Mucins are large glycoproteins with extensive O-linked glycosylation, typically found on the apical surface of epithelial cells. During carcinogenesis, the normal topology and polarity of epithelial cells both change markedly, and aberrantly glycosylated mucins are expressed on the whole cell surface, thus leading to the shedding of these carcinoma mucins to the bloodstream.1, 2 Sialyl Lewisa, also known as CA19-9, and sialyl Lewisx are sialylated fucosylated oligosaccharides that fall into this category. Both of these carbohydorate determinants present on mucins serve as ligands for the P-, L- and E-selectins expressed on the surface of platelets, leukocytes and vascular endothelial cells.3, 4 As a result, carcinoma cells expressing these carcinoma mucins can interact with platelets, leukocytes and vascular endothelial cells via selectins on their cell surfaces, and this interaction is believed to mediate the adhesion of carcinoma cells to the vasculature, thereby promoting metastasis.5, 6, 7, 8 Indeed, many studies have shown that the augmented expression of sialyl Lewisa and sialyl Lewisx on carcinomas correlate with a poor prognosis because of a high rate of metastasis.9, 10, 11
MUC1 is a transmembrane mucin and its expression is known to increase in lung, breast, ovary and colon carcinomas, thus suggesting that this aberration in the MUC1 expression is a common phenomenon of adenocarcinomas.1, 12, 13 We previously identified a mouse IgG1 monoclonal antibody directed against a lung adenocarcinoma cell line, which recognizes a mucin-like glycoprotein, designated KL-6.14 KL-6 is classified as “Cluster 9 (MUC1 mucin)” of lung tumor and differentiation antigens according to the findings of immunohistochemical and flow cytometry studies.15 Cluster 9 antibodies could contain a carbohydrate specificity not expressed by the MUC1 glycoforms. However, a previous study has shown that purified KL-6 is detectable by an anti-MUC1 core protein antibody, DF-3, thus, KL-6 can be considered as a MUC1-related glycoprotein.16 In addition, KL-6 has been shown to consist of multiple heterogenous submolecules.17, 18, 19 On the basis of the findings of previous studies demonstrating that sialyl Lewisa oligosaccharides are present on MUC1 mucin and interact with E-selectin on the vascular endothelial cells,20, 21, 22 we hypothesized that KL-6/MUC1 mucin also carries selectin ligands and the abundance of KL-6/MUC1 carrying selectin ligands is associated with both an advanced cancer stage and metastatic progression. To investigate this, we developed a sandwich ELISA system using anti-sialyl Lewisa and anti-KL-6/MUC1 antibodies to detect KL-6/MUC1 carrying sialyl Lewisa (designated SLAK), and then measured levels of SLAK in 103 serum samples obtained from the newly diagnosed patients with lung adenocarcinoma. We next assessed the association between its levels and the clinical characteristics and survival of the patients.
Material and methods
Patients and specimens
One hundred and three patients with previously untreated lung adenocarcinoma pathologically diagnosed and subsequently treated at Hiroshima University Hospital (Hiroshima, Japan) between 2001 and 2005 (n = 80) or Ehime University Hospital (Ehime, Japan) between 1994 and 2003 (n = 23), were included in this study. The clinical parameters investigated were as follows: age, sex, Eastern Cooperative Oncology Group (ECOG) performance status, serum KL-6 level and serum CA19-9 level. Each patient was classified on the basis of the tumor, node and metastasis (TNM) classifications of the International Union Against Cancer. The mean age was 63 ± 11 years (mean ± SD, range, 30–78), and 67 (65%) of the patients were male. The organs with metastasis in stage IV lung adenocarcinoma patients were lung (42.4%, 25 out of 59), brain (44.1%, 26 out of 59), bone (27.1%, 16 out of 59), liver (11.1%, 7 out of 59), adrenal gland (10.2%, 6 out of 59) and cervical lymph node (3.4%, 2 out of 59).
Patients with lung adenocarcinoma (n = 103) were treated either by operation (n = 17, 16.5%), chemoradiation (n = 2, 2.0%), radiation (n = 4, 3.9%), platinum-based chemotherapy (n = 40, 38.8%), nonplatinum-based chemotherapy (n = 23, 22.3%), or best supportive care (n = 17, 16.5%). Lung adenocarcinoma patients with distant metastasis (n = 59) were treated either with platinum-based chemotherapy (n = 29, 49.2%), nonplatinum-based chemotherapy (n = 21, 35.6%), or best supportive care (n = 9, 15.2%). The baseline characteristics of the patients are summarized in Table I.
Table I. Association between the Serum Slak Level and Clinical Characteristics in 103 Patients with Lung Adenocarcinoma
Informed consent was obtained from each patient prior to the collection of blood sample. Serum was isolated via centrifugation of blood samples and stored at −80°C until analysis.
We randomly selected 97 sex-matched healthy individuals (63 males and 34 females; mean age ± SD, 61.7 ± 6.5 years; range, 41–76). Healthy individuals with no history or present diagnosis of malignancy, who showed no abnormalities in complete blood cell counts, C-reactive proteins, erythrocyte sedimentation rate, liver function tests, renal function tests, urinalyses, fecal examinations, chest X-rays or electrocardiograms in medical check up, were included as controls.
Measurement of serum KL-6 and CA19-9 concentrations
The concentrations of KL-6 and CA19-9 in the sera were measured with EITEST KL-6 ELISA kit (Eisai, Tokyo, Japan) and CA19-9 ELISA KIT (Alpha Diagnostic international, San Antonio, TX) according to the manufacturer's instructions, respectively.
Human lung adenocarcinoma PC-3, ABC-1 and RERF-LC-KJ cells were purchased from the Health Science Research Resources Bank (Osaka, Japan). PC-3 and ABC-1 cells were cultured in Eagle's minimal essential medium with 10% fetal bovine serum (Biowest, Nuaille, France), and RERF-LC-KJ cells were cultured in RPMI1640 medium with 10% fetal bovine serum. The culture supernatants were collected after 7 days of incubation.
Immunoprecipitation and Western blot analysis
Five micrograms of monoclonal mouse IgG1 anti-KL-6 antibody14 or anti-sialyl Lewisa antibody (Chemicon International, Temecula, CA) were mixed with the culture supernatant (1 ml) of PC-3 cells and incubated on a shaker at 4°C overnight. Next, 20 μl of sepharose beads conjugated with rat-anti-mouse IgG1 antibody (Zymed, San Francisco, CA) were added and again incubated on a shaker at 4°C overnight. The sepharose beads were pelleted by centrifugation (12,500g) and washed 3 times with phosphate buffered saline (PBS). The washed pellets were resuspended in Laemmli buffer for loading on SDS-polyacrylamide gels.
The PC-3 culture supernatant and the immunoprecipitated proteins were resolved on 4% polyacrylamide gels and then were transferred to nitrocellulose membranes electrophoretically. After being blocked in TBST (10 mM Tris/HCl, 150 mM NaCl, 0.05% Tween 20, pH 7.5) with 5% bovine serum albumin (BSA), the membranes were incubated with a monoclonal mouse IgM anti-sialyl Lewisa antibody (Exibio, Prague, Czech Republic) at 4°C overnight. After washed 3 times with TBST, the membranes were incubated with an HRP-conjugated polyclonal goat-anti-mouse IgM antibody (Stressgen, Victoria, Canada) at room temperature for 60 min. Next, ECL reagents (ECL Western Blotting Analysis system; Amersham International, Buckunghamshire, UK) were applied according to the manufacturer's instructions, and the blots were then exposed to ECL-sensitive film (ECL Western Blotting Analysis system; Amersham International, Buckunghamshire, UK).
Enzyme-linked immunosorbent assay for SLAK
Microtiter plates (Maxisorp; Nunc, Roskilde, Denmark) were coated with 100 μl of diluted monoclonal mouse IgM anti-sialyl Lewisa antibody (4 μg/ml in 0.1 M Sodium Carbonate, pH 9.6) at 4°C overnight and then were blocked with 200 μl of PBS containing 1% BSA at 4°C overnight. After washed 3 times with washing buffer (PBS with 0.025% Tween 20 and 0.05% BSA), 100 μl of standards or diluted serum samples in ELISA buffer (PBS containing 1% BSA and 0.1% normal mouse serum; Chemicon international) was added to each well, and incubated at 4°C overnight. After washed 3 times with washing buffer, the plates were incubated with 100 μl of HRP conjugated monoclonal mouse IgG1 anti-KL-6 antibody solutions (Eisai, Tokyo, Japan, diluted 1:8000 in ELISA buffer) at room temperature for 60 min. Thereafter, the plates were washed 3 times with washing buffer and developed with 100 μl of O-phenylenediamine tablet (Sigma-Aldrich, St. Louis, MO) dissolved at 2.5 mg/ml in 0.05 M citrate-phosphate buffer, pH 4.9, containing 0.024% hydrogen peroxide for 20 min. The reaction was stopped by adding 100 μl of 2 N HCl, and optical densities were read at 490 nm. The culture supernatant of PC-3 cells was used as a standard reference sample. One unit per milliliter was defined as the SLAK level, which was calculated from the PC-3 culture supernatant diluted 1:2560. The lower limit of quantitation for this assay was 20 U/ml.
The association of 2 categorical variables was analyzed by Fisher's exact probability test or χ2 distribution. The Spearman rank-order correlation coefficient was used to assess the relation between SLAK and KL-6 levels in the culture supernatants of the adenocarcinoma cell lines. Comparisons of continuous variables were performed with the Mann–Whitney U test. The overall cancer-specific survival was defined from the date of blood sampling to the date of death due to cancer. Survival curves were estimated using the Kaplan-Meier method. The significance of the difference in survival between the 2 groups was estimated by the log-rank test. Several clinical factors were subjected to univariate and multivariate analyses using the Cox proportional-hazard regression model. The relative risk of death was compared using the assessment of the hazard ratio. Differences were considered to be significant when the p values were <0.05.
Identification of KL-6/MUC1 carrying sialyl Lewisa oligosaccharides
To clarify whether sialyl Lewisa oligosaccharides are present on KL-6/MUC1 mucin, Western blot analysis was performed for the culture supernatant of PC-3 cells using anti-sialyl Lewisa antibody and anti-KL-6 antibody. The immunobloting of the untreated PC-3 culture supernatant with anti-sialyl Lewisa antibody revealed a broad smear migrating between 100 and 250 kDa, a protein at larger than 250 kDa, and a very high-molecular-weight protein at larger than 500 kDa (Fig. 1, lane 1). These bands and smears were enhanced when immunoprecipitate of PC-3 culture supernatant with anti-sialyl Lewisa antibody was immunoblotted using anti-sialyl Lewisa antibody (Fig. 1, lane 3). However, only the band migrating at larger than 500 kDa was detected when the immunoprecipitate of PC-3 culture supernatant with anti-KL-6 antibody was immunoblotted using anti-sialyl Lewisa antibody (Fig. 1, lane 2). This Western blot analysis revealed that the very high-molecular weight band at larger than 500 kDa represents KL-6/MUC1 mucin carrying sialyl Lewisa oligosaccharides, designated SLAK.
Levels of SLAK in the culture supernatants of human adenocarcinoma cell lines
As described in Material and methods, we developed a sandwich ELISA system for the measurement of SLAK levels in various specimens. Using this ELISA system, we first attempted to evaluate the SLAK levels in the culture supernatants of the human adenocarcinoma cell lines, PC-3, ABC-1 and RERF-LC-KJ, all of which were confirmed in our laboratory to abundantly express MUC1 mucin (data not shown). To determine whether SLAK levels are correlated with KL-6 levels in these cell lines, KL-6 in the culture supernatants of these cells was also measured. As shown in Table II, SLAK was detectable only in PC-3 cells whereas all of 3 adenocarcinoma cells were found to variously produce KL-6. Particularly, the SLAK level in the culture supernatant of ABC-1 cells expressing KL-6 most among these 3 cell lines was found to be undetectable. In addition, the Spearman rank-order correlation revealed no correlation between the SLAK and KL-6 levels in the culture supernatants of these human lung adenocarcinoma cell lines, (p = 0.8597).
Table II. Levels of Slak in the Supernatant of the Human Adenocarcinoma Cell Lines
Human lung adenocarcinoma cell line
KL-6 level of the supernatant (U/ml)
SLAK level of the supernatant (U/ml)
Spearman rank-order correlation coefficient.
Levels of SLAK in the sera from healthy subjects and patients with lung adenocarcinoma
Using this ELISA system, SLAK levels were measured in the sera obtained from 97 healthy controls and 103 patients with lung adenocarcinoma. Circulating SLAK was detectable in 59 (57%) patients with lung adenocarcinoma and 11 (11%) healthy controls. As shown in Figure 2a, serum SLAK levels in the patients with lung adenocarcinoma were found to be significantly higher than those of healthy controls (p < 0.0001, Mann–Whitney U test). The serum SLAK levels in the patients with stage IA–stage IIIB lung adenocarcinoma were found to be significantly higher than those of healthy controls (p = 0.0035, Mann–Whitney U test, data not shown); however, serum SLAK levels in the lung adenocarcinoma patients with distant metastasis (TNM: stage IV) were found to be further significantly higher in comparison with those of the patients without distant metastasis (TNM: stage IA–stage IIIB) (Fig. 2b, p = 0.0002, Mann–Whitney U test).
Next, we intended to determine a cut-off level for serum SLAK. Because the serum SLAK levels of any healthy subject in the present study did not exceed 50 U/ml (Fig. 2a), we decided to establish this level as the cut-off value. Therefore, 50 U/ml as the threshold of the serum SLAK level was used for the subsequent analyses. When values greater than this cut-off level were defined as high, 34 of the patients with lung adenocarcinoma were thus found to fall into this category.
Clinical characteristics of the patients with lung adenocarcinoma and their correlation with the serum SLAK levels
On the basis of the cut-off value of 50 U/ml for serum SLAK level, the lung adenocarcinoma patients were divided into those with high (≥50 U/ml) and those with low (<50 U/ml) serum SLAK levels, and the association with their clinical characteristics was analyzed using the Fisher's exact probability test (Table I). A high level of serum SLAK in the patients with lung adenocarcinoma was found to significantly correlate with T factor (p = 0.0395), distant metastasis (p = 0.0014), the serum KL-6 level (p < 0.0001) and the serum CA19-9 level (p < 0.0001), but not with age (p = 0.1397), sex (p = 0.9591), ECOG performance status (p = 0.1030), and N factor (p = 0.4920). In addition, a high level of serum SLAK in the patients with stage IV lung adenocarcinoma was not correlated with a number of organs with metastasis (p = 0.9708).
To see whether the treatments subsequently given to the lung adenocarcinoma patients with high serum SLAK levels differed from those given to the patients with low serum SLAK levels, we evaluated the regimen of treatments. The treatments given to the patients with high serum SLAK levels were as follows; operation (2 of 34, 5.9%), platinum-based chemotherapy (18 of 34, 52.9%), nonplatinum-based chemotherapy (10 of 34, 29.4%), or best supportive care (4 of 34, 11.8%). The patients with low serum SLAK levels were subsequently treated either by an operation (15 of 69, 21.7%), chemoradiation (2 of 69, 2.9%), radiation (4 of 69, 5.8%), platinum-based chemotherapy (22 of 69, 31.9%), nonplatinum-based chemotherapy (13 of 69, 18.8%) or best supportive care (13 of 69, 18.8%). Statistically no significant difference was found between the treatments given to the lung adenocarcinoma patients with high and low serum SLAK level (Table I; p = 0.0543, χ2 distribution). In addition, the statistical evaluation of the treatments given to the lung adenocarcinoma patients with distant metastasis did not show any significant differences, between the high and low serum SLAK levels (Table I; p = 0.7178, χ2 distribution).
Next, to assess whether circulating SLAK is valuable as a serum marker, we calculated the sensitivity to separate lung adenocarcinoma patients from healthy subjects. The sensitivity as a diagnostic marker was calculated as 0.33 (34 of 103).
The serum SLAK level and survival rate in the patients with lung adenocarcinoma
The overall survival of the patients with lung adenocarcinoma was evaluated using the Kaplan-Meier method. To confirm that the patient selection was not biased, we first analyzed whether the tumor stage at diagnosis correlated with the overall survival in the studied patients. Tumor stage of the patients was found to be significantly correlated with overall survival (p < 0.0001, log-rank test, data not shown), and, thus, we could make sure that our patient selection had no significant bias. We thereafter investigated the relationship between the serum SLAK level and survival rate in the patients with lung adenocarcinoma. As shown in Figure 3a, the survival rate of the patients with a high serum SLAK level (≥50 U/ml) was significantly poorer than that of the patients with a low serum SLAK level (p < 0.0001, log-rank test). The median overall survival in the patients with a high and low serum SLAK level was 8.0 months and 26.3 months, respectively. Furthermore, the survival analysis restricted to the patients with distant metastasis showed the same trend (Fig. 3b, p = 0.0003, log-rank test). The median overall survival in the stage IV lung adenocarcinoma patients with high and low serum SLAK level was 6.3 months and 14.8 months, respectively. Furthermore, a high serum SLAK level was found to be significantly correlated with a poorer prognosis in the patients without distant metastasis (TNM: stage IA–stage IIIB) irrespective of the subsequent treatments (p = 0.039, log-rank test, data not shown).
Prognostic value of a high serum SLAK level in the patients with lung adenocarcinoma
To determine prognostic importance of clinical characteristics and the serum SLAK level in the patients with lung adenocarcinoma, we performed Cox proportional hazard regression analysis on these parameters. A univariate analysis revealed that high serum SLAK level, T2–4, N1–3, M1, 2–4 of ECOG performance status, and high serum KL-6 level were significant prognostic factors (Table III). Subsequently, a multivariate analysis was performed on these 6 parameters. After adjusting for these covariates, a high serum SLAK level remained significantly associated with an elevated risk of death in comparison with a low level of serum SLAK (Table IV, hazard ratio = 2.39, 95% confidence interval = 1.39–4.13, p = 0.0017). In addition, N1–3, M1 and 2–4 of ECOG performance status were also found to be significant and independent prognostic factors, while the serum KL-6 level was not (Table IV).
Table III. Univariate Cox Analysis of the Overall Survival of the 103 Patients with Lung Adenocarcinoma
95% confidence interval
Serum SLAK (high)
Age (≥70 y)
T factor (2–4)
N factor (1–3)
M factor (1)
Performance status (2–4)
Serum KL-6 (high)
Serum CA19-9 (high)
Table IV. Multivariate Cox Analysis of the Overall Survival of the 103 Patients with Lung Adenocarcinoma
95% confidence intervel
Serum SLAK (high)
T factor (2–4)
N factor (1–3)
M factor (1)
Performance status (2–4)
Serum KL-6 (high)
Soluble MUC1 has been demonstrated in the sera from patients with various MUC1-bearing adenocarcinomas, and it appears to represent a truncated form of membrane bound MUC1.1,, 23,, 24 We have previously reported that MUC1 recognized by anti-KL-6 monoclonal antibody (KL-6/MUC1) can be detected in the sera and tumors from patients with lung adenocarcinoma, suggesting that KL-6/MUC1 is shedding from lung adenocarcinoma cells producing KL-6/MUC1.14, 17, 18 Our previous studies have also shown that KL-6/MUC1 consists of multiple heterogenous submolecules.17, 18, 19 In the present study, we first revealed the presence of KL-6/MUC1 mucin carrying sialyl Lewisa oligosaccharides, designated SLAK, in the culture supernatant of a lung adenocarcinoma cell line, PC-3. This finding confirmed previous observations showing that MUC1 carries sialyl Lewisa epitopes,20, 21, 22 and also suggests that the shedding of SLAK takes place during the proliferation of cancer cells.
Following the establishment of a double sandwich ELISA system to measure the SLAK concentration using anti-sialyl Lewisa and anti-KL-6 antibodies, we first examined the SLAK levels in the culture supernatants of human adenocarcinoma cell lines, PC-3, ABC-1, and RERF-LC-KJ, all of which were confirmed in our laboratory to abundantly express MUC1 mucin. In comparison with the KL-6 levels in the culture supernatants of these cell lines, the SLAK level was not found to correlate with the KL-6 level. SLAK was only detectable in PC-3 cells, whereas all of 3 cell lines variously produced KL-6, indicating that KL-6/MUC1 mucin does not necessarily carry sialyl Lewisa oligosaccharides. This result also suggests that there is a heterogeneity in the productivity of SLAK among human adenocarcinoma cells.
We then examined the serum SLAK levels in 103 patients with lung adenocarcinoma and 97 healthy controls. The serum SLAK levels in the patients with lung adenocarcinoma were found to be significantly higher than those in controls, and the lung adenocarcinoma patients with distant metastasis showed a significantly higher serum SLAK levels in comparison with those without metastatic disease. We established a cut-off level for the serum SLAK based on the maximum value obtained among healthy subjects. Because the sensitivity of high serum SLAK level to separate healthy controls and lung adenocarcinoma patients was not very high (0.33), this level thus does not seem to be highly valuable as a diagnostic marker.
The clinical significance of a high serum SLAK level was further substantiated by its correlation with a shorter overall survival time. Although an elevated serum KL-6/MUC1 level was frequently observed in patients with lung adenocarcinoma or breast cancer,14, 19, 25 its correlation with survival of lung adenocarcinoma patients has not yet been determined. In the present study, a univariate analysis demonstrated that high serum levels of both SLAK and KL-6/MUC1 were significantly correlated with the survival of lung adenocarcinoma patients. In contrast, a high serum level of CA19-9 was not found to correlate with the survival of lung adenocarcinoma patients. In addition, a multivariate analysis revealed a high serum SLAK level but not a high serum KL-6/MUC1 level to be an independent prognostic factor, thus indicating that a high serum SLAK level is a stronger prognostic factor than a high serum KL-6/MUC1 level in patients with lung adenocarcinoma. Furthermore, even when the survival analysis was restricted to the patients with or without distant metastasis, a high serum SLAK level was found to significantly correlate with a poorer prognosis in both cases. These findings suggest that elevated serum SLAK level reflects the aggressiveness of the disease, while it also has a clinically significant ability to predict a poor prognosis in patients with lung adenocaricinoma.
The correlation of SLAK with the presence of distant metastasis and poor survival is possibly explained by its properties associated with the process of metastasis. Sialyl Lewisa oligosaccharide has been reported to serve as a ligand for selectins.3, 4 In fact, a correlation between an increased expression of sialyl Lewisa and a high metastatic potential has been reported in previous studies.26, 27 Recent studies also have shown that the interaction of carcinoma mucins with selectins expressed on the platelets, leukocytes and endothelial cells is strongly involved in the metastatic process.5, 6, 7, 8 Considering that SLAK is sialyl Lewisa oligosaccharides carried by KL-6/MUC1 core protein, these observations strongly suggest the potency of SLAK to facilitate the metastasis of carcinoma cells expressing this molecule. In addition, MUC1 mucin has also been shown to mediate cell adhesion by binding to adhesion molecules such as ICAM-1,28, 29 E-selectin21 and SIGLECS (sialic-acid-binding immunoglobulin superfamily lectins).30 Through these interactions, MUC1 can strengthen the binding of carcinoma cells expressing SLAK to the vasculature and facilitates the stasis and extravasation of the cells, thus resulting in the establishment of metastatic foci. Furthermore, MUC1 on the cell surface has a potential to suppress the aggregation of the cells and cell adhesion to extracellular matrix.31, 32 This property may promote the detachment of the carcinoma cells carrying SLAK from the tumor, thus facilitating the initial step of metastatic process. Judging from these suggested associations of SLAK with metastasis, the cells producing SLAK can thus be considered to possess a high metastatic potency.
In conclusion, the findings in the current study demonstrated the serum SLAK level to be useful for predicting a poor survival in patients with lung adenocarcinoma. These observations can be possibly explained by the potency of SLAK to interact with selectins. The measurement of SLAK is simple to perform and therefore it can be easily applied as a new clinical modality.