Oral squamous carcinoma cell lines including HSC-2 (oral cavity) and SAS (tongue) were obtained from the Japanese Collection of Research Bioresources Cell Bank (JCRB; Osaka, Japan). Chinese hamster ovary (CHO)-K1 was obtained from the American Type Culture Collection (ATCC). CD44v3-10 plus N-terminal PA16 tag-overexpressed CHO-K1 (CHO/PA16-CD44v3-10) was generated by transfection of pCAG/PA16-CD44v3-10 to CHO-K1 cells using the Neon Transfection System (Thermo Fisher Scientific, Inc.). The PA16 tag consists of 16 amino acids (GLEGGVAMPGAEDDVV) (27). HSC-2 and SAS cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Nacalai Tesque, Inc.), and CHO-K1 and CHO/PA16-CD44v3-10 were cultured in RPMI-1640 medium (Nacalai Tesque, Inc.), supplemented with 10% heat-inactivated fetal bovine serum (FBS; Thermo Fisher Scientific Inc.), 100 units/ml of penicillin, 100 µg/ml streptomycin, and 0.25 µg/ml amphotericin B (Nacalai Tesque, Inc.) at 37°C in a humidified atmosphere containing 5% CO₂.

Mouse anti-CD44s mAb C₄₄Mab-5 (IgG₁, kappa) was developed as previously described (27). Mouse IgG was purchased from Sigma-Aldrich Corp. (Merck KGaA). To generate recombinant C₄₄Mab-5 (recC₄₄Mab-5), cDNAs of C₄₄Mab-5 heavy and light chains were subcloned into pCAG-Neo and pCAG-Ble vectors (FUJIFILM Wako Pure Chemical Corporation), respectively. To generate 5-mG_2a-f, appropriate V_H cDNA of mouse C₄₄Mab-5 and C_H of mouse IgG_2a were subcloned into pCAG-Neo vector, and light chain of C₄₄Mab-5 was subcloned into pCAG-Ble vector. Vectors were transfected into ExpiCHO-S or BINDS-09 (FUT8-deficient ExpiCHO-S cells) using the ExpiCHO Expression System (28). recC₄₄Mab-5 and 5-mG_2a-f were purified using Protein G-Sepharose (GE Healthcare Bio-Sciences).

All animal experiments were performed in accordance with relevant guidelines (e.g. ARRIVE guidelines) and regulations (e.g. 3R regulations) to minimize animal suffering and distress in the laboratory (29,30). Seventy female BALB/c nude mice (6 weeks old, 15–18 g) were purchased from Charles River (Kanagawa, Japan). Animal studies for ADCC and antitumor activity were approved by the Institutional Committee for Experiments of the Institute of Microbial Chemistry (permit number: 2020-003). Mice were maintained in a pathogen-free environment (23±2°C, 55±5% humidity) on 11 h light/13 h dark cycle with food and water supplied ad libitum during the experimental period. Mice were monitored for health and weight every 1 or 5 days. Experiment duration was three weeks. A bodyweight loss exceeding 25% and a maximum tumor size exceeding 3,000 mm³ were identified as humane endpoints. Mice were euthanized by cervical dislocation, and the death was verified by respiratory arrest and cardiac arrest.

C₄₄Mab-5 and 5-mG_2a-f were immobilized on Nunc Maxisorp 96-well immunoplates (Thermo Fisher Scientific Inc.) at 1 µg/ml for 30 min. After blocking using SuperBlock buffer (Thermo Fisher Scientific Inc.) containing 0.5 mM CaCl₂, the plates were incubated with biotin-labeled lectins, such as Aleuria aurantia lectin (AAL; Vector Laboratories), Pholiota squarrosa lectin (PhoSL; J-OIL MILLS, Inc.) (31), and concanavalin A (ConA; Vector Laboratories), followed by 1:3,000 diluted peroxidase-conjugated streptavidin (Agilent Technologies). The enzymatic reaction was produced using a 1-Step Ultra TMB-ELISA (Thermo Fisher Scientific Inc.). The optical density was measured at 655 nm using an iMark microplate reader (Bio-Rad Laboratories, Inc.).

Cells were harvested by brief exposure to 0.25% trypsin/1 mM ethylenediaminetetraacetic acid (EDTA; Nacalai Tesque, Inc.). After washing with 0.1% bovine serum albumin (BSA) in phosphate-buffered saline (PBS), cells were treated with primary mAbs for 30 min at 4°C and subsequently with Alexa Fluor 488-conjugated anti-mouse IgG (1:1,000; Cell Signaling Technology, Inc.). Fluorescence microscopy data were collected using an EC800 Cell Analyzer (Sony Corp.).

Histologic sections (4-µm thick) of an oral cancer tissue microarray (catalogue number: OR481; US Biomax Inc.) were directly autoclaved in citrate buffer (pH 6.0; Agilent Technologies Inc.) for 20 min. Sections were then incubated with 1 µg/ml primary mAbs for 1 h at room temperature and treated using an Envision+ Kit (Agilent Technologies) for 30 min. Color was developed using 3,3′-diaminobenzidine tetrahydrochloride (DAB; Agilent Technologies Inc.) for 2 min, and sections were then counterstained with hematoxylin (FUJIFILM Wako Pure Chemical Corporation). Hematoxylin and eosin (H&E) staining (FUJIFILM Wako Pure Chemical Corporation) was performed using consecutive tissue sections. Leica DMD108 (Leica Microsystems GmbH) was used to examine the sections and obtain images.

Cells were suspended in 100 µl of serially diluted mAbs (0.3 ng/ml-5 µg/ml), followed by the addition of Alexa Fluor 488-conjugated anti-mouse IgG (1:200; Cell Signaling Technology, Inc.). Fluorescence microscopy data were collected using an EC800 Cell Analyzer (Sony Corp.). The dissociation constant (K_D) was calculated by fitting binding isotherms to built-in one-site binding models in GraphPad PRISM 8 (GraphPad Software, Inc.).

Cell lysates (10 µg) were boiled in sodium dodecyl sulfate (SDS) sample buffer (Nacalai Tesque, Inc.). Proteins were separated on 5–20% polyacrylamide gels (FUJIFILM Wako Pure Chemical Corporation) and transferred onto polyvinylidene difluoride (PVDF) membranes (Merck KGaA). After blocking with 4% skim milk (Nacalai Tesque, Inc.) in PBS with 0.05% Tween 20, the membranes were incubated with 10 µg/ml of an anti-CD44 mAb [clone C₄₄Mab-46 (mouse IgG₁, kappa)]; available from Antibody Bank of Tohoku University (ABTU; Miyagi, Japan); http://www.med-tohoku-antibody.com/topics/001_paper_antibody_PDIS.htm#antiCD44) or 1 µg/ml of anti-β-actin (clone AC-15; cat. no. A5441; Sigma-Aldrich Corp.; Merck KGaA). This was followed by incubation with peroxidase-conjugated anti-mouse immunoglobulins (Agilent Technologies Inc.). Finally, protein bands were detected with ImmunoStar LD (FUJIFILM Wako Pure Chemical Corporation) using a Sayaca-Imager (DRC Co., Ltd.).

Total RNAs were prepared from cell lines using an RNeasy Mini Prep Kit (Qiagen Inc.). The initial cDNA strand was synthesized using SuperScript IV Reverse Transcriptase (Thermo Fisher Scientific, Inc.) by priming nine random oligomers and an oligo(dT) primer according to the manufacturer's instructions. We performed 35 cycles of PCR for amplification using HotStarTaq DNA Polymerase (Qiagen Inc.) with 0.2 µM of primer sets: Human CD44 sense (5′-GAAAGGAGCAGCACTTCAGG-3′), human CD44 antisense (5′-ACTGCAATGCAAACTGCAAGC-3′), GAPDH sense (5′-CAATGACCCCTTCATTGACC-3′), and GAPDH antisense (5′-GTCTTCTGGGTGGCAGTGAT-3′).

Six six-week-old female BALB/c nude mice were purchased from Charles River (Kanagawa, Japan). After euthanization by cervical dislocation, spleens were removed aseptically and single-cell suspensions obtained by forcing spleen tissues through a sterile cell strainer (352360, BD Falcon, Corning, Inc.) using a syringe. Erythrocytes were lysed with a 10-sec exposure to ice-cold distilled water. Splenocytes were washed with DMEM and resuspended in DMEM with 10% FBS and used as effector cells. Target cells were labeled with 10-µg/ml Calcein AM (Thermo Fisher Scientific, Inc.) and resuspended in the same medium. The target cells (2×10⁴ cells/well) were plated in 96-well plates and mixed with effector cells, anti-CD44s antibodies, or control IgG (mouse IgG_2a) (Sigma-Aldrich Corp.; Merck KGaA). After a 5-h incubation, the Calcein AM release of supernatant from each well was measured. Fluorescence intensity was determined using a microplate reader (Power Scan HT) (BioTek Instruments) with an excitation wavelength of 485 nm and an emission wavelength of 538 nm. Cytolytic activity (as % of lysis) was calculated using the equation: % lysis=(E-S)/(M-S) ×100, where E is the fluorescence of the combined target and effector cells, S is the spontaneous fluorescence of the target cells only, and M is the maximum fluorescence measured after lysing all cells with a buffer containing 0.5% Triton X-100, 10 mM Tris-HCl (pH 7.4), and 10 mM of EDTA.

Cells in DMEM supplemented with 10% FBS were plated in 96-well plates (2×10⁴ cells/well), and incubated for 5 h at 37°C with either anti-CD44s antibodies or control IgG (mouse IgG_2a) (Sigma-Aldrich Corp.; Merck KGaA) and 10% rabbit complement (Low-Tox-M Rabbit Complement; Cedarlane Laboratories). To assess cell viability, an MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium; inner salt] assay was performed using a CellTiter 96^®AQueous assay kit (Promega Corp.).

3D cell proliferation was measured with the CellTiter-Glo^® 3D cell viability assay (Promega Corp.) according to the manufacturer's instructions. Briefly, the cells were plated (2,000 cells/100 µl/well) in triplicate in 96-well ultra low attachment plates (Corning Inc.) with PBS or 100 µg/ml of mouse IgG_2a and an anti-CD44 mAb (5-mG_2a-f) in DMEM containing 10% FBS. The cell viability was measured after 48 h of incubation. The CellTiter-Glo^® 3D reagent was added into wells in a 1:1 dilution (100 µl volume in well:100 µl of reagent) and then the plates were shaken for 5 min on an orbital shaker and incubated at room temperature for an additional 25 min. The luminescent signal was read using an EnSpire multi-plate reader (Perkin Elmer). Images were taken using an Evolution MP camera (Media Cybernetics). The proliferation rate was calculated relative to the control (PBS was added instead of the antibodies).

Sixty-four six-week-old female BALB/c nude mice were purchased from Charles River (Kanagawa, Japan) and used at 10 weeks of age. HSC-2 and SAS cells (0.3 ml of 1.33×10⁸ cells/ml in DMEM) were mixed with 0.5 ml BD Matrigel Matrix Growth Factor Reduced (BD Biosciences). One hundred microliters of this suspension (5×10⁶ cells) was injected subcutaneously into the left flank. After day 1 (protocol-1) or day 7 (protocol-2), 100 µg of 5-mG_2a-f and control mouse IgG (Sigma-Aldrich Corp.; Merck KGaA) in 100 µl PBS were injected intraperitoneally (i.p.) into treated and control mice, respectively. Additional antibodies were then injected on days 7 and 14 (protocol-1) or on days 14 and 21 (protocol-2). Nineteen days (protocol-1) or 27 days (protocol-2) after cell implantation, all mice were euthanized by cervical dislocation and tumor diameters and volumes were determined as previously described (32).

All data are expressed as mean ± standard error of the mean (SEM). Statistical analysis was carried out using ANOVA following Tukey-Kramer's test for ADCC and CDC. Sidak's multiple comparisons test was used for tumor volume and mouse weight, or Welch's t test for tumor weight and 3D cell proliferation assay using GraphPad Prism 7 (GraphPad Software, Inc.). P<0.05 was adopted as a level of statistical significance.

As mouse IgG_2a possesses high ADCC and CDC activities (33), we first produced a mouse IgG_2a version of mouse IgG₁ C₄₄Mab-5 by subcloning appropriate V_H cDNA of C₄₄Mab-5 and C_H of mouse IgG_2a into pCAG-Neo vector, and light chain of C₄₄Mab-5 into pCAG-Ble vector. This IgG_2a-type of C₄₄Mab-5 is henceforth referred to as 5-mG_2a. We additionally produced a core-fucose-deficient type of 5-mG_2a, henceforth referred to as 5-mG_2a-f, using the BINDS-09 cell line (FUT8-knockout Expi-CHO-S cell line) (28). Defucosylation of 5-mG_2a-f was confirmed using lectins such as Aleuria aurantia lectin (AAL, fucose binder) (34) and Pholiota squarrosa lectin (PhoSL, core fucose binder) (31). Concanavalin A (ConA, mannose binder) (35) was used as a control. Both C₄₄Mab-5 and 5-mG_2a-f were detected using ConA (Fig. 1A). C₄₄Mab-5, but not 5-mG_2a-f, was detected using AAL (Fig. 1B) or PhoSL (Fig. 1C), indicating that 5-mG_2a-f was defucosylated.

We examined the sensitivity of 5-mG_2a-f in CHO cells expressing CD44v3-10 plus N-terminal PA16 tag (CHO/PA16-CD44v3-10) and in oral squamous cell carcinoma (OSCC) cell lines (SAS and HSC-2) using flow cytometry. Both C₄₄Mab-5 and 5-mG_2a-f reacted with CHO/PA16-CD44v3-10 cells (Fig. 2A), but not with CHO-K1 cells (Fig. 2B). Both C₄₄Mab-5 and 5-mG_2a-f reacted with SAS cells (Fig. 2C) and HSC-2 cells (Fig. 2D), indicating that both mAbs showed high sensitivity against SAS and HSC-2 cells.

As shown in Fig. S1A, CD44 was not detected by an anti-CD44s mAb (C₄₄Mab-46) in both SAS and HSC-2 cells presumably because the CD44 expression level in those cells might be low for the detection in western blot analysis. Then, we performed RT-PCR analysis for detection of CD44. As shown in Fig. S1B, the multiple bands of CD44v were detected in SAS and HSC-2 cells using PCR, indicating that CD44v is expressed in SAS and HSC-2 cells.

Next, we performed immunohistochemical analysis on oral cancer cell lines. Representative images are shown in Fig. 3. Both C₄₄Mab-5 (Fig. 3A and B) and 5-mG_2a-f (Fig. 3C and D) stained the plasma membrane of oral cancer cells. The sensitivity of 5-mG_2a-f was similar with that of C₄₄Mab-5. Both C₄₄Mab-5 and 5-mG_2a-f stained 33/38 cases (86.8%) of OSCCs of the tissue microarray. Hematoxylin & eosin (H&E) staining of consecutive tissue sections of OSCC is shown in Fig. 3E and F.

We performed a kinetic analysis of the interactions of recC₄₄Mab-5 and 5-mG_2a-f with SAS and HSC-2 oral cancer cell lines using flow cytometry. As shown in Fig. 4A, the dissociation constant (K_D) for recC₄₄Mab-5 in SAS cells was 2.4×10⁻¹⁰ M. In contrast, the K_D for 5-mG_2a-f in SAS cells was 2.8×10⁻¹⁰ M (Fig. 4A). The binding affinity of 5-mG_2a-f in SAS cells was very similar to that of recC₄₄Mab-5. Likewise, the K_D for recC₄₄Mab-5 against HSC-2 was 2.3×10⁻⁹ M (Fig. 4B). In contrast, the K_D for 5-mG_2a-f in HSC-2 cells was 2.6×10⁻⁹ M (Fig. 4B). The binding affinity of 5-mG_2a-f in HSC-2 cells was very similar to that of recC₄₄Mab-5. The binding affinity of 5-mG_2a-f in SAS was 9.3-fold higher than that against HSC-2.

Because the mouse IgG₁ subclass of C₄₄Mab-5 does not possess ADCC or CDC activities, we synthesized a mouse IgG_2a subclass mAb, and further defucosylated it to augment those activities. In this study, we examined whether the developed 5-mG_2a-f induced ADCC and CDC in CD44-expressing oral cancer cell lines, such as SAS and HSC-2 cells. As shown in Fig. 5A, 5-mG_2a-f exhibited higher ADCC (16% cytotoxicity) in SAS cells compared with that of control PBS (3.4% cytotoxicity; P<0.01) and control mouse IgG_2a treatment (4.2% cytotoxicity; P<0.01). Similarly, 5-mG_2a-f exhibited higher ADCC (18% cytotoxicity) against HSC-2 cells compared with that of control PBS (3.1% cytotoxicity; P<0.01) and control mouse IgG_2a treatment (5.2% cytotoxicity; P<0.01), indicating that ADCC in SAS cells is similar with that in HSC-2 cells, despite the binding affinity of 5-mG_2a-f in SAS cells being 9.3-fold higher than in HSC-2 cells (Fig. 4). As shown in Fig. 5B, 5-mG_2a-f exhibited slightly higher CDC (33% cytotoxicity) in SAS cells compared with control PBS (21% cytotoxicity; P<0.01) and control mouse IgG_2a treatment (22% cytotoxicity; P<0.01). Similarly, 5-mG_2a-f exhibited slightly higher CDC (30% cytotoxicity) in HSC-2 cells compared with control PBS (18% cytotoxicity; P<0.01) and control mouse IgG_2a treatment (19% cytotoxicity; not significant). Although ADCC/CDC activities of 5-mG_2a-f in oral cancer cells are not outstanding, 5-mG_2a-f may exert antitumor activity against oral cancer cells in vivo.

Next, we investigated whether 5-mG_2a-f inhibits cell growth of SAS and HSC-2 cells in anchorage-independent condition. As shown in Fig. S2A, both SAS and HSC-2 cells grew in anchorage-independent condition for 48 h. In contrast, an anti-CD44 mAb (5-mG_2a-f) did not inhibit the growth of SAS or HSC-2 compared to control mouse IgG_2a (Fig. S2B), indicating that 5-mG_2a-f did not affect the cell growth of oral cancer cell lines in anchorage-independent condition.

SAS cells were subcutaneously implanted into the flanks of nude mice. In protocol-1, 5-mG_2a-f (100 µg) and control mouse IgG (100 µg) were injected i.p. three times into the mice, on days 1, 7, and 14 after SAS cell injections. Tumor volume was measured on days 6, 12, 15, and 19. Tumor development was significantly reduced in the 5-mG_2a-f-treated mice on days 12, 15, and 19 in comparison to the IgG-treated control mice (Fig. 6A). Tumor volume on day 19 was reduced by 27% after 5-mG_2a-f treatment. Tumors from 5-mG_2a-f-treated mice weighed significantly less than tumors from IgG-treated control mice (16.9% reduction, Fig. 6B). Resected tumors on day 19 are depicted in Fig. 6C. Control and 5-mG_2a-f-treated SAS xenograft mice are shown on day 19 in Fig. S3A and B, respectively. Total body weights did not significantly differ between the two groups (Fig. S3C).

In protocol-2 of the SAS xenograft models, tumor formation of 16 SAS-bearing mice was observed on day 7. Then, these 16 SAS-bearing mice were divided into a 5-mG_2a-f-treated group and a control group. On days 7, 14, and 21 after SAS cell injections into the mice, 5-mG_2a-f (100 µg) and control mouse IgG (100 µg) were injected i.p. into the mice. Tumor formation was observed in mice in both treated and control groups. Tumor volume was measured on days 7, 12, 15, 19, 22, and 27. 5-mG_2a-f-treated mice displayed significantly reduced tumor development on days 22 and 27 in comparison to IgG-treated control mice (Fig. 7A). Tumor volume reduction by 5-mG_2a-f was 43% on day 27. Tumors from the 5-mG_2a-f-treated mice weighed significantly less than tumors from the IgG-treated control mice (27.1% reduction, Fig. 7B). Resected tumors on day 27 are depicted in Fig. 7C. Control and 5-mG_2a-f-treated SAS xenograft mice are shown on day 27 in Fig. S4A and B, respectively. Total body weights did not significantly differ between the two groups (Fig. S4C). These results indicate that 5-mG_2a-f reduced the growth of SAS xenografts effectively, even when 5-mG_2a-f was injected 7 days post-SAS cell injections in mice.

In a second xenograft model of oral cancers, HSC-2 cells were subcutaneously implanted into the flanks of nude mice. In protocol-1 of HSC-2 ×enograft models, 5-mG_2a-f (100 µg) and control mouse IgG (100 µg) were injected i.p. three times into the mice, on days 1, 7, and 14 after HSC-2 cell injections into the mice. Tumor volume was measured on days 6, 12, 15, and 19. 5-mG_2a-f-treated mice displayed significantly reduced tumor development on days 12, 15, and 19 in comparison to IgG-treated control mice (Fig. 8A). Tumor volume reduction by 5-mG_2a-f was 53% on day 19. Tumors from 5-mG_2a-f-treated mice weighed significantly less than HSC-2 tumors from IgG-treated control mice (44.1% reduction, Fig. 8B). Resected tumors on day 19 are depicted in Fig. 8C. Control and 5-mG_2a-f-treated HSC-2 ×enograft mice are shown on day 19 in Fig. S5A and B, respectively. Total body weights did not significantly differ between the two groups (Fig. S5C).

In protocol-2 of the HSC-2 ×enograft models, tumor formation of 16 HSC-2-bearing mice was observed on day 7. Then, these 16 HSC-2-bearing mice were divided into a 5-mG_2a-f-treated group and a control group. On days 7, 14, and 21 after cell injections into the mice, 5-mG_2a-f (100 µg) and control mouse IgG (100 µg) were injected i.p. into the mice. Tumor volume was measured on days 7, 12, 15, 19, 22, and 27. 5-mG_2a-f-treated mice displayed significantly reduced tumor development on days 22 and 27 in comparison to IgG-treated control mice (Fig. 9A). Tumor volume reduction in 5-mG_2a-treated mice was 32% on day 27. Tumors from 5-mG_2a-f-treated mice weighed significantly less than tumors from IgG-treated control mice (27.1% reduction, Fig. 9B). Resected tumors on day 27 are depicted in Fig. 9C. Control and 5-mG_2a-f-treated HSC-2 ×enograft mice are shown on day 27 in Fig. S6A and B, respectively. Total body weights did not significantly differ between the two groups (Fig. S6C). These results indicate that 5-mG_2a-f reduced the growth of HSC-2 ×enografts effectively, even when 5-mG_2a-f was injected 7 days post-HSC-2 cell injections in mice.