The human HCC cell lines HepG2, Hep3B, Huh7, MHCC97L, MHCC97H and HCCLM3, and the human normal hepatic cell line L-02, were routinely maintained in high-glucose Dulbecco’s modified Eagle’s media (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 U/ml penicillin and 100 mg/ml streptomycin. All cell lines were cultured at 37°C in a humidified incubator in an atmosphere of 5% CO₂. Male athymic BALB/c nude mice (4-weeks old; Shanghai Institute of Material Medicine, Chinese Academy of Science, Shanghai, China) were raised under pathogen-free conditions. All animal care and experimental protocols were carried out in accordance with the guidelines established by the Shanghai Medical Experimental Animal Care Commission.

Patient samples were collected after obtaining informed consent according to an established protocol approved by the Ethics Committee of Fudan University. The data did not contain any information that could lead to the identification of the individual patients.

Samples for real-time polymerase chain reaction (PCR) studies were randomly collected from patients undergoing curative resection for HCC at the Liver Cancer Institute, Zhongshan Hospital, Fudan University, in March 2006. Samples were collected immediately after resection, transported in liquid nitrogen, and stored at −80°C. Sixteen frozen tissue samples were also obtained from the above patients for western blot studies.

Tumor specimens for tissue microarray (TMA) studies were obtained from 228 consecutive HCC patients who underwent curative resection without preoperative treatment at the Liver Cancer Institute, Zhongshan Hospital, Fudan University, between 2005 and 2006. Complete follow-up data were available for each patient, and the diagnosis of HCC was confirmed by pathological examination.

Total RNA was extracted from cell lines and frozen tumor specimens using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Total RNA (2 μg) was reverse transcribed using a RevertAid First-Strand cDNA Synthesis kit (Fermentas, Burlington, ON, Canada). Reverse transcription-PCR was performed before quantitative real-time PCR. Lamp2a mRNA expression was determined by real-time PCR using SYBR Premix Ex Taq (Takara Bio, Dalian, China). PCR amplification cycles were carried out for 10 sec at 95°C, followed by 40 cycles of 95°C for 5 sec and 60°C for 30 sec. Data were collected after each annealing step. β-actin was used as an endogenous control to normalize for differences in the amounts of total RNA in each sample. Relative gene expression levels were calculated and were expressed as 2^−ΔCt, as previously described (16). The following primers were used: β-actin 5′-CAACTGGGACGACATGGAGAAAAT-3′ and 5′-CCAGAGGCGTACAGGGATAGCAC-3′; Lamp2a 5′-AGACTGCAGTGCAGATGACGAC-3′ and 5′-GACCAATAAAATAAGCCAGCAAC-3′.

Western blot analysis was performed as previously described (16). Briefly, proteins from total cell lysates were separated by standard sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to polyvinylidene difluoride membranes. The membranes were washed, blocked and incubated with specific primary anti-human antibodies against β-actin (1:1,000; ab8226; Abcam, Cambridge, MA, USA) or Lamp2a (1:1,000; ab18528; Abcam), followed by incubation with horseradish peroxidase-conjugated secondary antibodies. The reactions were detected by enhanced chemiluminescence assay.

Lentiviral-mediated pGC-GFP-Lamp2a and pGCSIL-GFP-Lamp2a short hairpin RNAs (shRNAs) were constructed (Shanghai Genechem, Co., Ltd., Shanghai, China). The shRNA targeting sequence (5′-AAGCACCATCATGCTGGATAT-3′) for Lamp2a was used. Huh7 and HCCLM3 cells were transfected with lentivirus particles and the cell populations expressing GFP-Lamp2a or GFP-Lamp2a shRNA were isolated by flow cytometry (BD Biosciences, San Jose, CA, USA). Stable transfectant clones were further validated by real-time PCR, and immunoblotting was used to determine the expression levels of Lamp2a protein.

Cell viability was assessed using an MTT (3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide) kit (Trevigen, Gaithersburg, MD, USA), according to the manufacturer’s protocol. Cells (5×10³) were seeded in 96-well plates, incubated for 24 h at 37°C, and treated with the specified agents at defined time-points.

Cell migration was evaluated using the scratch-wound assay. Cells were cultured for 2 days to form a tight cell monolayer and then serum-starved for 6 h. After serum starvation, the cell monolayer was wounded using a 20-μl plastic pipette tip. The remaining cells were washed twice with culture medium to remove cell debris and incubated at 37°C with normal serum-containing culture medium. Migrating cells at the wound front were photographed at the indicated times, using an inverted microscope (Leica Microsystems, Wetzlar, Germany). The percentage of the cleared area at each time-point compared with time zero was measured using Image-Pro Plus software v6.2.

Filters coated with Matrigel (BD Biosciences) in the upper compartment were loaded with 100 μl medium containing 1×10⁵ cells and the lower compartment was filled with conditioned culture medium mixed with DMEM and supplemented with 10% FBS, NIH3T3 and HCC cell super supplements. After 36 h, migrated cells on the bottom surface were fixed with 4% paraformaldehyde and counted after Giemsa staining.

Autophagy was assessed using GFP-LC3 redistribution. Redistribution of GFP-LC3 was detected using an inverted fluorescence microscope. The number of GFP-LC3-positive dots per cell was determined in three independent experiments. Eight randomly selected fields representing 200 cells were counted.

HCCLM3-control, HCCLM3-Lamp2a, Huh7-control and Huh7-Lamp2a shRNA cells (5×10⁶) were suspended in 100 μl serum-free DMEM and Matrigel (BD Biosciences) (1:1), and then inoculated into the liver parenchyma or the right upper flank subcutaneous region of nude mice, as previously described (20). The mice were sacrificed 5 weeks after tumor implantation. At necropsy, the volumes of the largest (a) and smallest (b) tumors were measured and the tumor volume was calculated as: V = a × b² × π/6. The tumor sections were prepared for immunohistochemical staining. Immunoreactivity was analyzed using Ki-67 (1:1,000; ab15580; Abcam) and P62 (1:1,000; ab56416; Abcam) in tumor tissues. Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL) staining was performed using an In Situ Apoptosis Detection kit (VB-4005; GeneCopoeia, Rockville, MD, USA), according to the manufacturer’s instructions.

TMA was constructed as previously described (21). Briefly, all the HCC tissues were reviewed by two histopathologists, and representative areas free from necrotic and hemorrhagic material were premarked in the paraffin blocks. Two core biopsies (1 mm in diameter) were taken from the donor blocks and transferred to the recipient paraffin block at defined array positions. Three different TMA blocks were constructed, each containing 200 cylinders. Consecutive sections (4 μm in thickness) were placed on 3-aminopropyltriethoxysilane-coated slides (Shanghai Biochip, Co., Ltd., Shanghai, China).

Immunohistochemistry was performed with monoclonal rabbit antibodies against human Lamp2a (1:100; ab18528; Abcam), using a two-step protocol (Novolink Polymer Detection system; Novocastra/Leica Biosystems, Richmond, IL, USA), as previously described (21). Briefly, after microwave antigen retrieval, tissues were incubated with primary antibodies for 60 min at room temperature, followed by incubation for 30 min with the secondary antibody (RE7112; Novolink Polymer; Leica Microsystems GmbH, Wetzlar, Germany). The sections were developed in 3,3′-diaminobenzidine solution under microscopic observation and counterstained with hematoxylin. Negative control slides, in which the primary antibodies were omitted, were included in all assays.

Immunohistochemical staining was evaluated by 3 independent pathologists with no knowledge of the patient characteristics. Scores were assigned for the intensity and percentage of positive staining in the cytoplasm in the whole cylinder. Discrepancies were resolved by consensus among the three pathologists, using a multihead microscope. The criteria for Lamp2a-positivity included moderate or strong immunoreactivity in >20% of the cells. In the event of a difference between duplicate tissue cores, the higher score was considered to be the final score.

Means were compared between two groups using unpaired, two-tailed Student’s t-test, and among multiple groups using one-way analysis of variance. Categorical data were analyzed using χ² or Fisher’s exact tests. The Kaplan-Meier method was used to determine survival probability and differences were assessed by the log-rank test. Statistical significance was set at P<0.05. All analyses were performed using SPSS software (v.16.0).

In several HCC cell lines with different malignant phenotypes and invasive potentials as previously described (16), Lamp2a mRNA and protein expression levels were evaluated. Lamp2a mRNA expression levels were lower in HCC cells compared with normal hepatic L-02 cells, as confirmed at the protein level by western blotting (Fig. 1A). Among the HCC cell lines, Lamp2a expression levels were high in Huh7, intermediate in MHCC97L, MHCC97H and HepG2, and low in Hep3B and HCCLM3 cells, which did not correlate with their malignancy or invasiveness. We further evaluated the effect of Lamp2a expression in HCC cells in paired isogenic HCC cell lines (termed HCCLM3-Control/HCCLM3-Lamp2a and Huh7-Control/Huh7-Lamp2a shRNA) in which Lamp2a expression was modified by RNA interference or cDNA transfection (Fig. 1B). Light microscopy showed no significant difference in cell morphology in relation to Lamp2a expression (Fig. 1C).

We also assessed cellular viability, migration ability and invasion ability in paired HCC cell lines. Two groups of HCC cells were incubated under normal conditions for 6–36 h and cellular viability was measured using MTT assays. There was no significant difference in survival in cells with upregulated or downregulated Lamp2a expression compared with control cells (Fig. 2A). Cell migration as the first step of tumor metastasis is a characteristic of invasive tumors. Wound-healing migration assays were adopted to assess the effect of Lamp2a expression on HCC cell migration. No significant difference was revealed in wound-closure rate of HCCLM3-Lamp2a and Huh7-Lamp2a shRNA cells compared with the control cells under microscopic examination after 24 and 48 h (Fig. 2B). Invasive capacity of the HCC cells was assessed by Matrigel invasion assays. As shown in Fig. 2C, Lamp2a overexpression in HCCLM3 and Lamp2a silencing in Huh7 cells had no effect on cellular invasion compared with control cells.

CMA is activated maximally in response to stressors and changes in cellular nutritional status (22). We therefore explored the role of Lamp2a as the critical receptor of CMA in HCC during cell starvation. Isogenic paired HCC cell lines (HCCLM3-Control/HCCLM3-Lamp2a and Huh7-Control/Huh7-Lamp2a shRNA) were incubated in the absence of serum for 6–32 h and cellular viability was measured using MTT assays (Fig. 3A). There were no differences in viability between the cell line pairs after 6 and 12 h of starvation, but upregulation of Lamp2a in HCCLM3 cells markedly increased cell survival after prolonged starvation for 24 and 36 h by 10 and 16.5%, respectively, compared with control cells. Similarly, silencing Lamp2a expression in Huh7-Lamp2a shRNA cells abolished the protective effect of Lamp2a and reduced survival to 16.6 and 12.5%, respectively, compared with control cells.

Macroautophagy is also activated by starvation and is involved in the protective mechanism against environmental and cellular stress (23). We further quantified the level of autophagy by transfecting cells with GFP-LC3 fusion protein (Fig. 3B). Following starvation GFP-LC3 protein changed from a diffuse cytoplasmic distribution to punctate GFP-LC3 dots, indicating a formation of autophagy vacuoles. Morphometric analysis revealed similar numbers of GFP-LC3-positive dots per cell in Lamp2a-modified cells after short and prolonged starvation, indicating that the protection offered by Lamp2a was independent of macroautophagy.

Ortho-topic injection of HCCLM3-Control and HCCLM3-Lamp2a, or subcutaneous injection of Huh7-Control and Huh7-Lamp2a shRNA cells in nude mice led to tumor formation in all the groups. However, tumors from HCCLM3-Lamp2a-derived xenografts were significantly larger (2,164.6±332.0 mm³) than those from HCCLM3-Control xenografts (1,477.6±732.4 mm³; P<0.05) (Fig. 4A). Moreover, Huh7-Lamp2a shRNA-derived xenografts were significantly smaller (426.5±214.4 mm³) than Huh7-Control-derived tumors (880.3±293.3 mm³; P<0.01) (Fig. 4B).

Lamp2a overexpression caused a moderate decrease of TUNEL-positive tumor cells compared with HCCLM3-Control-derived tumors (4.0±1.5 vs. 4.9±1.7; P=0.002). Significantly more TUNEL-positive tumor cells were also observed in Huh7-shRNA-derived xenografts compared with control tumors (10.7±3.0 vs. 5.6±1.4; P<0.001) (Fig. 4C). Similarly, Ki-67 staining indicated a significant increase in proliferative activity in the Lamp2a-overexpressing HCCLM3 orthotopic model and control Huh7 subcutaneous model compared with the control HCCLM3 and Huh7-shRNA groups, respectively (69.0±6.6 vs. 62.0±8.9%; P<0.001, and 80.4±4.1 vs. 67.7±4.5%; P<0.001) (Fig. 4C). We examined the effect of Lamp2a expression on macroautophagy in tumor xenografts in vivo by investigating the expression of P62, which accumulates during defective autophagy. Immunohistochemical analysis revealed no significant difference of P62 expression between the two groups, suggesting that macroautophagy activity was independent of Lamp2a expression (Fig. 4C).

We compared Lamp2a mRNA expression in 42 HCC tissue samples and para-tumor tissues and 10 normal livers. Compared with para-tumor tissue samples and normal liver samples, Lamp2a expression was significantly decreased in tumor tissue samples (P<0.001 and P=0.001, respectively) (Fig. 5A). Lamp2a expression was lower in 83.3% (35/42) of tumor tissues compared with matched para-tumor tissues. Levels were decreased >3-fold in 10 samples, and 2- to 3-fold in 11 samples among the above tumor samples. These findings were further confirmed in 16 of the 42 HCC cases by western blot analysis (Fig. 5B). Lamp2a protein levels were decreased in 11 of the 16 samples compared with the para-tumor tissues, especially in cases of non-recurrent HCC. Higher Lamp2a mRNA expression was associated with HCC recurrence (P=0.036) and larger HCC size (P=0.016). Together with the results of the in vivo studies, these findings suggest that Lamp2a expression in HCC tissues contributes to tumor growth and may also influence the clinical prognosis.

We further validated our hypothesis using TMA to evaluate the correlation between Lamp2a and prognosis in 228 HCC patients who underwent curative resection. Strong Lamp2a staining in hepatocytes and tumor cells was observed in most para-tumor tissue samples (176/228) and some tumor tissue samples (47/228) (Fig. 5C). High Lamp2a expression, indicated by moderate or strong immunohistochemical staining in >20% of the cells, was present in 43% (99/228) of HCC patients. However, there was no correlation between Lamp2a expression and clinicopathological characteristics such as gender, liver cirrhosis, tumor number, tumor capsule, tumor differentiation or vascular invasion. Older patients and patients with high α-fetoprotein levels had lower Lamp2a expression. Moreover, high expression of Lamp2a was significantly correlated with large tumor size (P=0.001) (Table I). The overall 3-, 5- and 7-year cumulative recurrence rates in these HCC patients were 34.1, 47.9 and 54.5%, respectively, though patients with high Lamp2a expression had a significantly higher recurrence rate than patients with low Lamp2a expression (P=0.005) (Fig. 5D). The 3-, 5- and 7-year cumulative recurrence rates for Lamp2a^high and Lamp2a^low groups were 43.4 vs. 26.8, 58.2 vs. 39.6 and 62.2 vs. 46.2%, respectively.