Colon 38 SL4 cells, a highly metastatic variant of the mouse colon 38 colon cancer cell line, were used. This cell line was maintained in a 1:1 mixture of Dulbecco’s modified Eagle’s medium and Ham’s F-12 medium (DMEM/F12; Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal calf serum (FCS), 2 mM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin in a humidified atmosphere of 95% air and 5% CO₂ at 37°C.

Colon 38 SL4 cells were transfected by Nucleofector™ (Amaxa, Inc., Gaithersburg, MD, USA). Expression vectors (pcDNA 3.1(+) vector; Life Technologies Ltd., Japan) for mouse membrane-bound CXCL16, pcDNA 3.1(+)-CXCL16 were used. DNA was adjusted to 1 μg with empty vectors. After transfection, CXCL16-positive colon 38 SL4 cells were selected by using the antibiotic G148 (Invitrogen). CXCL16-stable expression cells (SL4-CXCL16) and control cells (SL4-Mock) were maintained in DMEM/F12 supplemented with 10% FBS and antibiotics.

Cell growth was quantified using the cell proliferation reagent WST-8 [2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt] (Dojindo, Kumamoto, Japan). SL4 cells were seeded in 96-well microplates at 2×10³ cells/well and incubated 24 h. The medium was then changed (90 μl), 10 μl of WST-8 solution was added, and absorbance was measured at 450 nm.

The migration assay was performed in Transwell cell culture chambers (Corning Incorporated Life Sciences, Tewksbury, MA, USA). Polyvinylpyrrolidone-free polycarbonate (PVFP) filters (8.0 μm pore size; Nuclepore, Pleasanton, CA, USA) were pre-coated with Matrigel (BD Biosciences) on the lower part. SL4-Mock or SL4-CXCL16 cell suspension [1×10⁵ cells/100 μl in DMEM/F12 with 0.1% bovine serum albumin (BSA)] was added to the upper compartment and incubated for 4, 8 and 12 h at 37°C. Cells that invaded the lower surface were stained with hematoxylin and eosin and counted under a microscope at ×400 magnification. The adhesion assay was performed in 96-well plates. Triplicate wells were precoated with Matrigel and after removing the Matrigel, SL4-Mock or SL4-CXCL16 cells (1×10⁴ cells/100 μl in DMEM/F12) were incubated for 30 min at 37°C. Cells were aspirated and fixed with 4% formaldehyde approximately 30 min. Adhered cells were stained with hematoxylin for 20 min and counted under the microscope in five predetermined fields at ×400 magnification.

C57BL/6 female and KSN/slc male 5-week-old nude mice were purchased from Sankyo Lab Service (Hamamatsu, Japan). The study was conducted in accordance with the standards established in the guidelines for the Care and Use of Laboratory Animals of Toyama University.

Colon 38 SL4 cells were harvested and resuspended in 1 mM EDTA in phosphate-buffered saline (PBS). For experimental liver metastasis, colon 38 SL4 cells were injected into the intraportal vein. The mice were sacrificed 17 days later and the livers were removed. The increase in tumor weight and the number of tumor colonies in the liver were measured to evaluate tumor metastasis.

Hybridomas producing a 53-6.72 CD8 T cell a specific monoclonal antibody (ATCC, Manassas, VA, USA) were injected intraperitoneally into KSN/slc mice. C57BL/6 female mice were intraperitoneally injected with the ascites fluid of KSN/slc mice 3 days prior to and 3 and 6 days following the inoculation with cancer cells. CD8 T cell depletion was verified by flow cytometry.

di-palmitoyl-phosphatidylethanolamine polyethyleneglycol (DPPE-PEG) which is used as a reagent of NKT cell depletion (38), purchased from NOF Corporation, was dissolved with saline and injected intraperitoneally (250 μg/50 μl/mouse/day) until the mice were sacrificed.

Total-RNA was extracted using an RNeasy Mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s directions. First-strand cDNA was synthesized from an RNA template (2 μg) using SuperScript II reverse transcriptase (Invitrogen). The primer sequences are as follows: CD4, 5′-ACACACCTGTGCAAGAAGCA-3′ (forward) and 5′-GCT CTTGTTGGTTGGGAATC-3′ (reverse); CD8, 5′-CTCACCT GTGCACCCTACC-3′ (forward) and 5′-ATCCGGTCCCCTT CACTG-3′ (reverse); Vα14-Jα281, 5′-TGGGAGATACTCAG CAACTCTGG-3′ (forward) and 5′-CAGGTATGACAATCA GCTGAGTCC-3′ (reverse); CD3ɛ, 5′-AACACGTACTTGT ACCTGAAAGCTC-3′ (forward) and 5′-GATGATTATGGC TACTGCTGTCA-3′ (reverse); β-actin, 5′-CTAAGGCCAACC GTGAAAAG-3′ (forward) and 5′-ACCAGAGGCATACAGG GACA-3′ (reverse). Real-time quantitative RT-PCR was performed using a SYBR Premix Ex Taq (Takara Bio, Inc., Otsu, Japan) and Lightcycler nano system (Roche, Pleasanton, CA, USA). All data were normalized to β-actin mRNA.

To isolate lymphocytes from the liver, normal and tumor tissues from the liver were dissected in 20% RPMI-1640 medium. Samples were then further homogenized through wire mesh and mononuclear cells were isolated using a 30% Percoll gradient (GE Healthcare, Little Chalfont, UK) 5 ml, 10X PBS 365 μl, 7% NaHCO₃ (Meylon) 65 μl, Heparin 50 μl, and RPMI-1640 10 ml/mouse. Red blood cells were lysed by lysis buffer (NH₄Cl solution: 155 mM NH₄Cl, 100 mM KHCO₃ and 1 mM EDTA-2Na to 1 L with DW, Tris-HCl solution: Trizma 20.6 g was dissolved in 900 ml DW, adjusted pH 7.6 with concentrated HCl, diluted to 1 L with DW mix 900 ml NH₄Cl solution and 100 ml Tris-HCl solution, autoclaved, stored at 4°C). To isolate lymphocytes from the spleen, the Percoll gradient was used. For flow cytometry, lymphocytes were first pre-incubated with CD16/32 (2.4G2) mAb and then incubated with a saturating amount of mAb. Antibodies to CD3ɛ (145-2C11), NK1.1 (PK136), and CD1d (RA3-6B2) were purchased from eBioscience (San Diego, CA, USA). The samples were analyzed with the FACSCanto II system (BD Biosciences).

Data were analyzed for statistical significance using the Student’s t-test. P-values <0.05 were considered to indicate statistically significant differences. The mean and SD were calculated for all variables.

We initially intended to obtain CXCL16 stable-expression colon 38 SL4 cells (SL4-CXCL16) by the antibiotic selection. Higher CXCL16 expression in SL4-CXCL16 than SL4-Mock cells was confirmed by RT-PCR and western blotting (Fig. 1A and B). Gene introduction into cells often causes various property changes, compared with control cells. Compared with SL4-Mock cells, there was no significant difference in cell growth (Fig. 1C), migration (Fig. 1D) and adhesion (Fig. 1E) in SL4-CXCL16 cells.

Colon 38 SL4 cells (Parent, Mock and CXCL16) were injected into the intraportal vein to confirm the effect of CXCL16 expression on liver metastasis. Mice were sacrificed on Day 17 and liver metastasis was evaluated by counting nodules. Metastasis into the liver was significantly decreased by SL4-CXCL16 cells (Fig. 2A). Also, the number of nodules was significantly decreased in CXCL16-expressing tumor tissue (Fig. 2B). Since CXCL16 acts as a chemoattractant as well as a cell adhesion molecule for CXCR6-expressing cells (21,22), we conducted a real-time PCR analysis with tumor tissues to confirm which immune cells were recruited to the liver by CXCL16. Among the CXCR6-expressing cells, mRNA levels of CD8 T cell (CD8), NKT cell (Vα14-Jα281) and T cell (CD3ɛ) markers, but not CD4 T cell marker (CD4), were increased 6-, 10- and 4-fold compared with SL4-Mock (Fig. 2C). These results suggested that CXCL16 expression in metastatic tissues recruited CXCR6-expressing cells and then inhibited metastasis to the liver.

To confirm that CXCR6-expressing T cells are involved in liver metastasis by CXCL16- expressing CRC cells, we performed intraportal vein (i.p.v.) injections of SL4 cells (7.5×10⁴ cells/200 μl PBS/mouse) into nude mice (T cell-deficient mice). Metastasis to the liver did not differ significantly in the SL4-CXCL16 group (Fig. 3). Inhibition of metastasis to the liver by CXCL16 expression was not observed in T cell-deficient mice. These results indicate that T cells mediated the inhibition of liver metastasis by CXCL16.

Activated CD8 lymphocytes were recruited to inflamed liver in a mouse model and human patients (39,40). Thus, we examined the role of CD8 T cells in the inhibition of metastasis by CXCL16 in CD8 T cell-depleted mice. To analyze lymphocytes in tumor tissues, we injected SL4-Mock and SL4-CXCL16 cells (1.5×10⁵ cells/200 μl PBS/mouse) and sacrificed the animals at 17 days. In the CD8 T cell-depleted group, liver metastasis was decreased by CXCL16 expression the same as in the control group. Tumor weight and tumor/liver weight percentage were not changed in the CD8 T cell-depleted mice (Fig. 4). These results suggested that CD8 T cells were not related to the inhibition of liver metastasis by CXCL16.

NKT cells present in the liver and the activation of NKT cells regulate tumor surveillance in the liver microenvironment and suppresses liver metastasis (24–30). Thus, we examined whether NKT cells are involved in the inhibition of metastasis by CXCL16. We injected SL4-Mock and SL4-CXCL16 cells (1.5×10⁵ cells/200 μl PBS/mouse) into C57BL/6 mice and DPPE-PEG-treated mice. DPPE-PEG has been reported to be used to inhibit the action of NKT cells (38). The expression of NK1.1 and CD1d in metastatic tumor tissues was measured by FACS analysis. Thus, we used DPPE-PEG as a NKT cell-specific antagonist. As expected, levels of NK1.1 and CD1d were decreased by the DPPE-PEG treatment (Fig. 5A). In addition, the inhibition of liver metastasis by CXCL16 recovered in NKT cell-depleted mice (Fig. 5B). Tumor weight and tumor/liver weight percentage were significantly recovered in NKT cell-depleted mice (Fig. 5C and D). These results indicate that NKT cells are necessary for the inhibition of liver metastasis by CXCL16-expressing cancer cells.