Genome-wide in vivo screening of circulating tumor cells identified SLIT2 as a regulator of metastasis

summary

Circulating tumor cells (CTC) break free from the primary tumor and spread through the circulatory system to metastatic tumors, which are the leading cause of

Brief introduction

Metastases account for more than 90% of cancer deaths, but are still poorly understood and largely incurable (1–3).

As circulating tumor cells (CTC) leave the primary tumor and seed metastases, the molecular characteristics of CTCs are essential

Here, we report a genome-wide CRISPR gene knockout (KO) screening designed to identify genes for CTC transmission in vivo

outcome

WHOLE GENOME CRISPR-CAS9 KO SCREENING FOR CTCs

Genome-wide CRISPR screening is a powerful tool for the systematic and unbiased identification of transfer-related genes (17, 18, 21).

For sgRNA enrichment analysis, we first analyzed the TKOv3 library depletion

Screening of highly enriched sgRNAs in CTCs

To analyze CTC-rich sgRNAs, we developed two selection criteria for robust candidate hit identification (Figures 2, A, and B).

Use the sub-library screen to functionally verify popular clicks

A customized subcumis crisprko library consists

Loss of layer 2 is associated with poor prognosis and in vivo induction of metastasis in patients with prostate cancer

SLIT2 is a family of secretory proteins that play a known role in axonal guidance through interactions with the cyclic receptor family (30, 31).

According to patient data collected from the Cancer Genome Atlas (TCGA), layer 2 has a significantly lower level of expression in tumor tissue compared to normal prostate tissue (Figure 3D).

Layer 2 KO-PC-3M cells in a prostate cancer model were investigated using CRISPR editing to detect the effects of loss of this gene in vivo (Figure S5A
).
The same amount of NTC or Layer 2 KO PC-3M batteries (1×106) immunocompromised mice are injected
in situ in the prostate.
At the end of tumor growth, the number of CTC present in the whole blood of the injected mice in layer 2 increased by a factor of 5 compared to the group of injected NTC cells (Figure 4A and Figures S5, D, and E).

Compared with the control group, KO cells also showed increased cancer cell proliferation in xenograft layer 2, resulting in a significant increase in tumor size (Figures S5, B, and C
).
Assess whether the KO model at layer 2 is simply because the tumor volume increases, and the number of KO xenograft tumors in layer 2 tumor size comparable to that of the control group decreases the number of KO cells at the time of layer 2 injection (Figure 4B ).
In this case, layer 2 compared to the control group, KO again showed a significant increase in the number of ctc in whole blood (Figure 4B) indicating that the inward migration of cells and dissemination enhanced layer 2 because in this case, there were fewer cells for injection and blood sample collection was earlier, To ensure that the tumor size between the control group and the control group is comparable to that of layer 2 KO cells, the total CTC count is lower
than the previous injection.
In addition, tumor cells are severely infiltrated within the lymph nodes in layer 2 KO models, and cases of submembrane and intravascular violations are found in multiple histological sections stained with hematoxylin and echinacea (H&E) (Figures 4, C, and D, and Figure S5F).

At harvest time, no invasion or surrounding infiltrates
of tumor cells were observed in the control cell-generated prostate cancer model.
Lymph nodes are used as representatives of metastatic models, and based on our previous observations, PC-3M is more likely to develop lymph node metastasis at advanced stages of metastasis than other prostate cancer cell lines such as LNCaP (35).
Our conclusion is that layer 2, in a prostate cancer model, increases the number of CTCs, thereby increasing the burden
of metastasis.

Loss of layer 2 results in enhanced migration and aggressive behavior of prostate cancer cells

Understanding the Layer 2 regulatory transfer process, we study Layer 2 KO and NTC cells using a variety of phenotypic analyses to assess cell proliferation, deformation, and aggressiveness
.
The cultured three-dimensional spheroid layer 2 KO and NTC cells are similar in size and composition, but the tissue structure of the layer 2 KO sphere is not good, and the cells on the outer surface of the spheroid are loosely attached together (Figure 5A as well as Figure S6
).
To assess the functional loss of Layer 2 for the migration potential of PC-3M cells, we designed a special microfluidic device whose migration channel connects the cell loading and nutrient solution storage pools (Figure S7).

The height of the channel is constant at 5 μm, but the width is reduced, so cells are required to remodel in order to squeeze nutrients through the narrow channel (Figure S7
).
A significant number of Layer 2 KO and NTC cells pass through a migration channel width greater than 10 μm, significantly more than other layer 2 KO cells can adapt to channels with a width of 8 μm or less (Figures 5B and S7B).

Invasive behavior layer 2 KO and control cells are also detected
with the Transwell cell invasion test.
More layer 2, compared to the control group, migrates KO cells from the top to the bottom of the membrane, regardless of the presence or absence of a chemical attractant to the lower chamber (Figure 5C).
After the addition of exogenous SLIT2 protein to the number of invasions in layer 2, KO cells resemble control cells (Figure 5C).
Cell deformation and aggressiveness test results were not related to cell proliferation rate, and there was no significant difference in the number of cells in the control and control groups Layer 2 found KO after 3 days of 2D culture (Figure S8).

Losing Layer 2 causes EMT activation

Interrogations using tumor antibody arrays on a group of cancer-associated proteins revealed several robust findings related to KO cells in Layer 2 to EMT (Figures 5D as well as S9).

Taking into account the highest protein family or biological pathway found from the array, mesenchymal marker waveform protein and snail protein levels are in layer 2 KO cells
.
Therefore, we hypothesize that KO produces more mesenchymal phenotypes
.
Decreased expression of epithelial markers (E-cadherin) and increased expression of mesenchymal markers (Vimentin, N-cadherin, and Snail) in Layer 2 KO cells compared to NTC cells suggest that EMT may be associated with cell metastatic phenotype deficiency layer 2 (.
2).
Figure 5E and Figure S10).

During screening, the expression level of epithelial marker EpCAM on the cell surface was also detected, and compared with the control group and the control group, no significant difference was found in layer 2 KO cells (Figure S11).

In a rescue experiment, exogenous SLIT2 was added to the medium, which explored the reversal of the mesenchymal phenotype and analyzed the levels of
waveform proteins.
SLIT2 had no significant effect on waveform protein expression in control PC-3M cells (Figures 5F and S12
).
However, the level of layer 2 KO-PC-3M cells of waveform proteins was initially twice that of control cells, but with the addition of exogenous SLIT2, the number of cells decreased significantly (Figure 5F as well as Figure S12
).
In addition, we also examined waveform protein levels in metastatic breast cancer cell lines (MDA-MB-231) and found that the presence of SLIT2 significantly reduced the expression of waveform proteins (Figure 5F and Figure S12), extending Layer 2 and EMT in different contexts
.

Loss of layer 2 leads to elevated complex I expression, hypersensitivity to rotenone

Elucidating other drivers of aggressive phenotypes in Layer 2 KO cells, we also on NTC and Layer 2 KO cells (Figures 6, A, and B).
In particular, layer 2 KO cells exhibit increased expression of one component of ATP synthase (ATP6 type) as well as multiple subtypes of complex I (MT-ND1 type, MT-ND2 type, and MT-ND5 type) belong to the mitochondrial electron transport chain (ETC), making metabolic processes and oxidative phosphorylation the richest genetic ontological terms (Figure 6B and Figure S13).

The high expression of the ETC component suggests the effect
of layer 2 KO cells on mitochondrial activity and oxidative metabolism.
Confirming this prediction, LAYER 2 was more sensitive to treatment with the complex I inhibitor rotenone compared to control cells (Figure 6C).
To detect whether the increase in complex I activity promotes an increase in the migration potential of PC-3M cells layer 2, we tested the blocking migration capacity of rotenone layer 2 using the KO and control unit of the above microfluidic device (Figure S7).

After rotenone treatment, the number of deformed cells in both groups decreased significantly at a wider channel width (Figure 6D).
However, for narrower channels, rotenone is more effective at inhibiting the migration of cells layer 2 KO cells stronger than NTC cells (Figure 6D).
The sensitivity of two other prostate cancer cell lines, LNCaP and PC3 to rotenone, also enhanced layer 2 knockdown (Figures S14 and S15
).
Rotenone significantly reduces the aggressiveness of LNCaP and PC3 layer 2 KO (Figure S16
).

discuss

Previous observations about layer 2 metastases producing contradictory results in different experimental systems have limited research on prostate cancer models (33, 34, 36–39).
Promoter hypermethylation has previously been found to inhibit Layer 2 associated with aggressive prostate, breast, and lung cancers (33, 34).
A recent study showed that layer 2 is in endothelial cells but inhibits expression due to hypermethylation of the tumor chamber, suggesting that layer 2 acts as a driving force for the release of tumor cells into blood vessels in the tumor microenvironment (33).
As a hit for our genome-wide CTC CRISPR KO screening and sub-library screening, Layer 2 is considered an emphasis node factor
for prostate cancer invasion and metastasis.
Pro-metastatic effect layer 2 In our study, KO was also confirmed
in a mouse model of prostate cancer.
Subcutaneous injection of layer 2 compared to injection of the same number of control cells, KO cells lead to tumor enlargement, while the number of ctCs in whole blood increases
significantly.
However, a KO model of a tumor size similar to that of layer 2 and the control group still resulted in a significant increase in the number of ctCs in the bloodstream, suggesting that it may be layer 2 regarding spread and spread
.
In addition, our study also showed layer 2 compared to a control group, suggesting that KO cells have a potential inhibitory effect Layer 2 intravenously and disseminated
.
However, since metastasis is a very complex process, Layer 2 As for other pathways that lead to an increase in the number of CTCs, including vascular extravasation and cell viability, it can be further studied to fully understand the effects of CTC on Layer 2 during
transfer.

There was no previous pathway associated with the aggressive behavior of prostate cancer cells carrying the loss-of-function mutation Layer 2, except that Layer 2 is regulated by transcription inhibitor enhancers of the zeste homolog 2 (EZH2) (34).
Our EMT and transcriptome studies have shown that a variety of EMT markers and abnormal expression of ETC components in Layer 2 may lead to metastatic KO cells
.
In various studies, mitochondrial activity has been linked to cancer progression (40–44), and complex I inhibitors have been reported as promising anti-tumor therapies (45, 46).
With the validation of two other prostate cancer cell lines, LNCaP and PC-3, our study shows that prostate cancer cells lack Layer 2 This expression increases sensitivity to complex I inhibition, revealing a potential therapeutic strategy layer 2 loss of function mutation
.

Batch sequencing of tumors has uncovered many oncogenic mutations, some of which are being used to develop new targeted therapy drugs
.
Identifying genetic mutations that are functionally associated with tumor progression is critical
to designating new cancer targets.
An important limitation is to assess gene function in rare cell populations, such as CTCs or immune cell subsets in the tumor microenvironment
.
In this study, we performed in vivo genome-wide CRISPR-cas9ko screening
for CTCs of prostate cancer origin in mouse models.
Due to the extreme rarity and fragility of ctCs in patients, making it less feasible to perform genome-wide CRISPR screening without large-scale in vitro CTCs culture, we injected a large number of tumor cells into mice to simulate the production process of CTC and obtain a sufficient number of tumor cells from mouse whole blood for NGS
.
The development of this experimental pipeline is portable and can be easily applied to find unique pro-cancer factors in other types of tumors
.
Once a suitable animal model is established, the platform can also be further applied to explore genetic factors that influence the behavior of immune cells or other rare cell types in the tumor microenvironment
.
Building on this research, we expect that combining effective capture of model rare cells with gene-editing methods will provide new insights into the progression of various diseases, including cancer, and will lead to ideas
for new therapies.

Materials and methods

cell culture

The prostate cancer cell line PC-3M was derived from A.
Allan of the London Academy of Health Sciences, Ontario, Canada, and cultured
in RPMI 1640 (Wisent).
LNCaP and PC-3 were cultured
in RPMI 1640 and Dulbecco Modified Eagle Medium (DMEM)/F12 (Wisent), respectively.
MDA-MB-231 (ATCC) and lenti-x293t cells (Takara Bio) are grown in DMEM (Sigma-Aldrich) with 4-mM stable glutamine (Sigma-Aldrich) added
.
All media were added with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (Wisent) and cells were stored at 37 °C, 5%_CO2

Lentiviral production

All lentiviruses are produced
from leno-x293t cells.
For TKOv3 genome-wide library production, 8.
0 μg of TKOv3 mixed plasmid (Addgene, #90294) (22) was co-transfected into 80% confluent lenti-x293t cells with 4.
8 μg psPAX2 (Addgene, #12260) and 3.
2 μg pMD2.
g (Addgene, #12259
).
Plasmid pSPAX2 and pMD2.
G are gifts
from Didier Trono.
Mix the plasmid (16.
0 μg in 50 μl ddH_)2O) to add 48 μl of X-tremegene9dna transfection reagent (Roche) and pre-culture in 750 μl optim (Gibco) for 5 min
.
The mixture of plasmid transfection reagent-optim is cultured further for 25 min and then driped into lenti-X cells
.
After 18 h, the medium is replaced with fresh DMEM
.
Virus particles are collected 48 h after media exchange, snap-frozen in liquid nitrogen, and stored in ? 80 degrees Celsius
.
Customized sub-libraries (plasmid pools synthesized by GenScript) lentiviruses are prepared
using the same method as TKOv3 libraries.
Individual layer 2 CRISPR-KO lentivirus is produced
according to the polyenediamine transfection protocol (Addgene).
The sgRNA sequences for layer 2 genes that were cloned to the lentil protein rv2 (Addgene, #52961)(47), 20-base pairs (bp) were cggttggtcgtactca for KO1 and acggaaagcttttcttggga
for KO2 。 For NTC, the sgRNA sequence used is cccgaattctatcgtgcggg
.

Lentiviral transduction and infection multiplicity assay

Lentiviral titration toPC-3M cells on 2×1056-well plates, 2 ml of complete RPMI 1640 supplemented polyblen (8 μg/ml; Sigma-Aldrich）
。 After overnight infection, the medium was replaced with puromycin supplementation (2 μg/ml; Sigma-Aldrich' fresh RPMI
.
After antibiotic screening for 48 h, the multiplicity (MOI)
of infection is determined by comparing the number of cells retained at a specific viral load, using 0.
1% crystal violet staining (Sigma-Aldrich) and absorbance readings at 590 nm.
Import the CRISPR library into PC-3M cells
with a MOI of 0.
3.

Genome-wide CTC-CRISPR-KO screening

PC-3M batteries (1.
0×108) were converted
with polybutadiene as a transduction reagent with TKOv3 libraries (71090 sgRNAs) at a MOI of 0.
3.
After two rounds of purinemycin screening, 3×¹⁰⁷⁶ male 6-8 weeks old (1.
5×10) male nude mice (nu/nu) were injected subcutaneously into transgenic cells7 per mouse side).

Three other mice were injected with unmodified parent PC-3M cells for tumor size comparison
.
Tumor volume is measured twice a week with calipers and 2/2 is calculated with the length × width formula 2/2.
Mice are sacrificed 3 weeks after transplantation, tumor and whole blood samples (cardiac puncture)
are collected.
All animal work is carried out
in accordance with the guidelines of the Canadian Animal Care Board and in accordance with the protocol approved by the Animal Care Committee of the University of Toronto.

Design and manufacture of prism chips

The initial iteration of the prism chip (24) was modified to increase throughput to 3×107 cells per hour per chip and to divide the cells into zero/low expression, medium expression, and high expression
.
Magnetically labeled cells undergo a lateral shift
based on the expression level.
A guide rail made of a high permeability alloy is used so that it flows at an angle with the fluid (medium is 5°, height is 20°
).
In the case where the permanent magnet generates a magnetic field, a local region with a high magnetic flux appears at the guide edge, causing the attraction of magnetic nanoparticles (48).
The balance of magnetic force and Stokes resistance from fluid flow causes the magnetically marked battery to follow the deflection rail until the angle changes or the end of the rail is on the side
of the device.
Changes in the angle of the guide cause the resistance component to increase and the media expression elements to separate
.
The chip is manufactured on a soda lime glass wafer
.
Use lithography and wet etching to graphically the guides and create channels
using SU-8 photoresist (Kayaku-AM).
Polydimethylsiloxane (PDMS) plates (Dow-Sylgard) are bonded together to package the device according to existing protocols (49).
Simultaneous setup of multiple chips simultaneously isolates CTCs from multiple blood samples
.
To prepare prism chips for CTC isolation, overnight under constant hydraulic pressure with sterile water containing 0.
1% Pluronic F68 (Sigma-Aldrich) before use
.
The nonionic surfactant Pluric F68 prevents non-specific capture of cells in the microchannel, with constant hydraulics removing small bubbles
.
For immunomagnetic-based cell sorting, the prism chip is first mounted on a permanent magnet (BY088-N52, K&J Magnetics
).
Before sample loading, hanks equilibrate the salt solution (HBSS; GIBCO) and 2% bovine serum albumin (BSA; A buffer solution consisting of Sigma-Aldrich) and 5 mM EDTA (BioShop) is introduced into the prism chip to remove pruriconic water
.

CTC collection and cell lysis

To isolate CTCs, each blood sample is diluted with HBSS buffer in a 1:1 ratio and shaken with 10% EpCAM microspheres (Miltenyi Biotec) for 1 h
at room temperature.
Subsequently, an equal amount of buffer solution and a blood sample containing magnetically labeled CTC are loaded into the buffer inlet and sample inlet, respectively, and then removed
from the chip at a flow rate of 2 ml/h.
After sample handling, load 300 μl of buffer solution into the sample and buffer inlet and extract at the same flow rate to rinse the remaining samples
in the microchannel.
Inject 1.
5 ml of centrifugal microtubes (Axygen) into the solution collected from the high expression outlet (~500 μl) of each chip and leave on a DynaMag-2 magnet (Thermo Fisher Scientific) for one h
.
After the cell pellet, slowly pipette up the clear liquid with a pipette, without touching the bottom or wall of the tube, leaving only the minimum solution containing CTCs (5 to 10 μl).

CTCs were lysed by alkali lysis (25).
Simply put, add 10 μl of 200 mM KOH and 50 mM dithiothreitol to the CTC solution and heat at 65 °C for 10 min
in a thermal circulator.
The lysis products
were then neutralized with 10 μl of 900 mM tris HCl (pH 8.
3) with 300 mM KCl and 200 mM HCl.

Genomic DNA extraction and PCR amplification of sgRNA

Extract gDNA from the CRISPR library - transform cell pools (3×10) ⁷ to 5 ×107 cells) and tumor tissue (100 to 200 mg) using the QIAamp DNA Blood Sample Analysis Kit (Qiagen) as per the manufacturer's instructions
。 For extracting gDNA from tumor tissue, cut the tumor into small pieces using a razor blade and incubate overnight with 1.
8-ml tissue lysis buffer ATL (Qiagen) and 0.
2 ml protease K (Qiagen) at 56 °C, and then treat
with the QIAamp DNA blood sample kit.
GDNA was extracted from the initial cell pool and primary tumor samples examined in 230:280 and 260:280 ratios before being sent to the Princess Margaret Genome Center (PMGC; Toronto, Ontario, Canada) to perform PCR amplification of the sgRNA region and generation
of barcode sgRNA libraries.
For CTC samples, there is no need to extract gDNA, and 10 μl of neutralizing lysate is extracted directly from each CTC sample for PCR amplification
of the sgRNA region.
PcR settings per 100 μl are as follows: 10 μl of 10× buffer (provided by Ex-Taq), 8 μl of deoxynucleotide triphosphate (provided by Ex-Taq), 5 μl of Amp1 primer mixture (10 μM), 10 μl of neutralizing CTC lysate, 65.
5 μl of water, and 1.
5 μl of Ex-Taq DNA polymerase (Takara Bio).

The parameters of the thermal cycle adopted are as follows: Step 1: 95 °C, 1 min; Step 2: 95 °C, 30 sec; Step 3: 52 °C, 30 sec; Step 4: 72 °C for 20 sec (repeat steps 2 to 4 34 times); Step 5: 72 °C, 2 min; Step 6: Hold
at 4 °C.
The PCR product is run on a 1% agarose gel to confirm the size of the product (expected size: 235 bp) and then sent to the PMGC for barcode sgRNA library generation
.
NGS testing was performed using Illumina NovaSeq 6000, with 1 million reads
per sample.

CTC subpool CRISPR KO screen

A sub-library (400sgrnas) of the TKOv3 whole genome library (lentiCRISPRv2) was synthesized and cloned with GenScript and cloned onto the
same vector.
The cloned sub-libraries are deep sequenced to verify the expression
of sgRNA in the plasmid pool.
Lentiviral production, MOI assays, and in vivo screening are performed in the same way as TKOv3 libraries, except that GPI-anchored lentiviruses are introduced into PC-3M cells to display Myc
on the cell surface before subbanking is introduced into cells.
Thus, ctCs in blood samples are labeled for 1 h with 1% biotinylated anti-Myc antibody (Abcam, ab197139) and 5% biotin microspheres (Miltenyi Biotech) for 1 h and then isolated
on a prism chip.
CTCs are lysed and sgRNA-PCR amplified in sub-library screening using the same method
as genome-wide screening.

Analysis of NGS data and sgRNA enrichment

Genomic widescreen data is plotted by PMGC, and custom sub-library data is plotted in accordance with the MAGeCK pipeline (50).
The number of reads in each sample is normalized
by converting the original sgRNA count to reads in parts per million (rpm).
The rpm value is then recorded down ₂ downstream analysis of the conversion
.
A non-negative matrix decomposition (NMF) R packet was used to generate a correlation heat map (51), and the Pearson correlation 2-rpm count
between samples was calculated with log.
The empirical cumulative distribution map was generated
using the Python libraries Pandas, Seaborn, and Matplotlib.
Biogenetic enrichment analysis using Metascape (52).
For sgRNA enrichment analysis, fold changes were obtained using MAGeCK and the first 100 candidate genes were identified from CTC-sgRNAs-based genome-wide screening using two criteria: (i) count: If in the standardization of any CTC sample (100 sgRNAs×6 replicates), one sgRNA is counted as the first 100 Due to the exclusion of overlapping sgRNAs in 6 repeat sequences, 162 unique sgRNAs instead of 600 were resulted; (ii) Replication: If CTC sgRNA is enriched (fold change ≥1.
0) in the corresponding primary tumor sample compared to its counterpart (n=6).

For sub-library data, sgRNA counts were normalized and enriched using DugZ analysis (53).

Tap In Vivo Verification at the top of the screen

PC-3M NTC or layer 2 KO cells (1.
0×106 each) are injected into the prostate
by the right dorsolateral lobe of 6-8 week-old male nude mice (nu/nu).
After 3 weeks, the sacrificed mice collect the primary tumor (size comparison), blood (CTC count), and lymph nodes (metastatic histology
).
Mice are collected from peripheral lymph nodes, including 2 inguinal lymph nodes, 2 axillary lymph nodes, and 2 arm lymph nodes
.
Histologically examined lymph nodes are formalin fixed, paraffin-embedded, sliced, H&E stained (prepared by the University Health Network Pathology Research Project laboratory
).
Prepared tissue slices are imaged and examined by pathologists in Charles River laboratories to evaluate and grade metastatic lesions
.
Lymph node metastases are divided into three groups: no tumor cells are seen in the inner or outer tissues of the lymph nodes (none); Tumor cells are observed in the lymph nodes or in the peripheral tissues (partially); Tumor cells are visible in the lymph nodes and account for more than half of the cross-sectional area (severe).

Development and performance of eight-zone CTC capture device

As mentioned earlier, eight-zone profiling chips (54–56) were fabricated using 3D stereolithography and standard soft lithography Before use, the microfluidic chip was processed overnight
with a 0.
1% Prurian F-68.
Place a rectangular array of two N52 neodymium magnets up and down connected to the syringe pump (Chemyx) and then extract
during sample processing.
Blood samples are prepared in the same way as prism chips, then loaded onto an octopus device, where CTC is collected and analyzed
for EpCAM expression.
Extract phosphate-buffered saline (PBS; 500 μl), replace Pruonic F-68 solution
.
Treat samples
at a flow rate of 750 μl/h.
The sample is then immobilized in PBS with 150 μl of 4% paraformaldehyde (Sigma-Aldrich) and osmotic in
PBS with 150 μl of 0.
2% Triton X-100 (Sigma-Aldrich).
Wash the chips
with 125 μl cliniMACS PBS-EDTA solution (Milton i) between each step.
Captured cells with AF488 containing anti-pancellular keratin (30 μg/ml; Abcam, ab277270), anticellular keratin 18-isothiocyanate fluorescein (6 μg/ml; LifeSpan BioSciences, LS-C46335), anticellular keratin 19 AF488 (30 μg/ml; BioLegend, 628508) antibody mixture for immunostaining, and anti-mouse CD45 fluorophore phycocyanin (3 μg/ml; BD Bioscience, 559864), add 1% BSA and 0.
1% Tween 20 (Sigma-Aldrich) at a flow rate of 200 μl/h
.
The captured cells were then stained with 4′,6-diamino-2-phenylindole (R37606) and then fluorescence imaging and scanning
with a Nikon Ti-E eclipse microscope with Anorran Neo-sCMOS camera.
CTCs were identified as nuclearly present and EpCAM⁺CK⁺CD45 type^? Cells
.

Spherical analysis

Ko cells made of NTC or layer 2 are cultured according to the method previously described (57).
Simply put, RPMI 1640 adds 0.
75% methylcellulose (Sigma-Aldrich) and cultures spheroids
in pretreated ultra-low attachment plates (Corning).
To determine the number of cells containing spheroids in each well, follow the manufacturer's instructions using the CellTiter Glo 3D cell viability assay (Promega
).
Briefly, after placing the plate at room temperature, mix
vigorously with the cell culture medium in a 1:1 ratio.
After 25 min incubation at room temperature, record the glow
of each well.

Migration chip preparation and cell migration/deformation analysis

The migration chip consists of a sample loading channel in the middle and two stimulus loading channels
on both sides.
There are many narrow migration channels (5 μm constant height) ranging in width from 6 μm to 20 μm, connecting sample channels and stimulation channels
for cell migration.
The chip consists
of a PDMS substrate with microchannels and a glass cover.
PDMS substrates were prepared using soft lithography
.
Simply put, mix the liquid base and reagents of the PDMS in a 10:1 ratio, cast on a silicon mold with a SU8 negative photoresist pattern, and then incubate at 70 °C for 2 h for polymerization
.
After plasma treatment, the cured PDMS is stripped off the mold and bonded
to the glass cap.
In cell deformability assays, the migration chip is first filled with 70% ethanol and then treated with ultraviolet light
.
The degassed and sterilized chip is washed with PBS and then with cell culture medium
.
PC-3M NTC (3×106) or Layer 2 load KO cells from 500 μl of complete medium into the sample channel and incubate overnight at 37 °C to allow cells to attach
.
The next day, wash the sample channel with PBS and load the serum-free medium
.
Fresh cell culture medium containing 10% FBS is injected within the stimulation channel to create a nutrient gradient
along the narrow migration channel.
After 6 h of incubation, fluorescence labeling is performed by loading 5 μM SYTO24 (Invitrogen) into samples and stimulation channels and incubated for 15 min
.
Observe the migration behavior
of cells under a fluorescence microscope.
Cells that enter the migration channel or reach the stimulation channel are considered deformed and migratory cells
.
Calculate the number of migrated cells and compare the deformation capacity and migration potential
of different sample cells.
To test the effect of rotenone on cell deformation capacity and migration, rotenone (IC20) is added to the medium loaded onto the chip at an inhibitory concentration of
20%.

Transwell cell invasion assay

To prepare the invasive laboratory, add 750 μl of 10% FBS-supplemented RPMI 1640 to each of the 24 wells of the Transwell plate (Corning) and add 100 μl of matrix gel (Corning) to the cell culture insert (8 μm).

Petri dishes are cultured in a humidified incubator at 37 °C with 5% CO _{for 2} h
.
KO cells from serum-free RPMI 1640 are added to the matrix gel-coated insert at 500 μl pc-3M NTC or layer 2 at a ratio of 1×10 to the matrix gel-coated insert ⁵ cells/ml
.
Depending on the conditions tested, the medium in the well contains or does not contain 10% FBS and secretes N-SLIT2 (1 μg/ml; R&D System, 8616-SL-050).

Cells are incubated overnight
in a matrix gel-coated insert in an incubator.
The next day, remove the medium
from the insert and well.
Remove the non-migrating cells on the upper side of the membrane with a cotton swab and rinse
with PBS.
Fix the cells
remaining on the membrane by immersing the membrane in wells filled with cold methanol for 20 min and air-drying for 30 min.
The membrane is then immersed in 750 μl of 0.
1% crystal violet (Sigma-Aldrich) for 30 min at room temperature to stain the cells and then wash thoroughly with distilled water
.
After air drying, immerse the membrane in 750 μl of 10% acetic acid (VWR) and shake well until the stain is completely dissolved
.
The optical density
of dissolved stains in each well is determined with a microplate reader at 590 nm.

Tumor proteomic antibody array

Tumor antibody arrays (R&D systems) are performed according to the manufacturer's instructions
.
Simply put, NTC or layer 2 KO cell lysate is diluted and incubated overnight with a nitrocellulose membrane that has been preincapsulated in two parts with antibodies against 84 tumor-associated proteins
.
Wash the membrane before incubation with a mixture of biotinylated detection antibodies, and then add streptavidin-horseradish peroxidase
.
The amount of protein bound at each capture point is detected with a chemiluminescent substrate and the signal
is captured with a chemiluminescent imaging system (Bio-Rad).
The intensity of the snap point is quantified
with ImageJ.

Western blot analysis

Whole cell extracts
were prepared with radioimmunoprecipitation assay buffer (Thermo Fisher Scientific) and protease inhibitor cocktail (Sigma-Aldrich).
Proteins are separated on 4 to 15% precast protein gels (Bio-Rad) and transferred to polyvinylidene difluoroethylene
membranes.
The main antibodies used are anti-SLIT2 (Abcam, ab134166), anti-waveform protein (Abcam, ab92547), anti-N-calocrin (BD Bioscience, 610920), anti-E-cadherin (R&D system, AF648) and anti-snail (NEB, 3879).

Secondary antibodies are purchased from cell signaling
technology.
Use Thermo Fisher Scientific to visualize
blots using a chemiluminescent imaging system.
The strength of the protein bands is quantified
with ImageJ.

RNA sequencing and data analysis

Follow the manufacturer's instructions to extract the RNA using the RNeasy Plus kit (Qiagen) and submit it to the PMGC for complementary DNA library preparation and RNA sequencing
using Illumina NovaSeq 6000.
RNA sequencing data is processed using
the *STAR program.
The data was generated
using standardized read counts from R and RStudio (R Project, Revolution Analytics), Metascape, and gene enrichment analysis tools (Broad Institute).

Cell viability assay

NTC or layer 2 KO-PC-3M cells are seeded on 96-well plates at a density of 2×10, triplicate ⁴ number of
cells per well.
The next day, rotenone is titrated
in a 96-well plate at a concentration of 0 to 10 μM in a double serial dilution.
After 48 h, follow the manufacturer's instructions to determine the number of
viable cells using cell counting kit 8 (Dojindo).

Clinical and patient data analysis

RNA sequencing and clinical data were retrieved
from portals of TCGAs, human protein atlases, GEPIA2, and COSMIC.
The median layer 2 expression is 2.
45 fragments/thousand base million (FPKM) as the threshold defined for low or high expression in patients
with layer 2 prostate cancer (prostate cancer).
Layer 2 detects expression levels in tumors and normal tissues of 31 tumor types to assess folding changes
.
We retrieved functional mutation loss information from the Cosmic Cancer Mutation Survey (CMC) and compared it with the well-known levels of loss-of-function mutations in TSGs and OGs (29).

Data collection and analysis

The counts of sgRNAs were collected by toronto PMGC and analyzed
using the aforementioned MAGeCK or DugZ.
Screen data quality control analysis and graphical preparation are carried out using the software package specified in the corresponding section above
.
Microscopic images were obtained with zeiss Axio Observer and made
with ZEISS Zen blue.
Protein band strength is analyzed and quantified
with ImageJ.
Biopath analysis
using Metascape.
Use statistical tests to analyze the data and specify the sample size
in each chart legend.
Use Graphpad Prism v7.
0 for statistical testing and graphical representation
of data.
All data are expressed as averages, with SD represented
by error bars.
The statistical significance is set to P<0.
05 (double-sided test
).