The clinical data of 30 patients with early-stage BC who underwent treatment at Kochi Medical School Hospital (Nankoku, Japan) from January 2015 to December 2019 and completed at least 1 year of follow-up were retrospectively reviewed. All 30 patients were women, with a median age of 64 years (range, 40–91 years). The study was reviewed and approved by the Ethical Review Board of Kochi Medical School (approval no. 2023-146; April 19, 2024), which waived the requirement for obtaining individual written informed consent and approved the use of an opt-out procedure as an alternative. All patients underwent surgical treatment, and the resected tissues were formalin-fixed at room temperature (about 20–25°C) in 20% neutral buffered formalin for 12–24 h and then embedded in paraffin blocks. Immunohistochemical (IHC) evaluations of the estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor type 2 (HER2) status were obtained from clinicopathological reports at the time of initial diagnosis. ER and PR positivity was defined as positive staining in ≥1% of tumor cells. HER2 expression was evaluated using a four-tiered scale (0, 1+, 2+, or 3+) in accordance with the 2018 American Society of Clinical Oncology/College of American Pathologists guidelines. Patients with a score of 2+ were re-evaluated by in situ hybridization, and HER2 positivity was defined as a score of 3+ or a positive in situ hybridization result.

A total of 9 tissue samples obtained from patients who developed metastases or recurrence within the 1-year follow-up period were sectioned into 4-µm-thick slices and subjected to heat-induced antigen retrieval using ULTRA Cell Conditioning 1 Solution (Roche Tissue Diagnostics) at 95°C for 30 min. IHC staining was performed with an anti-KL-6 mouse monoclonal antibody (cat. no. LNM-KR-067; Cosmo Bio Co., Ltd.) using the Ventana automated system at a dilution of 1:2,000, with incubation at 4°C overnight (18 h). A total of 21 consecutive tissue samples from patients without metastasis or recurrence were processed identically and served as controls. KL-6 expression was evaluated using the Allred scoring system (13,14). Briefly, the proportion of positively stained cells was scored as follows: 0 (none), 1 (<1%), 2 (1–10%), 3 (11–33%), 4 (34–66%) and 5 (≥66%). Staining intensity in the most predominant areas was rated as 0 (none), 1 (weak), 2 (intermediate), or 3 (strong). The Allred score (range: 0–8 points) was calculated by summing the proportion and intensity scores. Immunostaining assessments were independently performed by two experienced pathologists (Mitsuko Iguchi and Makoto Toi).

For immunocytochemical staining, MCF-7 and MDA-MB-231 cells (Health Science Research Resources Bank), each at a density of 4×10⁴ cells/ml, were cultured in slide chambers, fixed with 4% paraformaldehyde (PFA; Nacalai Tesque, Inc.) at 4°C for 30 min, and washed with phosphate-buffered saline (PBS). The cells were then incubated overnight at 4°C with an anti-KL-6 mouse monoclonal antibody diluted at 1:2,000, followed by incubation with fluorescein isothiocyanate-conjugated anti-mouse immunoglobulin (IgG) antibodies (cat. no. F-2761; 1:200; Molecular Probes; Thermo Fisher Scientific, Inc.) for 1 h at room temperature. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; 0.001 mg/ml; MilliporeSigma). Fluorescence images were acquired using a BX53 fluorescence microscope (Olympus Corporation).

Total protein was extracted from cells using a radioimmunoprecipitation assay buffer (FUJIFILM Wako Pure Chemical Corporation). Protein concentrations were measured using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Inc.). Equal amounts of protein (20 µg per lane) were separated by 10% sodium dodecyl-sulfate polyacrylamide gel electrophoresis (Bio-Rad Laboratories, Inc.) and transferred to polyvinylidene difluoride membranes using the Trans-Blot Turbo Transfer System (Bio-Rad Laboratories, Inc.). Membranes were blocked at room temperature for 1 h with Blocking One (cat. no. 03953-93; Nacalai Tesque, Inc.) and subsequently incubated overnight incubation at 4°C with an anti-KL-6 mouse monoclonal antibody (1:2,000) and a glyceraldehyde 3-phosphate dehydrogenase mouse monoclonal antibody (1:2,000; cat. no. 20035, ProMab Biotechnologies, Inc.). Subsequently, membranes were incubated for 1 h at room temperature with a horseradish peroxidase (HRP)-conjugated anti-mouse polyclonal antibody (1:2,000; cat. no. P0447; Dako, Agilent Technologies, Inc.). Protein bands were visualized using Enhanced Chemiluminescence Prime Western Blotting Detection Reagents (Amersham; Cytiva) and detected with an LAS-4000 Lumino-Image Analyzer (FUJIFILM Wako Pure Chemicals Corporation).

Wound healing assays were performed as previously described (15). Trypsinized MCF-7 cells were counted using a C-Chip™ disposable hemocytometer (SKC, Inc.; Thermo Fisher Scientific, Inc.), seeded at a density of 2×10⁴ cells/ml into 35-mm diameter dishes, and incubated at 37°C for 48 h. Upon reaching ~90% confluency, three linear scratches were created in the cell monolayer of each dish using a 20-µl pipette tip. Phase-contrast images were captured at 0, 12, 24 and 36 h post-scratch at 50 randomly marked locations per 35-mm dish using an Olympus CKX41 microscope (original magnification, ×200; Olympus Corporation). The scratch areas devoid of cell migration were quantified using the ImageJ software version 1.53 (National Institutes of Health). The average migration across the 50 sites per dish was calculated. The relative cell migration rates at each time point were normalized to the 0-h measurement, which was set as 1.0.

MDA-MB-231 cells (1×10⁴ cells/ml) were prepared as described in Fluorescent immunocytochemical staining. Cellmatrix (Nitta Gelatin Inc.) was prepared according to the manufacturer's instructions. Each cell group was mixed with 500 µl of Cellmatrix^® and plated into 35-mm dishes pre-coated with Cellmatrix, followed by incubation at 37°C for 30 min. The gels were overlaid with 200 µl of Dulbecco's Modified Eagle Medium (DMEM; MilliporeSigma) and cultured at 37°C for 10 days. After incubation, the gels were fixed overnight in 20% buffered formaldehyde at room temperature and embedded in paraffin. Hematoxylin-eosin staining was performed using Mayer's hematoxylin (10 min) and 1% eosin (5 min) at room temperature. IHC staining for KL-6 and HIF-1α was subsequently performed. Paraffin-embedded tissue sections (4-µm thick) were subjected to antigen retrieval by immersion in Immunosaver (Nisshin EM Co., Ltd.) at 98°C for 30 min. Endogenous peroxidase activity was quenched by incubation in 0.3% hydrogen peroxide in methanol for 10 min at room temperature. Then, sections were blocked with Blocking One for 1 h at room temperature and incubated overnight at 4°C with the following antibodies: anti-KL-6 mouse monoclonal antibody (1:2,000) and anti-HIF-1α mouse monoclonal antibody (1:100; cat. no. AP2054a, Abgent, Inc.). After washing with PBS, the sections were incubated for 1 h at room temperature with Universal ImmPRESS HRP Reagent (MP-7500; Vector Laboratories, Inc.), followed by additional PBS washes. Color development was achieved using the DAB Peroxidase Substrate Kit (SK-4100; Vector Laboratories, Inc.), and the nuclei were counterstained with Mayer's hematoxylin for 1 min at room temperature. Optical microscopic images were captured using an Olympus BX53 microscope equipped with the cellSens imaging system (Olympus Corporation).

A total of 1×10⁴ parental, MCF-7 and MDA-MB-231 cells per well were counted as aforementioned and seeded into PrimeSurface^® 96-well ultra-low attachment round-bottom plates (Sumitomo Bakelite Co., Ltd.). The cells were cultured at 37°C in a 5% CO₂ atmosphere for 6 days to form multicellular spheroids. Following culture, spheroids were fixed and subjected to IHC staining for KL-6 and HIF-1α, as aforementioned and observed under an Olympus BX53 microscope.

Hypoxia-induced changes in KL-6 protein expression were evaluated in MDA-MB-231 cells using immunocytochemistry. Cells (1×10⁴) were cultured for 2 days in three 2-well slide chambers (cat. no. 192-002; Watson Co., Ltd.). In each chamber, one well was treated with cobalt chloride (CoCl₂) to induce hypoxia, whereas the other well served as an untreated control. The cells were incubated in DMEM supplemented with CoCl₂ (2 µmol/l; cat. no. 15862-1ML-F; MilliporeSigma) at 37°C for 8 h. Following incubation, the cells were washed twice with PBS and fixed in 4% PFA at 4°C for 30 min. Each of the three slides was then assigned to a specific assay: One for KL-6 immunocytochemical staining, one for HIF-1α immunocytochemical staining, and one for hypoxia detection. Hypoxia was visualized using the LOX-1 hypoxia probe (cat. no. NC-LOX-1S; MBL Co., Ltd.). Briefly, after the 8-h incubation, LOX-1 was added to the DMEM at a concentration of 10 µl/ml. The cells were subsequently fixed again in 4% PFA (4°C for 30 min), washed twice with PBS, counterstained with DAPI, and examined under a fluorescence microscope.

Immunoelectron microscopy was performed as previously described (16). MDA-MB-231 cells and spheroids were cultured on chamber slides or PrimeSurface 96-well ultra-low attachment plates and fixed with 4% PFA containing 0.01% glutaraldehyde (Wako Pure Chemical Industries) at room temperature for 10 min. Endogenous peroxidase activity was blocked by incubating the samples in methanol containing 0.3% H₂O₂ for 10 min. After washing with PBS, non-specific binding was blocked by incubation in PBS supplemented with 10% normal goat serum for 10 min. The samples were then incubated overnight at 4°C with a mouse monoclonal antibody against KL-6. After washing with PBS, the slides were incubated at room temperature for 60 min with the ImmPRESS detection reagent (Vector Laboratories, Inc.). The peroxidase reaction was visualized using DAB substrate solution (Vector Laboratories, Inc.) for 10 min. After several rinses with distilled water, the slides were treated with 0.01% aqueous hydrogen tetrachloroaurate(III) tetrahydrate (HAuCl₄·3H₂O; MilliporeSigma) for 10 min, rinsed again in distilled water, and dried using an air blower. The slides were then incubated in a humid chamber at 37°C for 15 h, followed by air-drying. As a control, parallel slides were incubated under dry air conditions. All samples were examined under a JSM-6010LV microscope (JEOL, Ltd.).

All quantitative data were obtained from at least three independent experiments and are expressed as the mean ± standard deviation. To compare KL-6-positive cell rates between the MCF-7 and MDA-MB-231 cell lines, Levene's test was initially used to assess the homogeneity of variances. As unequal variances were observed, Welch's t-test was applied for statistical comparison. Western blotting was performed to evaluate the KL-6 expression in MCF-7 cells treated with or without CoCl₂, a chemical hypoxia-inducing agent. The Shapiro-Wilk test confirmed normal distribution in both groups, whereas Levene's test revealed unequal variances; therefore, Welch's t-test was used for statistical comparison.

In the wound healing assay, the relative area covered by cells was measured at 0 and 36 h post-treatment. As the data from both treatment groups (KL-6 neutralizing antibody and non-immune IgG control) were not normally distributed, according to the Shapiro-Wilk test, comparisons were performed using the non-parametric Mann-Whitney U test. All statistical analyses were performed using the R software (version 4.4.3, http://www.R-project.org/) with a significance level set at α=0.05. P<0.05 was considered to indicate a statistically significant difference.

The association between BC recurrence or metastasis and the IHC expression of KL-6 and HIF-1α in surgical pathology specimens was investigated. Among the 30 patients, eight experienced recurrence or metastasis, whereas 20 showed no evidence of disease progression (Table SI).

Expression levels were quantified using the Allred scoring system (range, 0–8). Based on the observed score distribution, the top three score categories were defined as ‘high’ expression, and the remaining categories were classified as ‘low’ expression.

For KL-6, 7/9 patients in the recurrence/metastasis group were classified as having high expression, whereas 2/9 were classified as having low expression. This pattern is suggestive but should be interpreted with caution given the small sample (Fig. 1A-C).

For HIF-1α, 4/9 patients in the recurrence/metastasis group were classified as having high expression, whereas 5/9 were classified as having low expression. In the non-recurrence/metastasis group, 1/21 patients were classified as having high expression, whereas 20/21 were classified as having low expression. These counts indicate a numerical enrichment of higher HIF-1α expression in patients with recurrence or metastasis (Fig. 1D-F).

A single-center analysis was conducted with a limited number of events (n=9), yielding an events-per-variable (EPV) below commonly accepted thresholds for stable multivariable modeling. To prevent overfitting and unstable estimates, statistical analyses, including multivariable modeling and hypothesis testing, were not performed; accordingly, inferential statistics (for example, P-values or confidence intervals) were not reported in this subsection. The findings are therefore descriptive and hypothesis-generating and require validation in larger, multicenter cohorts with adequate EPV.

KL-6 protein expression in the BC cell lines was evaluated using immunofluorescence staining and western blot analysis (Fig. 2). Immunofluorescence staining (Fig. 2A) demonstrated that the majority of MCF-7 cells exhibited KL-6-positive signals (Fig. 2Aa), whereas only a small proportion of MDA-MB-231 cells were positive for KL-6 expression (Fig. 2Ab). Subcellular localization analysis revealed that KL-6 in MCF-7 cells was present as punctate staining along the cell membrane (Fig. 2Ac). By contrast, KL-6 localization in MDA-MB-231 cells was primarily observed at the membrane edges and cellular protrusions (Fig. 2Ad).

Western blot analysis (Fig. 2B) demonstrated markedly lower KL-6 expression in MDA-MB-231 cells compared with MCF-7 cells. Quantitative analysis (Fig. 2C) confirmed a higher proportion of KL-6-positive cells in MCF-7 than in MDA-MB-231 cells.

Prior to group comparison, data distribution was assessed using the Shapiro-Wilk test, which confirmed normality in both cell lines. Levene's test indicated unequal variances; therefore, Welch's t-test was applied. A significantly higher percentage of KL-6-positive cells was observed in MCF-7 cells compared with MDA-MB-231 cells (P<0.01).

To elucidate the role of KL-6 in cell migration, a wound healing assay was conducted. MCF-7 cells were selected for this analysis due to their markedly higher KL-6 expression compared with MDA-MB-231 cells, as demonstrated using western blot analysis.

Following the creation of a linear wound in the cell monolayer, two treatment groups were established: cells treated with a KL-6 neutralizing antibody and control cells treated with non-immune IgG. Wound closure was assessed 36 h after treatment. Substantial wound closure was observed in the control group, whereas closure was notably inhibited in the KL-6 antibody-treated group (Fig. 3A).

For quantitative analysis, the relative wound area covered at 0 and 36 h post-treatment was calculated and compared between the groups. The Shapiro-Wilk test indicated that data from both groups were not normally distributed; therefore, the non-parametric Mann-Whitney U test was employed. The analysis demonstrated that the KL-6 neutralizing antibody-treated group exhibited a significantly smaller relative covered area compared with the control group, indicating reduced migratory capacity (P<0.05; Fig. 3B).

These findings suggested that KL-6 positively regulates MCF-7 cell migration, supporting its role as a functional contributor to BC cell invasiveness.

A three-dimensional spheroid culture model was used to evaluate KL-6 expression in the BC cell lines MCF-7 and MDA-MB-231 through IHC and western blot analysis (Fig. 4).

IHC analysis (Fig. 4A) demonstrated that KL-6 positivity in MCF-7 spheroids was predominantly localized to the peripheral regions of the spheroid structure (Fig. 4Aa and c). By contrast, strong KL-6 positivity in MDA-MB-231 spheroids was observed throughout the entire spheroid, including the central regions (Fig. 4Ab and d).

Quantitative assessment by western blotting (Fig. 4B) demonstrated higher KL-6 protein levels (Fig. 4B) in MDA-MB-231 spheroids compared with MCF-7 spheroids. These findings indicated that KL-6 expression was markedly upregulated in MDA-MB-231 cells under three-dimensional culture conditions.

To further examine the relationship between KL-6 expression and hypoxia, the MDA-MB-231 spheroids were analyzed at various time points (Fig. 5). IHC analysis demonstrated a mosaic-like pattern of KL-6 expression 2 days after seeding (Fig. 5A); by day 6, the majority of cells exhibited KL-6 positivity (Fig. 5B). Nuclear expression of HIF-1α, a marker of hypoxia, was detected on day 6 (Fig. 5C). Furthermore, hypoxia probe-based detection demonstrated strong fluorescence signals within the spheroid core at this time point (Fig. 5D). These findings suggested that hypoxia develops progressively during spheroid culture and is associated with increased KL-6 expression in MDA-MB-231 cells.

KL-6 expression was subsequently assessed in MDA-MB-231 cells and spheroids embedded within a three-dimensional extracellular matrix (Fig. 6). Immunofluorescence analysis demonstrated that most embedded cells exhibited nuclear localization of HIF-1α, as evidenced by co-localization with DAPI staining (Fig. 6A and B). IHC analysis further revealed prominent KL-6 expression in cellular protrusions extending from the spheroids into the surrounding matrix (Fig. 6C and D). These results suggested that KL-6 contributes to the invasive behavior of MDA-MB-231 cells under hypoxic conditions within a three-dimensional matrix environment.

To examine the effect of hypoxic conditions on KL-6 expression in BC cells, CoCl₂, a chemical hypoxia inducer, was used to treat the cells, and changes in the expression of the hypoxia markers LOX-1 and HIF-1α, as well as KL-6, were evaluated by immunofluorescence staining (Fig. 7A).

MCF-7 cells were incubated with CoCl₂ at concentrations of 2 and 4 µmol/l for 6 h at 37°C. Immunofluorescence analysis demonstrated that the fluorescence intensities of LOX-1 (Fig. 7Aa and b) and HIF-1α (Fig. 7Ac and d) were markedly increased in the CoCl₂-treated groups, indicating effective induction of hypoxic conditions. KL-6 fluorescence was also significantly enhanced in CoCl₂-treated cells (Fig. 7Af) compared with the untreated controls (Fig. 7Ae).

Quantitative analysis of the proportion of KL-6-positive cells relative to the total cell population revealed a significant increase in the CoCl₂-treated group compared with the untreated group (P<0.001), as determined by Welch's t-test (Fig. 7B).

To elucidate the ultrastructural localization of KL-6, immunogold labeling combined with low-vacuum scanning electron microscopy was performed on MDA-MB-231 cells and spheroids (Fig. 8).

Immunoelectron micrographs of MDA-MB-231 monolayer cells and spheroids (Fig. 8A and C) demonstrated KL-6-positive gold particles prominently localized on the cell membrane, within cellular protrusions, and at sites of cell-cell contact. Surface imaging of spheroids (Fig. 8B and D) further demonstrated the accumulation of KL-6-associated gold particles in regions enriched with protrusive structures.