ICAM2 initiates trans-blood-CSF barrier migration and stemness properties in leptomeningeal metastasis of triple-negative breast cancer

Leptomeningeal metastasis (LM) occurs when tumor cells spread to the leptomeningeal space surrounding the brain and the spinal cord, thereby causing poor clinical outcomes. The triple-negative breast cancer (TNBC) has been associated with symptoms of LM and mechanism remained unclear. Through proteomic analysis, we identified high expression of ICAM2 in leptomeningeal metastatic TNBC cells, which promoted the colonization of the spinal cord and resulted in poor survival in vivo. Two-way demonstration indicated that high levels of ICAM2 promoted blood–cerebrospinal fluid barrier (BCB) adhesion, trans-BCB migration, and stemness abilities and determined the specificity of LM in vivo. Furthermore, pull-down and antibody neutralizing assay revealed that ICAM2 determined the specificity of LM through interactions with ICAM1 in the choroid plexus epithelial cells. Therefore, neutralizing ICAM2 can attenuate the progression of LM and prolong survival in vivo. The results suggested that targeting ICAM2 is a potential therapeutic strategy for LM in TNBC.

Following washing with 1X PBS. Secondary antibody conjugated with fluorescence dye was used to hybrid with primary antibodies for fluorescence detection.
The LeptoM3, two independent ICAM2 shRNA knock-down (sh#1 and sh#4) LeptoM3, BrM3, MDA-MB-231, ICAM2 overexpressing MDA-MB-231, and Choroid Plexus Epithelial cells were seeded at a density of 1×10 5 in cover slides per well. The IF staining was performed by primary antibodies including the anti-ICAM1 polyclonal (A19300, ABclonal) and anti-ICAM2 monoclonal (#13355, Cell Signaling) antibodies incubated at 4℃ overnight. Following washing with 1X PBS. Secondary antibody conjugated with fluorescence dye was used to hybrid with primary antibodies for fluorescence detection.

BCB adhesion assay
Poly-L-lysine coated wells were seeded with human Choroid Plexus Epithelial cells (#1310, Science Cell) and maintained in a growth medium (#4101, Science Cell).
One day later, Cancer cells were seeded on the top of the Choroid Plexus Epithelial cells monolayer and adhesion for 1-10 mins. Adhesive cell numbers were counted by the luciferase activity of cells.
Poly-L-lysine coated wells were seeded with human Choroid Plexus Epithelial cells (#1310, Science Cell) and maintained in a growth medium (#4101, Science Cell).

Brain slices preparation and brain slices adhesion
The removal of the brain tissue from the mice following subject for the Rodent Brain Matrix (RBM-2000C), which was designed to aid in the free-hand dissection of the brain with 1.0 mm thickness for each slice. The physiological conditions of the brain slices were maintained in the DMEM/F12 culture medium for the brain slices adhesion study. 1×10 5 MDA-MB-231, BrM3, and LeptoM3 cells were seeded on the brain slices and cell adhesion was tested for 60 mins. Following, 1XHBSS was used to remove the non-adhesive tumor cells. IVIS and IHC staining were used to investigate the adhesive cell numbers and location of adhesive cells in brain slices ex vivo.

Sphere forming assay
Ultra-low 96 well plates were prepared and the cells were seeded at a density of 1×10 4 per well in the DMEM/F12 supplemented with 1% N2 supplement, 20 ng/mL epidermal growth factor (EGF), and 20 ng/mL basic fibroblast growth factor (bFGF).
The plate was incubated at 37℃ in the incubator for 10 days to allow sphere formation, and the sphere was counted by bright field microscopy.
Ultra-low 96 well plates were prepared and the LeptoM3 cells were seeded at a density of 1×10 3 or 5×10 3 per well in the DMEM/F12 supplemented with 1% N2 supplement, 20 ng/mL EGF, and 20 ng/mL bFGF. Following, 5 µg/mL isotype control IgG or Goat anti-ICAM2 polyclonal (AF244, R&D) antibodies were treated in cultured medium every two days. The plate was incubated at 37℃ in the incubator for 10 days to allow sphere formation.

Trans-BCB migration assay
Poly-L-lysine coated 6.5 mm transwell supports (Corning Costar) with 3.0mm pore size were seeded with human Choroid Plexus Epithelial cells (#1310, Science Cell) and maintained in a growth medium (#4101, Science Cell). Three days later, the permeability of the barriers was tested with sodium fluorescein dye (518-47-8, Merck) (0.01%) diluted with a culture medium and incubated for 15 and 30 minutes at 37℃.
Medium in the bottom chamber was collected, and absorbance was measured by Elisa reader, the barrier was deemed ready for assays. Cancer cells were seeded on the top of the chamber, were coated with human Choroid Plexus Epithelial cells for 3 days, and cultured at 37℃ for 24 hours. The trans-migrated cells were attracted by DMEM+1% CCS and luciferase activity of cells counted migrated cell numbers.

Invasion assays
Trans-well (8um pore size, BD Biosciences) was performed on invasion assay.
Trans-wells were placed in 24 well plates and each well filled with 1mL of DMEM medium containing 1% CCS and 1% P/S. For invasion assay, each top of chamber was coated the Matrigel (1mg/mL, BD Biosciences) for 3 hours at 37℃. Cancer cells were seeded on the top of chamber and cultured at 37℃ for 18 hours. subsequently, top of chamber was removed from the plate and washed with 1X HBSS three times. Lower side of the filter membrane was fixed using methanol for 20 minutes and stained with 10% Giemsa solution (Sigma-Aldrich, St. Louis, MO) for 10 minutes. Finally, the Lower side of the filter membrane was washed with distilled water three times and cells were counted in bright-field microscopy.

Gel assisted digestion
The protein solutions were mixed with SDS sample buffer and eluted at 95°C for 10 min. The proteins were loaded and analyzed by 10% SDS-PAGE (1 cm). The excised gel was first de-stained, and then reduced with 10 mM dithiothreitol (DTT, Merck) at 60°C for 45 min, followed by cysteine-blocking with 55 mM iodoacetamide (IAM, Sigma) at 25°C for 30 min. Samples were digested with sequencing-grade modified porcine trypsin (Promega) at 37°C for 16 hours. The peptides were then extracted from gel, dried by vacuum centrifugation, and reconstituted with 0.5% Formic acid before analyzing.

Protein identification.
The data analysis was carried out using Proteome Discoverer software (version 1.4, Thermo Fisher Scientific). The MS/MS spectra were searched against the SwissProt database using the Mascot search engine (Matrix Science, London, UK; version 2.5). For peptide identification, 10 ppm mass tolerance was permitted for intact peptide masses, and 0.5 Da for CID fragment ions with allowance for two missed cleavages made from the trypsin digestion: oxidized methionine and acetyl (protein Nterminal) as variable modifications; carbamidomethyl (cysteine) as static modification.
Peptide-spectrum match (PSM) were then filtered based on high confidence and Mascot search engine rank 1 of peptide identification to ensure an overall false discovery rate below 0.01. Proteins with single peptide hit were kept.

Extraction of Membrane and Cytosolic fractions
The protocol is according to the Mem-PER™ Plus Membrane Protein Extraction Kit (#89842, Thermo Fisher). Resuspend 5 × 10 6 cells in the growth media by scraping the cells off the surface of the plate with a cell scraper. Centrifuge harvested cell suspension at 300 × g for 5 minutes. Wash cell pellet with 3mL of Cell Wash Solution

LC-MS/MS analysis
The digested peptides were diluted in HPLC buffer A (0.1% formic acid) and loaded onto a reverse-phase column (Zorbax 300SB-C18, 0.3 × 5 mm; Agilent Technologies). The desalted peptides were then separated on a homemade column 3,000 ions (MS/MS) for maximum accumulated time or ions, respectively. All datasets will be publicly available for querying and downloading through the GEO database when this paper is accepted.

Pull down assay
Purified ICAM2-His(6x) proteins were used to incubate with membrane fractions of Choroid Plexus epithelial cells at 4°C overnight. Following, a mixture or only membrane fractions of Choroid Plexus epithelial cells were incubated with Ni-beads (#30210, QIAGEN) at 4°C for 1 hour. Following, washing the mixture with 10mM imidazole two times. Finally, the 500mM imidazole will be used to eluate the ICAM2-His(6x) proteins and their interacting partners from Ni-beads for further investigation.

Xenograft animal model
The leptomeningeal metastasis mice model of NOD/SCID female age-matched mice aged 4-6 weeks were obtained from Laboratory Animal Center (National Cheng Kung University) and randomly used for xenograft studies. TNBCs were injected into the left ventricle of the heart of NOD/SCID mice through intracardiac injection. The tumor metastasis and growth were detected by IVIS measurement weekly. After 4 weeks of IC injection, the IVIS combined with Micro-CT was used to detect the specificity of spinal cord metastasis in mice (IVIS signal indicating the distribution of tumor cells and Micro-CT image indicating the location of signal). Animals were sacrificed at 7-10 weeks and then the organ can be collected for ex vivo detection.
The MDA-MB-231, BrM3, and LeptoM3 cells were IC injected into the NOD/SCID mice and metastasis was detected by the IVIS system every week. All criteria of euthanasia were in accordance with animal welfare regulations. First, according to the spirit of 3R, all the groups were sacrificed by CO2 euthanasia at endpoint, because our results demonstrated that the occurrence of leptomeningeal metastasis in mice was detected by IVIS in vivo and ex vivo at the 7th week after IC injection. Second, we sacrificed mice with severe metastasis including brain, lung, and spinal cord (photon flux signal higher than 1x10 7 ) accompanied with mobility, and significant weight loss issues, which is evaluated and informed by a professional veterinarian, by CO2 euthanasia before the 7 weeks.
For evaluating LM in normal background mice, 1×10 5 or 1×10 4 ICAM2 overexpressing 4T1 cells and vector control cells carries the luciferase were IC injected into the BALB/c mice and distribution of cancer cells were monitored by IVIS every two to three days. According to the spirit of 3R, we will sacrifice mice with sever metastasis including brain, lung, and spinal cord (photon flux signal higher than 1x10 7 ) accompanied with mobility issues by CO2 Euthanasia at the human endpoint every day.
All mice were sacrificed on the humanitarian end point.
The 1 × 10 5 LeptoM3 cells, which were preincubated with 5 µg/mL IgG control or ICAM2 antibodies (AF244, R&D), following injecting into the NOD/SCID mice (n=6/each group). Then, 5 µg/mL ICAM2 antibodies (AF244, R&D), which dissociated with 1xHBSS, were IC administered every 2 days following metastasis were monitored by IVIS every week. After 7 weeks, all mice were sacrificed and metastatic lesions in organ were detected by IVIS system. All criteria of euthanasia were in accordance with animal welfare regulations. First, according to the spirit of 3R, all the groups were sacrificed by CO2 euthanasia at endpoint, because our results demonstrated that the occurrence of leptomeningeal metastasis in mice was detected by IVIS in vivo and ex vivo at the 7th week after IC injection. Second, we sacrificed mice with severe metastasis including brain, lung, and spinal cord (photon flux signal higher than 1x10 7 ) accompanied with mobility, and significant weight loss issues, which is evaluated and informed by a professional veterinarian, by CO2 euthanasia before the 7 weeks.
The 1 × 10 3 LeptoM3 cells, which were preincubated with 5 µg/mL IgG control or ICAM2 antibodies (AF244, R&D), were injected into the NOD/SCID mice (n=6/IgG control group, n=5/ICAM2 antibodies treated group) following metastasis was monitored by IVIS every week. After 12 weeks, all mice were sacrificed and metastatic lesions in organ were detected by IVIS system. All criteria of euthanasia were in accordance with animal welfare regulations. First, according to the spirit of 3R, all the groups were sacrificed by CO2 euthanasia at endpoint, because our results demonstrated that the occurrence of leptomeningeal metastasis in mice was detected by IVIS in vivo and ex vivo at the 12th week after IC injection. Second, we sacrificed mice with severe metastasis including brain, lung, and spinal cord (photon flux signal higher than 1x10 7 ) accompanied with mobility, and significant weight loss issues, which is evaluated and informed by a professional veterinarian, by CO2 euthanasia before the 12 weeks.

Western Blotting
Cell lysates are loaded on SDS-polyacrylamide gel for electrophoresis and transferred to PVDF membranes. Protein expressions are examined after incubation with primary antibodies (Supplementary Table S1) followed by HRP-conjugated secondary antibodies and detected by X-ray film.