Galectin-9 is a member of a growing family of animal lectins defined by their β-galactoside-binding activity and by a conserved specific sequence motif called lectin domain.1–4 We have purified and cloned a unique T cell-derived eosinophil chemoattractant (ecalectin) and found that ecalectin is the correct galectin-9.5, 6 A matrix metalloproteinase and protein kinase c are, at least, partly involved in the release of galectin-9 from T cells7 and the carbohydrate recognition domains but not the linker peptide of galectin-9 are required for its potent and selective eosinophil chemoattractant activity.8 With regards to chemoattractant activity, there is no doubt that the target cell for galectin-9 is eosinophil.
Galectins appear to modulate a variety of cellular functions, such as cell proliferation and cell death other than chemoattractant activity.9–13 Thus, it cannot be excluded the possibility that galectin-9 acts on other cells to exhibit different biological functions other than chemoattraction. Indeed, galectin-9 exhibits proapoptotic activity on thymocytes and activated T lymphocytes,14 probably CD8-positive T lymphocytes.15 Galectins are related to not only apoptosis but also cell adhesion.16–20 It is not surprising that galectin-9 also plays a crucial role in cell adhesion or aggregation in addition to its chemotactic activity for eosinophils and proapoptotic activity for thymocytes and T lymphocytes.
Galectin-121 and -322–28 probably play a role in invasion and metastasis of cancer cells. Despite evidence that galectin expression links to tumor progression,19 little information about galectin-9 in melanocytic tumors is available. We report evidence that galectin-9 induces cell aggregation and apoptosis of melanoma cells in vitro and that high expression of galectin-9 in melanoma cells in the patients with melanoma is associated with a better prognosis.
MATERIAL AND METHODS
Human melanoma cell lines
Human melanoma cell lines, such as MM-AN, MM-BP, MM-EP, MM-LN, MM-MC and MM-RU established by Byers29, 30 were maintained in 10% FCS-containing MEM supplemented with antibiotics (Sigma, St. Louis, MO).
Polyclonal antibody against human galectin-9 was prepared from rabbits immunized with the C-terminal domain of human galectin-96 and purified using a Sepharose column conjugated with the C-terminal domain of galectin-9. The antibody used in the present experiments exhibited no or little cross-antigenicity with other galectins, such as galectin-1 and -3. The DAKO EnVision+™, Peroxidase, Rabbit was purchased from DAKO Corporation (Carpinteria, CA).
Total RNA from melanoma cell lines was isolated using TRIZOL reagent (Gibco BRL, Rockville, MD). Using a Gene Amp RNA PCR kit (Perkin-Elmer, Branchburg, NJ), 0.5 μg of total RNA was reversed transcribed to DNA in 1 step followed by a polymerase chain reaction (PCR) to amplify transcripts of human galectin-1, -3, -9 and G3PDH. Reverse transcription (RT) reaction and PCR steps were done following the manufacture's instruction. We used the following primer sequences synthesized at Amersham Pharmacia Biotech (Buckinghamshire, England): galectin-1 sense, TGGTCGCCAGCAACCTGAATCTCA; galectin-1 antisense, TAGTTGATGGCCTCCAGGTTGAGG; galectin-3 sense, ACCCATCTTCTGGACAGCCAAGTG; galectin-3 antisense, CACTGCAACCTTGAAGTGGTCAGG; galectin-9 sense, CGTCAATGGCTCTGTGCAGCTGTC; and galectin-9 antisense, AGATCCACACTGAGAAGCTCTGGC.
Thirty PCR cycles were used for amplification of all transcripts and reactions were performed in a GeneAmp PCR System 9600 (Perkin-Elmer). PCR products were run on a 1.5% agarose gel. After purification of the respective PCR products, sequencing of galectin-9 PCR products was performed using an ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction Kit (Perkin-Elmer). For each reaction, the following reagents were added to a tube: 8 μl of Terminator Reaction Mix (Perkin-Elmer), 500 ng of PCR product, 3.2 pmol of galectin-9 primers and deionized water. Sequencing of DNA was performed on a GeneAmp PCR System 2400 (Perkin-Elmer).
Flow cytometric analysis
The intracellular expression of galectin-9 was assessed by a modification of the methods of Krug et al.31 In brief, the cells were incubated in 30 mM lactose-containing phosphate buffered saline containing 0.05% NaN3 and 2% fetal calf serum (PBS+) on ice for 30 min to eliminate surface-bound galectin-9 and fixed with ice-cold PBS containing 4% paraformaldehyde for 10 min. After washing with PBS+, the cells were resuspended in 95 μl saponin buffer (PBS+ containing 0.1% saponin and 0.01 M HEPES buffer) at pH 7.4. After the addition of 25 μg/ml of rabbit-anti-human galectin-9 antibody in saponin buffer, the cells were incubated for 30 min at room temperature, followed by an incubation of 45 min on ice with fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG antibody (Santa Cruz Biotechnology, Santa Cruz, CA). As a control, we used normal rabbit IgG.
To assess cell surface galectin-9 expression, cells were centrifuged and washed with PBS+, followed by an incubation of 1 hr on ice with 50 μg/ml of rabbit-anti-human galectin-9 antibody. After several washing with PBS+, the cells were incubated with FITC-conjugated goat anti-rabbit IgG antibody (Santa Cruz Biotechnology) for 45 min on ice.
Surface and cytoplasmic galectin-9 expressions were analyzed with a flow cytometer (EPICS XL-MCL, Coulter, Miami, FL).
Experiments were done to clarify whether exogenous galectin-9 induces cell aggregation of MM-RU proliferating without colony formation. The cells (5 × 104/ml) were cultured in the presence or absence of galectin-9 at various concentrations (1 × 10−9 M to 3 × 10−7 M) for 24 hr. Microscopic observation was performed at various intervals (2, 6, 12 and 24 hr) after the incubation. As controls, galectin-1 and -3 were used.
For quantitative analysis, experiments were further performed by a modification of the method by Urushihara et al.32 In brief, we suspended MM-RU (4 × 105) with 0.5 ml of galectin-9 at various concentrations in a 24-well culture plate previously coated with 1% BSA and incubated on a gyratory shaker (80 rpm) for 12 hr at 37°C. To clarify whether lactose suppresses galectin-9-induced aggregation, cells were incubated in the presence of 30 mM lactose together with galectin-9. As a control, sucrose was used instead of lactose. The total particle number of each well was counted using a Coulter counter (Coulter). The degree of aggregation was represented by an aggregation index defined by Nt/No, where No is the total cell number per well and Nt is the total particle number per well at 12 hr culture.
Flow cytometric analysis was done to clarify whether galectin-9 exhibits apoptotic activity on melanoma cells. In short, MM-RU cells (1 × 10−6 cells in 300 μl PBS) were treated with galectin-9 for 72 hr at 37°C. After 24, 48 and 72 hr culture, the cells were mixed with 700 μl of ice-cold ethanol and incubated for 30 min at 4°C. After centrifugation, the cells were resuspended with 1 ml PBS and treated with 50μg RNase A (Sigma) for 30 min at 37°C. After centrifugation, the cells were further resuspended in 50 μl/ml propidium iodide (PI, Calbiochem, Inc., San Diego, CA) for 10 min at 4°C in the dark and analyzed with a flow cytometer. To competitively inhibit the binding of galectin-9, lactose (30 mM) was added to the culture media and sucrose was used as a control.
Proapoptotic activity was also assessed by Annexin V binding and analysis using Annexin V-FITC kit (Immunotech, Marseille, France) according to the manufacturer's instructions. MM-RU cells were harvested after 72 hr of exposure to galectin-9. The cells were washed with ice-cold PBS after centrifugation for 5 min at 500g at 4°C. The cell pellet was finally resuspended in binding solution (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl2) at a concentration of 105 to 106 cells/ml. Annexin V-FITC solution (5 μl) and dissolved PI (2.5 μl) were added to 100 μl of the cell suspension. After mixing gently, the cell suspension was incubated on ice for 10 min in the dark. The cells were then washed and resuspended in binding buffer and analyzed by flow cytometry.
Melanocytic lesions were obtained from patients who underwent surgery in the Department of Dermatology, Kumamoto University School of Medicine, Kumamoto, Japan. We examined the levels of immunoreactive galectin-9 in the lesions of 20 melanocytic nevi, 70 primary melanoma and 31 metastatic melanoma. Tumor staging was based on the histopathological TNM classification system.33 Among the patients with primary melanoma, 4 patients had Stage 0, 21 Stage 1, 17 Stage 2, 24 Stage 3 and 4 Stage 4. Histopathological classification revealed that 4 patients were melanoma in situ (IS), 35 acral lentiginous melanoma (ALM), 12 superficial spreading melanoma (SSM), 7 nodular melanoma (NM), 5 mucous melanoma and 7 lentigo maligna melanoma (LMM).
Tissues were processed within 15 min after surgical removal. Each tumor tissue was fixed in 10% buffered formaldehyde and processed for routine histopathology.
Indirect immunoperoxidase staining
Indirect immunoperoxidase staining of formalin-fixed and paraffin embedded tissue sections with anti-galectin-9 antibody was performed utilizing the DAKO EnVision+™, Peroxidase, Rabbit System following the manufacturer's instructions. Briefly, after deparaffinization and rehydration of 4 μm formalin-fixed and paraffin-embedded tissue sections, endogenous peroxidase activity was blocked with 0.03% peroxide. The tissue sections were incubated with primary antibody (1:50 dilution) for 30 min at room temperature and then incubated with EnVision+™, Peroxidase, Rabbit for 30 min at room temperature. 3,3′-Diaminobenzidine tetrahydrochloride was used as the chromogen. We used an immunoglobulin fraction from preimmune rabbit sera as a negative control. All sections were counterstained with Giemsa's solution. We estimated the intensity and percentage of the galectin-9-positive tumor cells in each section independently by two observers. The mean percentage, rounded off to the nearest multiple of 10, was used to express the results.
The immunostaining intensity was graded as 0 (no staining was detectable), 1 (staining was weak), 2 (staining was clearly positive) and 3 (staining was strongly positive).
The immunohistochemical evaluation incorporating both the percentage and intensity of stained cells (HSCORE; histochemical score) was used34 and HSCORE was calculated by a following formula:
where ¡ = 1, 2, 3 and Pi varies from 0–100%. An HSCORE greater than 100 was defined as high galectin-9 expression.
Statistical analysis was done using StatView software, v. 4.5. The differences in galectin-9 expression were analyzed using the χ2 test. Correlation between the expression of galectin-9 in primary melanoma lesions and lymph node metastasis, recurrence or survival was tested using the χ2 and Fisher exact tests. The disease-free and survival curves were calculated using the method of Kaplan and Meier. Differences between disease-free and survival curves were analyzed with a log rank test. Multivariate analysis using Cox's proportional-hazards regression model was performed to study the effects of different variables on survival. All p-values were based on two-tailed statistical analysis and p-values of <0.05 were considered statistical significant.
Galectin mRNA expressions in human melanoma cell lines
Proliferation profile was assessed in human melanoma cell lines such as PRM-EP, PRM-MC, MM-AN, MM-BP, MM-LN and MM-RU in vitro. MM-BP is a small pleomorphic cell line and proliferated in vitro with colony formation (Fig. 1a). In contrast, MM-RU is pleomorphic cell line and failed to form colonies during their proliferation (Fig. 1a). These characteristics were consistent with their original cell lines.29 Other melanoma lines such as MM-AN, MM-LN, PRM-EP and PRM-MC also failed to exhibit evident colony formation compared to MM-BP (data not shown). Therefore, we used MM-BP and MM-RU for further experiments.
RT-PCR was assessed to clarify whether the melanoma cell lines express mRNA of galectin-1, -3 and -9. Figure 1b showed that mRNA expression of galectin-9 was evident in MM-BP though that in MM-RU was very weak. In contrast, those lines expressed evident mRNAs of galectin-1 and -3 and there was no significant difference in mRNA expressions of galectin-1 and -3 between MM-BP and MM-RU. Furthermore, MM-BP had mRNAs of the medium-sized (483 bp) and the long-sized (560 bp) gaelctin-9 isoforms though that of short-sized galectin-9 isoform was not detectable. We thus raised the hypothesis that galectin-9 is involved in colony formation in melanoma cells.
Galectin-9 protein expression in human melanoma cell lines
Western blot analysis was performed to compare the protein levels of galectin-9 in MM-BP and MM-RU, but protein levels of galectin-9 were too low to compare the differences (data not shown). Therefore, flow cytometric analysis was performed to compare the expression of galectin-9 in MM-BP and MM-RU. We thus found that cytoplasmic level of galectin-9 in MM-BP (43.4% positive) was more evident than that of MM-RU (9.7% positive) (Fig. 2).
Galectin-9 expression on the surface of those melanoma lines was further assessed. MM-BP weakly but significantly expressed galectin-9 on the surface (10.9% positive) (Fig. 2). In contrast, no or little surface galectin-9 expression was detected in MM-RU (1.5% positive) (Fig. 2). From the findings described above, it is likely that the surface but not cytoplasmic galectin-9 plays a role in melanoma cell aggregation.
Effects of galectin-9 on melanoma cell aggregation
Experiments were performed to clarify whether exogenous galectin-9 could induce cell aggregation of MM-RU that proliferated without colony formation (Fig. 1a). Thus, MM-RU (5 × 104/ml) was cultured in the presence or absence of galectin-9 at various concentrations (1 × 10−9 M to 3 × 10−7 M). Microscopic observation was performed at various intervals (2, 6, 12 and 24 hr). MM-RU failed to form cell aggregation during the culture in the absence of galectin-9 at anytime (Fig. 3a). In contrast, MM-RU formed cell aggregation when they were cultured with galectin-9 (3 × 10−8 M) even after 2-hr culture (Fig. 3a). After 6 hr culture, the aggregation became more evident in the cells treated with galectin-9 (Fig. 3a). After 12 hr culture, MM-RU cells began to adhere to the surface of the culture dish (Fig. 3a). MM-RU, however, began to be detached from the surface of culture dish after 24 hr culture (data not shown). Galectin-1 and -3 induced cell aggregation weakly only at higher (greater than 3 × 10−6 M) concentrations (Fig. 3a). Furthermore, lactose but not sucrose suppressed gaelctin-9-induced cell aggregation (data not shown). The cell aggregation induced by exogenous galectin-9 (Fig. 3a), however, looks different from that of MM-BP (Fig. 1b). Taken together, melanoma cells may require some unknown mechanisms to form complete cell aggregation resulting in colony formation though galectin-9 plays a role in cell aggregation.
For quantitative analysis of galectin-9-induced cell aggregation, the total particle number of MM-RU incubated with or without galectin-9 was counted. Figure 3b showed that galectin-9 induced cell aggregation in a dose dependent manner confirming aggregation-inducing activity of galectin-9. Lactose (30 mM) but not sucrose (30 mM) suppressed the galectin-9-induced cell aggregation (Fig. 3b), indicating that galectin-9 plays a role, at least, partly in cell aggregation of melanoma cells through its lectin properties.
Effects of galectin-9 on melanoma cell apoptosis
Experiments were done to clarify whether galectin-9 is associated with apoptosis of melanoma cells. MM-RU was treated with various concentrations of galectin-9 (1 × 10−9 to 1 × 10−6 M) for 24, 48 and 72 hr. Although 24-hr incubation of MM-RU with galectin-9 even at 1 × 10−6 M failed to induce evident apoptosis of MM-RU, 72-hr incubation with galectin-9 could induce evident apoptosis in a dose-dependent manner (Fig. 4a). Galecitn-9-induced apoptosis was suppressed by 30 mM lactose (52.5%) but not 30 mM sucrose (86.9%). Apoptosis mediated by galectin-9 at 1 × 10−6 M was comparable with that mediated by anti-Fas antibody (galectin-9, 93% and anti-Fas antibody, 96.5%). These findings suggested that galectin-9 is involved in not only cell aggregation but also apoptosis. We further confirmed the proapoptotic activity by galectin-9. Figure 4b showed that galectin-9 increased the percentage of the cells in early apoptosis (35.2%).
Expression of galectin-9 protein in melanocytic lesions and clinical significance
High galectin-9 expression (greater than 100 HSCORE) was defined as positive and low expression (100 or less) was defined as negative. As summarized in Table I, galectin-9 positive expression was found in 14 (70%) lesions and negative expression was found in 6 (30%) lesions in melanocytic nevi. In 70 primary melanoma lesions, positive and negative galectin-9 expression was found in 23 (32.9%) and 47 (67.1%) lesions, respectively. In contrast, in 31 metastatic lesions, only 3 (9.7%) lesions were positive and 28 (90.3%) lesions were negative. It is, thus, suggested that galectin-9 expression is down-regulated with malignant transformation in the melanocytic cell lineage (melanocytic nevi vs. primary melanoma, p = 0.0029 and primary melanoma vs. metastatic melanoma, p = 0.0105; Table I, Fig. 5a–c).
Table I. Immunoreactive-Galectin-9 in Melanocytic Lesions and Galectin-9 Expression in Primary Melanoma1
HSCORE was used to determine positive (>100) or negative (100 or less). Averages represented the mean HSCORE ± SD.
Melanocytic nevi (n = 20)
152.0 ± 56.7
Malignant melanoma (n = 101)
Primary lesions (n = 70)
77.7 ± 70.8
Meatastatic lesions (n = 31)
49.7 ± 44.7
Primary melanoma histopathological type
Melanoma in situ (n = 4)
Acral lentiginous melanoma (n = 35)
92.6 ± 77.2
Superficial spreading melanoma (n = 12)
70.0 ± 58.2
Nodular melanoma (n = 7)
125.7 ± 58.6
Mucous melanoma (n = 5)
68.0 ± 36.3
Lentigo maligna melanoma (n = 7)
20.0 ± 42.9
Immunoreactive galectin-9 was mostly detected in the cytoplasm of melanoma cells. In some cases, galectin-9 was also detected in the nucleus, especially in metastatic melanoma cells. In addition, most of the melanoma cells exhibiting nest formation showed more immunoreactive galection-9 in the cytoplasm than those exhibiting no or little nest formation confirming that galectin-9 plays a role in nest formation of melanocytic cells.
According to the histopathological types of primary melanoma lesions (Table I) the highest expression was detected in NM (57%) with 125.7 HSCORE, whereas the lowest expression was found in LMM (0%) with 20 HSCORE and IS (0%). Moreover, an inverse correlation of galectin-9 expression was found with the thickness of primary melanoma lesions as follows. Seventeen of 26 lesions with a thickness less than 1.5 mm including melanoma in situ did not express galectin-9 (average HSCORE 23.1), whereas only 3 of 19 lesions with a thickness between 1.5 mm and 4 mm (average HSCORE 118.9) and no lesions with more than 4 mm (average HSCORE103.2) was negative for galectin-9, respectively. Therefore, prognostic analysis was confined to the primary lesions with a thickness of more than 1.5 mm (44 cases), because clinical significance of galectin-9 expression in early melanoma lesions was masked by surgical procedure. The statistical analysis shown in Figure 6a revealed that low or negative galectin-9 expression in primary melanoma lesions was significantly associated with the presence of regional lymph node metastasis (p = 0.025), recurrence (p = 0.005) and mortality (p = 0.002). Moreover, high galectin-9 expression in primary lesions was correlated with the 5-year disease-free interval and survival time with p-values of 0.032 and 0.041, respectively (Fig. 6b).
Multivariate analysis using Cox's proportional-hazards regression model shown in Table II proved that galectin-9 and tumor thickness were significant factors for prognosis and low galectin-9 expression was higher relative risk for prognosis of melanoma patients than tumor thickness. Values of relative risk of galectin-9 and tumor thickness for disease free survival were 5.967 (p = 0.030) and 5.388 (p = 0.004), respectively and those for over-all survival were 6.128 (p = 0.023) and 5.139 (p = 0.007), respectively. Other factors, including nodal status, tissue type, stage, age and sex were not significant (data not shown).
D-FS, disease-free survival; OS, overall survival; RR, relative risk; CI, confidential interval. Galectin-9 and tumor thickness were significant factors for prognosis, and low galectin-9 expression was higher relative risk for prognosis of melanoma patients than tumor thickness.
Our present study has demonstrated that cultured human melanoma cells express galectin-1, -3 and -9 by RT-PCR, Western blotting and flow cytometric analysis. Although galectin-1 and galectin-3 were expressed similarly in the melanoma cell lines used in the present experiment, the expression of galectin-9 is different according to the melanoma lines (Fig. 1a,b). By flow cytometric analysis, surface galectin-9 was only detected in a colony forming cell line such as MM-BP (Fig. 2). Exogenous galectin-9 (1 × 10−7 M) could induce cell aggregation of MM-RU proliferating without colony formation (Fig. 3a). Furthermore, lactose but not sucrose suppressed the cell aggregation (Fig. 3b). Taken together, it is suggested that surface bound galectin-9 is crucially required for cell aggregation of melanoma cells by its lectin properties. It is of note, however, that the cell aggregation induced by exogenous galectin-9 (Fig. 3a) looks different from that of MM-BP (Fig. 1b). Such difference may reflect that MM-RU requires not only surface galectin-9 but also another substance to form complete cell aggregation like MM–BP though the exact mechanisms of aggregation of melanoma cells should be clarified.
Galectins also play a role in apoptosis. For instance, galectin-1 induces cell apoptosis and galectin-3 prevents apoptosis in T cells and some cancer cells.9–11 It was also reported that galectin-9 has a potential activity for induction of apoptosis in thymocytes and T cells.12 In the present experiments, we have found that MM-BP grow slowly compared to MM-RU (data not shown). Therefore, we raised the hypothesis that galectin-9 is, at least, partially involved in cell growth. Indeed, exogenous galectin-9 apparently induced melanoma cell apoptosis after cell aggregation (Fig. 3a,b). The apoptotic activity of galectin-9 was potent (Fig. 4a,b) and comparable to that of anti-Fas antibody (data not shown). The mechanisms of apoptosis induced by galectin have not been fully understood, although activation of the AP-1 transcription factor and downregulation of Bcl-2 is probably implicated in molecular mechanisms of galectin-1-induced apoptosis.35
We immunohistopathologically examined resected benign and malignant lesions of melanocytic origin and found that loss of immunoreactive galectin-9 was significantly correlated with melanoma progression. Such findings paralleled to the previous reports showing the down regulation of galectin-3 in prostate, colorectal, breast, ovary and head and neck squamous cell carcinoma.36–38 Separated studies have shown that galectin-3 expression is controversially enhanced during the progression of tumors.24, 25, 39, 40 This may reflect the difference of tumor origin, patients or methodology in each study.
We found that high galectin-9 expression in primary melanoma cells was associated with a preferable prognosis in the present experiments, because patients with high galectin-9 expression showed significantly lower lymph node metastasis, recurrence and death than those of the low expression (Fig. 6a). Moreover, high galectin-9 expression in primary lesions was significantly associated with disease free and survival time (Fig. 6b). These findings suggest that expression of galectin-9 may be an additional prognostic factor in primary melanoma lesions. Upregulation of galectin-9 expression in melanoma cells may be a way to interfere with the invasive or spreading process of melanoma cells through the induction of cell aggregation. In addition, galectin-9-induced apoptosis may also contribute for a better prognosis of patients with melanoma.
Although the surface galectin-9 is required for induction of melanoma cell aggregation and apoptosis, the immunohistopathological analysis revealed that cytoplasmic galectin-9 might be associated with a preferable prognosis. Further studies are needed to clarify the functional role of gelactin-9 in the cytoplasm and association of cytoplasmic galectin-9 expression with the surface expression. We also observed a decreased galectin-9 expression during the early stage of melanoma such as melanoma in situ similar to galectin-3 expression in colorectal carcinoma,37 suggesting that the lack of galectin-9 expression is also directly or indirectly involved in the transformation from normal cells to malignant cells.
Recent studies suggest that galectin-9 is an integral membrane protein and acts as a urate transporter in living epithelial cells.41–43 In our experiment, little or no galectin-9 is detected on the surface of MM-RU (Fig. 2). Although MM-BP expresses low but significant surface galectin-9 (Fig. 2), galectin-9 on MM-BP is almost completely shed by lactose but not sucrose (data not shown). Therefore, melanoma cells may not possess galectin-9 as an integral membrane protein differing from epithelial cells though further studies are required to ascertain.
Our present study suggests that galectin-9 could be an important molecule to implement a novel therapeutic strategy through induction of cell aggregation and apoptosis in melanoma cells.
Supported in part by grants-in-aid for Scientific Research from the Ministry of Education, Science and Culture, Japan.