Morris rat hepatoma 3924A cells (MH3924A) were purchased from the Institut für angewandte Zellkultur, (Munich, Germany). Cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM), with 10% heat-inactivated fetal bovine serum (FBS; Invitrogen Life Technologies, Carlsbad, CA, USA) at 37°C in a humidified incubator with an atmosphere of 5% CO₂. When the fusion rate reached >80%, the cells were digested with 25% trypsin and serial passages were performed at a dilution rate of 1:3.

Tumor cell proliferation was determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (MTT; Sigma-Aldrich, St. Louis, MO, USA). Briefly, after MH3924A cells had reached the logarithmic growth phase, a 0.2-ml cell suspension at 1×10⁴ cells/ml was added into each well of a 96-well plate and cultured in DMEM with 10% FBS, 10 μg/l vascular endothelial growth factor and 0.1 g/l heparin for 24 h. When adherent growth was established, different concentrations of FK506 (10, 100 and 1,000 μg/l), AMD3100 (10, 50 and 100 μg/l) and FK506 (0 and 100 μg/l) + AMD3100 (0, 10, 50 and 100 μg/l) were added into the plates. Untreated cells cultured in medium alone were used as controls. After culturing for 48 h, 10 μl MTT (5 g/l) were added, and each well was incubated for 6 h; next, 150 μl/well dimethyl sulfoxide were added, followed by measurements of the absorbance at 570 mm on a spectrophotometer reader (Dynatech MRX, Elvetec, Genas, France). Each well was measured three times, and each sample was assayed in triplicate.

MH3924A cells were cultured on sterile cover glasses (8×8 mm²) for 1 day, then cultured in DMEM with 100 μg/l FK506, or 50 μg/l AMD3100, or 100 μg/l FK506 + 50 μg/l AMD3100. When the cover glasses were covered with cells, the morphologic changes of MH3924A cells were observed with SEM, while a number of cover glasses were used for the immunohistochemical assay to determine the expression of CXCR4. This assay was performed using the anti-CXCR4 antibody (1:100 dilution; Wuhan Boshide Biotechnology Co., Wuhan, China) as the primary antibody and a horseradish peroxidase-conjugated secondary antibody (1:100; DakoCytomation, Glostrup, Denmark). The formed immunocomplex was visualized using the 3,3′-diaminobenzidine (DAB) reagent. The level of CXCR4 was quantified by a computer-assisted image system, which included a Leica CCD camera DFC420 connected to a Leica DM IRE2 microscope (Leica Microsystems, Wetzlar, Germany) (23,24). Image analysis yielded integrated optical density (IOD) values.

Invasion assays were performed in 8-μm, 24-well, BioCoat Matrigel invasion chambers (Corning, Cambridge, MA, USA). Briefly, 1×10⁴ cells/ml were suspended in serum-free medium, and 50 μl of the suspension were added to each chamber, while 500 μl of the culture medium containing different concentrations and combinations of FK506, AMD3100 and SDF-1α (Table I) were added to the bottom chamber. Cells were allowed to invade for 12 h at 37°C in a humidified incubator with an atmosphere of 5% CO₂; the cells that migrated to the top chamber were stained with Giemsa (Sigma-Aldrich) and were counted at a ×400 magnification under an electron microscope (Olympus, Tokyo, Japan). Assays were performed 3 times using triplicate wells.

Experiments were performed in 16 healthy August Copenhagen Irish rats (male, 16–20 weeks, weighing 240–300 g). The animals were handled in accordance with the regulations for laboratory animals. All animal experiments were in accordance with national guidelines and approved by the ethical committee of Zhongshan Hospital (Shanghai, China). The rat model of liver tumor was established as follows: First, MH3924A cells were collected and injected into the alar skin of rats. The tumors were removed from alar skin when grown to 2×1×1 mm³, and intrahepatic tumor implantation of rats was performed under aseptic conditions as described previously (25,26). Five days later, rats were randomly divided into two groups: one group was subcutaneously injected with normal saline for 14 days (NS group, n=8, 3 mg/kg/day), and the second group was subcutaneously injected with FK506 for 14 days (FK506 group, n=8, 0.3 mg/kg/day). Forty days following implantation, rats were sacrificed, and the weight of tumor, the volume of the fluid in the ascites, the incidence of lymphatic metastasis in the abdominal cavity and of abdominal wall metastasis were measured. In addition, the lungs were irrigated with 15% Indian ink, followed by counting of the number of metastatic nodules in the lung. The tumor and adjacent tissues, as well as healthy liver tissues, were harvested and preserved in 4% formalin for later use.

The tumor and adjacent tissues along with healthy liver tissues were sectioned (4 μm) according to the EliVision method (27). First, all sections were dried at 92°C for 30 min, treated with a 100% dimethylbenzene, 3% formalin-H₂O₂ solution, and washed with phosphate-buffered saline. Second, the sections were incubated with polyclonal antibodies of CXCR4 and SDF-1α purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA) and Boster Biotechnology (Wuhan Boster Biological Technology, Ltd., Wuhan, Hubei, China), respectively, colored with DAB reagent and counterstained with hematoxylin and eosin. Following washing in PBS and drying and mounting in buffered glycerol-saline (90% glycerol, 10% PBS), the sections were observed under an inverted confocal microscope (DMRIE2; Leica Microsystems, Wetzlar, Germany). Sections were observed at a ×500 magnification, and 5 pictures were randomly selected and saved. The analysis of these images yielded IOD values.

Data with a normal distribution are expressed as mean ± standard deviation (SD), while median values are used to represent data that were not normally distributed. We used the Windows version of the SPSS 16.0 software (SPSS, Inc., Chicago, IL, USA) to perform one-way analysis of variance for the comparison of normally distributed data, and the Mann-Whitney rank sum test implemented in the Image-Pro Plus (IPP) v6.0 software (Media Cybernetics, Inc., Bethesda, MD, USA) to compare the data that were not normally distributed, as for example the IOD values from the immunohistochemical assay. P<0.05 and P<0.01 were considered to indicate statistically significant differences.

As shown by the MTT assay (Table II), treatment with a low concentration of FK506 (10 μg/l) did not significantly affect the proliferation of MH3924A cells (P=0.135). Upon treatment with higher concentrations of FK506 (100–1,000 μg/l), the proliferation of MH3924A cells was significantly enhanced (P<0.01). Treatment with AMD3100 at any concentration (10, 50 or 100 μg/l), had no obvious effect on MH3924A cell proliferation (P>0.05). However, when different concentrations of AMD3100 were combined with 100 μg/l FK506, the in vitro proliferation of MH3924A cells was increased (P<0.01, Table III). These data suggested that FK506 (≥100 μg/l) can promote the proliferation of MH3924A cells and that AMD3100 has no effect on the FK506-induced increase in proliferation.

The effects of FK506 and AMD3100 on the expression of CXCR4 were studied in MH3924A cells with an immunohistochemical assay (Fig. 1). In non-treated MH3924A cells (Fig. 1A), CXCR4 expression was detectable at medium levels. CXCR4-stained particles were mainly distributed in the cytoplasm and the extracellular matrix, were occasionally present on the cytomembrane, and were absent from the nucleus. Following treatment with 100 μg/l FK506 (Fig. 1B), stronger staining of CXCR4 was observed, although this increase was not significant (P>0.05). By contrast, upon treatment with 50 μg/l AMD3100 (Fig. 1C), no or weaker expression of CXCR4 was observed compared to non-treated cells (P<0.05). The expression of CXCR4 was also negative or weaker in MH3924A cells treated with 100 μg/l FK506 combined with 50 μg/l AMD3100, compared to non-treated cells (P<0.05). These results suggested that the expression of CXCR4 is marginally increased by treatment with 100 μg/l FK506, but decreased by treatment with AMD3100 or FK506 + AMD3100.

Examination of the cell morphology was performed with SEM (Fig. 2). The shape of non-treated MH3924A cells was regularly elliptical, and their surface was covered with microvilli and a few immobile processes (Fig. 2A). The MH3924A cells that were treated with 100 μg/l FK506 displayed a higher number of microvilli, longer immobile processes, and their shape was irregularly expanded compared to non-treated cells (Fig. 2B), suggesting stronger invasive ability. The cells treated with 50 μg/l AMD3100 were regularly elliptical similar to non-treated cells, while the number of microvilli and immobile processes did not significantly change (Fig. 2C). In addition, the morphology of MH3924A cells that were treated with 100 μg/l FK506 and 50 μg/l AMD3100 (Fig. 2D), was similar to that of cells treated with 100 μg/l FK506, displaying a higher number of microvilli and immobile processes, and irregular expansions; this morphology indicates increased cell invasiveness.

MH3924A cells are invasive. Therefore, their invasive ability was measured in vitro (Fig. 3). In the FK506 group, the number of invasive cells was significantly increased compared to the NS group (P<0.01). AMD3100 treatment had no significant effect on invasive ability compared to the NS group (P=0.09), but it clearly decreased the invasiveness of cells compared to treatment with FK506 (P=0.046) or SDF-1α (P=0.032), similarly to the NS group. In the SDF-1α group, the number of invasive cells was significantly increased compared to the NS group (P<0.01), but no significant difference was observed in the comparison with the FK506 group (P=0.881).

The number of invasive cells after treatment with the combination FK506 + AMD3100 was significantly increased compared to the NS group (P<0.01), but was not significantly affected compared to the FK506 (P=0.607) and SDF-1α (P=0.507) groups. In addition, the number of invasive cells was significantly increased in the FK506 + SDF-1α compared to the NS group (P<0.01), and was increased compared to the FK506 and the SDF-1α groups, but this increase was not significant (P=0.653 and P=0.548, respectively). Moreover, the number of invasive cells in the AMD3100 + SDF-1α group did not significantly change compared to the NS group (P=0.864). In the FK506 + AMD3100 + SDF-1α group, the invasive ability of MH3924A cells was significantly enhanced compared to the NS group (P<0.01), but not significantly enhanced in comparison to the FK506 (P=0.983) and the SDF-1α (P=0864) groups. These results showed that both FK506 and SDF-1α can enhance the invasive ability of MH3924A cells, and that AMD3100 can reduce the invasive ability of MH3924A cells when combined with SDF-1α.

Following treatment with NS or FK506, none of the rats died or failed to keep the tumor implant until the day of observation. Tumor growth was examined after the implantation (Table IV). The liver of rats of the NS group was large, with an average weight at 15.56±11.17 g (Fig. 4A), and the liver of rats of the FK506 group was oversize, with an average weight at 28.19±3.89 g (Fig. 4B); no significant difference was found between the two groups (P=0.041). Regarding the number of metastatic nodules in the lungs, this was significantly increased in the FK506 compared to the NS group (6.50±4.63 vs. 1.39±1.25, P=0.012) (Fig. 4C–D). Moreover, the ascite fluid volume was increased in rats of the FK506 group compared to the NS group, although this change was not statistically significant (21.25±6.94 vs. 13.13±21.87 ml, P=0.317). The rate of lymph node metastasis, as well as the number of pulmonary nodules, were significantly increased in the FK506 compared to the NS group (P=0.002 and P=0.012, respectively). However, no significant difference in the rate of abdominal wall metastasis was observed between the two groups (P=0.442).

Immunohistochemical staining was performed in order to detect the expression of CXCR4 in tumor HCC tissues and of SDF-1α in the tissues adjacent to the HCC ones (Table IV). In the NS group, CXCR4 was expressed at medium levels, with an average IOD at 1.48±0.29 (Fig. 5A), but in the FK506 group, the expression of CXCR4 was higher, with an average IOD at 2.50±0.62 (Fig. 5B). The increased expression of CXCR4 following FK506 treatment was significant (P=0.048). In addition, the expression level of SDF-1α in tissues adjacent to the HCC ones was medium in the NS group, with an average IOD at 1.46±0.39 (Fig. 6A). In the FK506 group (Fig. 6B), the expression of SDF-1α in the adjacent tissues was significantly higher compared to the NS group (P=0.026), with an average IOD at 2.54±0.94.