Human liver sinusoidal endothelial cells promote intracellular crawling of lymphocytes during recruitment: A new step in migration

The recruitment of lymphocytes via the hepatic sinusoidal channels and positioning within liver tissue is a critical event in the development and persistence of chronic inflammatory liver diseases. The hepatic sinusoid is a unique vascular bed lined by hepatic sinusoidal endothelial cells (HSECs), a functionally and phenotypically distinct subpopulation of endothelial cells. Using flow‐based adhesion assays to study the migration of lymphocytes across primary human HSECs, we found that lymphocytes enter into HSECs, confirmed by electron microscopy demonstrating clear intracellular localization of lymphocytes in vitro and by studies in human liver tissues. Stimulation by interferon‐γ increased intracellular localization of lymphocytes within HSECs. Furthermore, using confocal imaging and time‐lapse recordings, we demonstrated “intracellular crawling” of lymphocytes entering into one endothelial cell from another. This required the expression of intracellular adhesion molecule‐1 and stabilin‐1 and was facilitated by the junctional complexes between HSECs. Conclusion: Lymphocyte migration is facilitated by the unique structure of HSECs. Intracellular crawling may contribute to optimal lymphocyte positioning in liver tissue during chronic hepatitis. (Hepatology 2017;65:294‐309).


Movie 2
Time lapse recordings of endothelial monolayer during lymphocyte migration across TNFα stimulated HSEC under shear stress. Same sequence as in Movie 1 with the blue (lymphocyte) signal omitted. Note the endothelial cytoplasm is undisturbed during lymphocyte crawling and migration.

Movie 3
Time lapse recordings of lymphocyte migration across TNFα and IFNγ stimulated HSEC monolayer under shear stress. Lymphocytes were prelabelled with CellTracker BMQC (blue) and HSEC prelabelled with CellTracker CFMDA (green). Lymphocytes (blue) can be seen crawling below and above the endothelial surface (green) but also migrating intracellularly. The video is a representative recording from 3 separate experiments with different HSEC.

Movie 4
Time lapse recordings of endothelial monolayer during lymphocyte migration across TNFα and IFNγ stimulated HSEC under shear stress. Same sequence as in Movie 3 with the blue (lymphocyte) signal omitted. Note the redistribution of the endothelial cytoplasm (green) during lymphocyte crawling and migration.

Movie 5 Time lapse recordings of lymphocyte migration across TNFα and IFNγ stimulated HSEC monolayer under shear stress demonstrating 'intracellular crawling'.
Lymphocytes were prelabelled with CellTracker BMQC (orange) and HSEC prelabelled with CellTracker CMFDA (green). White border box demonstrates lymphocytes redistributing endothelial cytoplasm during crawling from one to endothelial cell to another. Intracellular crawling of lymphocytes across HSEC occurs in the presence of a chemokine gradient. (A) 3D reconstruction of an orthogonal (XZ) projection performed during live-cell imaging of peripheral blood lymphocytes migrating across TNFα and IFNγ stimulated HSEC undershear stress. HSEC were pre-labelled with CellTracker CFMDA (green) and lymphocytes with CellTracker BMQC (red). (B) Magnified images of boxed area separated into colours and merged image demonstrating a crawling lymphocyte (red) surrounded by endothelial cytoplasm (green). (C,D) Representative confocal images of lymphocytes adherent to TNFα and IFNγ treated HSEC monolayers grown on a collagen gel containing chemokine (IP-10). Endothelial cells were pre-labelled with CellTracker CFMDA (green), nuclei with DAPI (blue) and immunofluorescent staining with VE-Cadherin (red). Orthogonal (xz) projections are shown below each overlay image. Bar 10 μm (B,C,D).

Supplementary Figure 6
Junctional molecule expression in HSEC does not change significantly with TNFα and IFNγ stimulation. (A) Cell-based ELISA of junctional molecule expression in unstimulated and TNFα and IFNγ treated HSEC. Data are the mean of five experiments and values represent the mean optical density at 490nm of three replicate wells minus the optical density of an isotype matched control Ab. Statistical significance was determined by two-tailed Ttest. Table 1 Pathways upregulated in HSEC compared to HUVEC. Results of gene ontology analysis of the microarray data on genes which were significantly upregulated in HSEC compared to HUVEC.
Harvested lymphocytes were washed and resuspended in RPMI1640/10% FCS, monocytes were depleted by plastic adherence.

Antibodies and Immunostaining
A mAb against Stabilin-1 (9-11) has been described (2) and was used at
Lymphocytes (1 × 10 6 cells/ml) were perfused through the microslide over the endothelial cells at a shear stress of 0.05 Pa.

Quantification of Actin and Microtubule fibres in HSEC.
Images were taken of monolayers of HSEC labeled for actin or microtubules as outlined above. Zstack images were taken of random fields. For each field 10 orthogonal projections were saved and threshold analysis performed with ImageJ software version 1.42q (NIH).

Electron Microscopy
HSEC were seeded at confluence on at confluence on rat tail collagen-coated coverslips and then cytokine-stimulated for 24 hours. Following this peripheral blood lymphocytes (1 × 10 6 cells/ml) were overlaid on HSEC for 20 minutes and then non-adherent lymphocytes were washed with PBS. The cells were then fixed at room temperature in 2.5% EM grade gluteraldehyde buffer followed by secondary fixation in osmium tetroxide. The samples were then dehydrated and embedded. Sections were cut using a diamond knife and stained with uranyl acetate and lead citrate. Sections were visualized with a JEOL electron microscope.

Quantitative real-time PCR
Total RNA was isolated and purified from cell pellets of HSEC and HUVEC using the RNeasy ® Mini Kit (Qiagen) and RNase-Free DNase Set (Qiagen), following the manufacturer's protocol. RNA quantity and purity was determined with a Nanophotometer™ (Implen GmbH) and cDNA was synthesised using SuperScript ® II Reverse Transcriptase (Thermo Fisher). mRNA expression of target proteins was assessed by quantitative real-time PCR (qPCR), utilising predesigned TaqMan ® Gene Expression Assays (Applied Biosystems ® ) (as detailed in supplemental Table 4) and 2X TaqMan ® Universal PCR Master Mix (Applied Biosystems ® ). qPCR was performed on a Roche Lightcycler 480 (Roche) using the following programme: 95°C for 10 min and 45 cycles of 95 °C for 10 s, 60 °C for 1 min, 72 °C for 1 s. Target mRNA levels were normalised to the housekeeping gene (GAPDH) and a fold change of relative expression from the appropriate unstimulated control was calculated utilising the 2 -ΔΔCt method (3).
Subsequently, cell monolayers were subjected to cytokine stimulation or relevant media control and incubated for a further 24 h. Transendothelial electrical resistance (TEER) was measured using a Millicell-ERS2 Volt-Ohm Meter (EMD Millipore) and expressed in Ωcm 2 . Alternatively, 1 mg/ml FITCdextran (Sigma-Aldrich) was added to the apical surface of the transwell insert and incubated at 37 °C for 1 h. Following this, the fluorescence of the basolateral media was measured (excitation λ = 485 nm, emission λ = 520 nm) and was expressed in relative fluorescent units (RFU).

Microarray
For microarray analysis, total RNA was isolated as described above from HUVEC and HSEC batches which were untreated and batches which were challenged with TNF-α and IFN-γ for 24 h at 10 ng/ml, two replicates were Fold change >2 with adjusted P<0.05 was considered for gene changes.
Gene ontology analysis was also performed with the GeneSpring software on genes that were significantly differently expressed.

Microarray data submission to public repository
As per recommendations made by the Microarray Gene Expression Data society, our microarray data has been deposited with the Gene Expression Omnibus, accession number GSE78020.