
increased (Figure 29).
Figure 29 Fluorescence intensity measured at different times with different concentrations of miRNA.
In addition, we also set up a series of reporter concentration at 0 nM, 100 nM, 125 nM, 250 nM and 500 nM for
experiments. After we plotted the double inverse of fluorescence intensity and reporter RNA concentration, the
experimental results were found to be consistent with Michaelis-Menten equation[15], which was in line with our
expectations. (Figure 30).
Figure 30 (A) Fluorescence intensity measured at different times with different concentrations of reporter RNA.
(B) The fitting curve of Lineweaver-Burk.
Lateral flow strip
After completing a series of fluorescence experiments that validated the cleavage activity of our CRISPR/Cas13a
system, we began to simplify the method that was used to detect the cleavage activity of the CRISPR/Cas13a
system with lateral flow strips.
To facilitate the lateral flow strip to display the results, the BHQ1 modification of our RNA reporter 3' was
changed to biotin, while the sequence remained unchanged, still 6 U. In the experiment of lateral flow strip
test, we adopted the detection method of "line elimination". On the probe pad of the lateral flow strip, there
is streptavidin-labeled colloidal gold, which binds to the biotin at the RNA reporter. The position of the Test
line (T line) at the lateral flow strip is planted with FAM antibodies, which binds to FAM group of the RNA
reporter. In the case that the RNA reporter is uncleaved, the RNA reporter forms a complex with the
streptavidin-labeled colloidal gold through the interaction between biotin and streptavidin. As this complex
flows to the T line, it is captured by the FAM antibody that binds to the FAM group of the RNA reporter. As a
result, the colloidal gold accumulates at the T line, which can be observed by naked eyes. In addition, the
gold-labeled mouse antibodies are included in the lateral flow strip detection system as controls. They are
captured by the sheep-anti mouse antibodies at the C line position to show the samples are added successfully on
the lateral flow strip.
In the case that miR-21-5p is not expressed, the CRISPR/Cas13a system is not activated, and the RNA reporter is
not cleaved. The uncleaved RNA reporter with streptavidin-labeled colloidal gold is captured by the FAM antibody
at the T-line position, and the gold labeled mouse antibodies is captured by the sheep-anti mouse antibodies at
the C line position. As a result, the lateral flow strip shows both C line and T line, indicating negative
result (Figure 31).
In the case that miR-21-5p is expressed low, the CRISPR/Cas13a system is not fully activated, the reporter is
not completely cleaved. The uncleaved RNA reporter with streptavidin-labeled colloidal gold can still be
captured by the FAM antibody at the T-line position, while the cleaved RNA reporter can not link
streptavidin-labeled colloidal gold with FAM antibody anymore. As a result, the lateral flow strip shows C line
and a weak T line, indicating negative result.
Figure 31 Principle of our lateral flow strips.
Next, the lateral flow strip was tested. The CRISPR/Cas13a system was composed of working solution and sample
solution, and they were mixed for reaction. The composition of the working solution included 10×Cas13a Reaction
Buffer (the recipe was shown in Table 3), RNase Inhibitor, LwaCas13a protein, crRNA-Full length and RNase-free
water, and it has a total volume of 10μL (Table 4). The composition of the sample solution includes miR-21-5p,
reporter and so on, its volume was also 10μL (Table 5). The concentration of miRNA in the reaction system was
changed to test the effects of strips.
| Component |
Concentration |
| Tris-HCl |
10mM |
| KCl |
50nM |
| MgCl2 |
1.5mM |
| pH |
8.3 |
Table 3 Components of Cas13a Reaction Buffer (10×).
| Reagent |
Quantily μL (for 10μL mix) |
| Cas13a Reaction Buffer (10×) |
2μL |
| RNase Inhibitor |
1μL |
| LwaCas13a (2μM) |
1μL |
| crRNA full length (0.4μM) |
1μL |
| RNase-free water |
5μL |
| Total |
10μL |
Table 4 Components of the working solution.
| Reagent |
Quantily μL (for 10μL mix) |
| Cas13a Reaction Buffer (10×) |
2μL |
| RNase Inhibitor |
1μL |
| miR-21-5p (1μM) |
xμL |
| Reporter (2μM) |
2.4μL |
| RNase-free water |
to 10μL |
| Total |
10μL |
Table 5 Components of the sample solution.
After mixing the working solution and the sample solution, the reaction solution was obtained (Figure 32), and
the reaction droplets were added to the sample pad of the test strip, the result was recorded in 3 minutes.
Figure 32 The reaction solution with different concentration of miRNA.
The results of the lateral flow strip detection were as in Figure 33. The C lines of the five test strips were
clearly visible, indicating that these test strips were good to use. In the case that the miRNA concentration in
the reaction mix was 0 nM, T line was obvious. While the miRNA concentration increased to 5 nM, T line was
decreased. A further increase of miRNA concentration to 25 nM made the T line almost disappeared. In the case
that the miRNA concentration increased to 50 and 100 nM, T line completely disappeared. These results
demonstrated that the lateral flow worked as we expected, which could distinguish low and high concentration of
miR-21-5p easily (Figure 33).
Figure 33 High and low concentrations of miRNA were distinguished successfully by lateral flow strip.
Subsequently, the outer shell of the lateral flow strip was manufactured to make it easy to carry and work with
(Figure 34-35).
Figure 34 Diagram of the lateral flow strip and its outer shell.
Figure 35 The outer shell of the lateral flow strip facilitates sample loading and result interpretation.
Future plan
In this year's project, we engineered a CRISPR/Cas13a system to detect miR-21-5p for breast cancer screening.
In the future, we set to make a breast cancer screening kit that can simultaneously detect multiple breast
cancer-associated miRNAs, such as miR-21-5p, miR-3923, miR-885-5p and miR-142-5p, so as to achieve higher
sensitivity and accuracy of breast cancer detection. In addition, we would like to apply our CRISPR/Cas13a
system to screening of other types of cancers.
References
[1] https://www.datacamp.com/tutorial/introduction-to-shap-values-machine-learning-interpretability.
[2] Padroni, L., et al. (2023). "An Epidemiological Systematic Review with Meta-Analysis on Biomarker Role of
Circulating MicroRNAs in Breast Cancer Incidence." Int J Mol Sci 24(4).
[3] Weng, S., et al. (2022). "Highly sensitive and reliable detection of microRNA for clinically disease
surveillance using SERS biosensor integrated with catalytic hairpin assembly amplification technology." Biosens
Bioelectron 208: 114236.
[4] Fan, T., et al. (2018). "Branched rolling circle amplification method for measuring serum circulating
microRNA levels for early breast cancer detection." Cancer Sci 109(9): 2897-2906.
[5] Li, Q., et al. (2018). "Exosomal miR-21-5p derived from gastric cancer promotes peritoneal metastasis via
mesothelial-to-mesenchymal transition." Cell Death Dis 9(9): 854.
[6] Kercher, M. A., et al. (1997). "Lac repressor-operator complex." Curr Opin Struct Biol 7(1): 76-85.
[7] Kuo, D., et al. (2014). "SUMO as a solubility tag and in vivo cleavage of SUMO fusion proteins with Ulp1."
Methods Mol Biol 1177: 71-80.
[8] Tropea, J. E., et al. (2009). "Expression and purification of soluble His(6)-tagged TEV protease." Methods
Mol Biol 498: 297-307.
[9] Matsumura, I. (2022). "Golden Gate assembly of BioBrick-compliant parts using Type II restriction
endonucleases." Biotechniques 72(5): 185-193.
[10] Wang, J., et al. (2021). "Phase separation of OCT4 controls TAD reorganization to promote cell fate
transitions." Cell Stem Cell 28(10): 1868-1883.e1811.
[11] Kellner, M. J., et al. (2019). "SHERLOCK: nucleic acid detection with CRISPR nucleases." Nat Protoc 14(10):
2986-3012.
[12] Abudayyeh, O. O., et al. (2017). "RNA targeting with CRISPR-Cas13." Nature 550(7675): 280-284.
[13] Shan, Y., et al. (2019). "High-Fidelity and Rapid Quantification of miRNA Combining crRNA Programmability
and CRISPR/Cas13a trans-Cleavage Activity." Analytical Chemistry 91(8): 5278-5285.
[14] Jares-Erijman, E. A. and T. M. Jovin (2003). "FRET imaging." Nat Biotechnol 21(11): 1387-1395.
[15] Ramachandran, A. and J. G. Santiago (2021). "CRISPR Enzyme Kinetics for Molecular Diagnostics." Anal Chem
93(20): 7456-7464.
