Establishment of a Novel Anti-CD44 Variant 10 Monoclonal Antibody C44Mab-18 for Immunohistochemical Analysis against Oral Squamous Cell Carcinomas

Head and neck squamous cell carcinoma (HNSCC) is the most common type of head and neck cancer, and has been revealed as the second-highest expression of CD44 in cancers. CD44 has been investigated as a cancer stem cell marker of HNSCC and plays a critical role in tumor malignant progression. Especially, splicing variant isoforms of CD44 (CD44v) are overexpressed in cancers and considered a promising target for cancer diagnosis and therapy. We developed monoclonal antibodies (mAbs) against CD44 by immunizing mice with CD44v3–10-overexpressed PANC-1 cells. Among the established clones, C44Mab-18 (IgM, kappa) reacted with CHO/CD44v3–10, but not with CHO/CD44s and parental CHO-K1 using flow cytometry. The epitope mapping using peptides that cover variant exon-encoded regions revealed that C44Mab-18 recognized the border sequence between variant 10 and the constant exon 16-encoded sequence. These results suggest that C44Mab-18 recognizes variant 10-containing CD44v, but not CD44s. Furthermore, C44Mab-18 could recognize the human oral squamous cell carcinoma (OSCC) cell line, HSC-3, in flow cytometry. The apparent dissociation constant (KD) of C44Mab-18 for CHO/CD44v3–10 and HSC-3 was 1.6 × 10−7 M and 1.7 × 10−7 M, respectively. Furthermore, C44Mab-18 detected CD44v3–10 but not CHO/CD44s in Western blotting, and endogenous CD44v10 in immunohistochemistry using OSCC tissues. These results indicate that C44Mab-18 is useful for detecting CD44v10 in flow cytometry and immunohistochemistry.


Introduction
Head and neck cancer is the seventh most common cancer type globally, and exhibits a profound impact on patients and their quality of life after surgical ablation and therapies [1]. Head and neck squamous cell carcinoma (HNSCC) is the most common type of head and neck cancer. The treatment of HNSCC includes surgery, chemotherapy, radiation therapy, immunotherapy, molecular targeted therapy, or a combination of those modalities [2]. Although survival can be improved through the development of treatments, cancer metastasis and resistance to drugs remain the main causes of death [3]. The rate of 5-year survival remains stagnant at approximately 50% [4].
CD44 is a multifunctional type I transmembrane glycoprotein that mediates metastasis and drug resistance in tumor cells. HNSCC is the second-highest CD44-expressing tumor in the Pan-Cancer Atlas [5]. The alternative splicing of CD44 mRNA produces the various isoforms [6]. The constant exons including the first five (1 to 5) and the last five (16 to 20) are present in all CD44 variants and make up the standard isoform (CD44s). The CD44 (Nacalai Tesque, Inc., Kyoto, Japan) containing 10% heat-inactivated fetal bovine serum (FBS; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and antibiotics (100 U/mL penicillin, 100 µg/mL streptomycin, and 0.25 µg/mL amphotericin B). A human OSCC cell line (HSC-3, the Japanese Collection of Research Bioresources, Osaka, Japan) was cultured in Dulbecco's Modified Eagle Medium (DMEM) (Nacalai Tesque, Inc., Kyoto, Japan) supplemented as indicated above. All cell lines were grown in a humidified incubator at 37 • C with 5% CO 2 .

Determination of Apparent Dissociation Constant (K D ) via Flow Cytometry
The serially diluted C 44 Mab-18 at the indicated concentrations was suspended with 2 × 10 5 of HSC-3 and CHO/CD44v3-10 cells. The cells were further treated with antimouse IgG conjugated with Alexa Fluor 488 (1:200). Fluorescence data were analyzed, and the apparent dissociation constant (K D ) was determined by fitting binding isotherms to built-in one-site binding models of GraphPad Prism 8 (GraphPad Software, Inc., La Jolla, CA, USA).

Immunohistochemical Analysis of Formalin-Fixed Paraffin-Embedded (FFPE) Tissues
Antigen retrieval of an OSCC tissue array (OR601c; US Biomax Inc., Rockville, MD, USA) was performed using EnVision FLEX Target Retrieval Solution High pH (Agilent Technologies, Inc.). SuperBlock T20 (Thermo Fisher Scientific, Inc.) was used for blocking. The sections were incubated with 1 µg/mL of C 44 Mab-18 and 1 µg/mL of C 44 Mab-46 at room temperature for 1 h. The sections were further treated with the EnVision+ Kit for a mouse (Agilent Technologies Inc.) at room temperature for 30 min. The chromogenic reaction and counterstaining were performed using 3,3 -diaminobenzidine tetrahydrochloride (DAB; Agilent Technologies Inc.) and hematoxylin (FUJIFILM Wako Pure Chemical Corporation), respectively.

Western Blot Analysis
To assess the sensitivity of C44Mab-18 in Western blot analysis, we analyzed the cell lysates from CHO-K1, CHO/CD44s, and CHO/CD44v3-10. C44Mab-18 mainly detected

Immunohistochemical Analysis Using C 44 Mab-18 against Tumor Tissues
Since HNSCC is revealed as the second highest CD44-expressing tumor in the Pan-Cancer Atlas [5], we examined the reactivity of C 44 Mab-18 and C 44 Mab-46 in immunohistochemical analyses using FFPE sections of OSCC tissue array. As shown in Figure 6, C 44 Mab-18 was able to distinguish tumor cells from stromal tissues. In contrast, C 44 Mab-46 stained both. We summarized the data of immunohistochemical analyses in Table 1 Since HNSCC is revealed as the second highest CD44-expressing tumor in the Pan-Cancer Atlas [5], we examined the reactivity of C44Mab-18 and C44Mab-46 in immunohistochemical analyses using FFPE sections of OSCC tissue array. As shown in Figure 6, C44Mab-18 was able to distinguish tumor cells from stromal tissues. In contrast, C44Mab-46 stained both. We summarized the data of immunohistochemical analyses in Table 1; C44Mab-18 stained 41 out of 50 cases (82%) in OSCC. These results indicate that C44Mab-18 applies to the immunohistochemical analysis of FFPE tumor sections.
Since the CD44 protein is modified by a variety of N-glycans and O-glycans, the molecular weight of CD44v isoforms surpasses 200-kDa [37]. C 44 Mab-18 recognized both more than 180-kDa and~70-kDa bands ( Figure 5A) in the lysate from CHO/CD44v3-10. The 70 kDa is approximately identical to the predicted molecular weight of CD44v3-10 from the amino acid sequence. Therefore, C 44 Mab-18 could recognize CD44v3-10 regardless of the glycosylation. The detailed epitope mapping and the influence of glycosylation on C 44 Mab-18 recognition should be investigated in future studies.
CD44v8-10 was shown to interact with xCT, a glutamate-cystine transporter, and regulate the level of reduced glutathione in tumor cells. The interaction is important for the stabilization of xCT on the cell surface, which promotes the defense against reactive oxygen species [17]. Furthermore, the interaction failed in CD44v8-10 (S301A), an Nlinked glycosylation consensus motif (Asn-X-Ser/Thr) mutant in the variant 10-encoded region [17]. Therefore, it is worthwhile to investigate whether C 44 Mab-18 interferes with the interaction between CD44v8-10 and xCT in future studies. Furthermore, several studies have revealed that CD44v9 is used as a predictive marker for recurrence [38] and a biomarker for patient selection and efficacy of xCT inhibitors, sulfasalazine in gastric cancer [39]. Further investigations are also required to clarify the clinical significance of CD44v10 expression using C 44 Mab-18.
The mAbs against CD44 have been considered a therapeutic option for solid tumors and leukemia [12]. However, anti-pan-CD44 mAbs can affect normal tissues such as the epithelium and hematopoiesis. In a preclinical study using a murine thymoma model, a comparative study between an anti-pan-CD44 mAb (IM-7) and an anti-murine CD44v10 mAb (K926) was conducted in CD44v10-transfected EL4 thymoma (EL4-v10) [40]. The results showed that a blockade of CD44v10 by K926 was superior to that of IM-7 in intra-marrow EL4-v10 growth retardation. Furthermore, K926 hardly disturbed the hematopoietic stem cell (HSC) interaction with the bone marrow stroma. In contrast, IM-7 strongly affected the embedding of HSC in the bone marrow stroma [40]. These results indicated that the therapeutic use of anti-pan-CD44 mAbs should be avoided in favor of CD44v-specific mAbs as far as leukemic cells express CD44v isoforms.
In a humanized mouse model, CD44v8-10 was elevated during chronic myeloid leukemia progression from chronic phase to blast crisis [41]. Furthermore, increased transcription of CD44 mRNA was observed in human acute myeloid leukemia (AML) patients with FLT3 or DNMT3A mutations through the suppression of CpG islands methylation in the promoter [42]. An anti-CD44v6 mAb (BIWA-8) derived from VFF-18 [43] was engineered to develop chimeric antigen receptors (CARs) for AML with FLT3 or DNMT3A mutations. The CD44v6 CAR-T cells exhibited potent anti-leukemic effects [42]. We have established class-switched and defucosylated IgG 2a recombinant mAbs and evaluated the antitumor activity in xenograft models [44]. Therefore, the production of class-switched and defucosylated C 44 Mab-18 is an important strategy to evaluate the antitumor effect in preclinical models.
Since anti-pan-CD44 and anti-CD44v mAbs still have the possibility of causing side effects by affecting normal tissues, the clinical applications are limited. This study used tumor cell-expressed CD44v3-10 as an immunogen. This strategy is critical for the development of cancer-specific mAbs (CasMabs). We developed podocalyxin-targeting CasMabs [45] and PDPN-targeting CasMabs [46], which react with the aberrantly glycosylated targets selectively expressed in cancer [47]. Anti-PDPN-CasMabs have been applied to CAR-T therapy in preclinical studies [48][49][50]. For CasMab development, we should perform a further selection of our established anti-CD44 mAbs by comparing the reactivity against normal cells and tissues. Anti-CD44 CasMabs could be applicable for designing the modalities, including antibody-drug conjugates and CAR-T.

Conclusions
In this study, we established an anti-CD44v10 mAb (C 44 Mab-18). We also established an anti-CD44v8 mAb (C 44 Mab-94) (manuscript submitted, see Supplementary Materials). Therefore, we have established an anti-CD44 mAb library that covers almost all CD44 variants. This library could contribute to the diagnosis of not only carcinoma, but also hematopoietic malignancies. Since we have already cloned the V H and V L cDNA of anti-CD44 mAbs, the production of recombinant mAbs or CARs could contribute to the development of novel tumor therapies.