Development of a Novel Anti-CD44 Variant 4 Monoclonal Antibody C44Mab-108 for Immunohistochemistry

CD44 has been known as a marker of tumor-initiating cells, and plays pro-tumorigenic functions in many cancers. The splicing variants play critical roles in the malignant progression of cancers by promoting stemness, cancer cell invasion or metastasis, and resistance to chemo- and radiotherapy. To understand each CD44 variant (CD44v) function is essential to know the property of cancers and the establishment of the therapy. However, the function of the variant 4-encoded region has not been elucidated. Therefore, specific monoclonal antibodies (mAbs) against variant 4 are indispensable for basic research, tumor diagnosis, and therapy. In this study, we established anti-CD44 variant 4 (CD44v4) mAbs by immunizing mice with a peptide containing the variant 4-encoded region. We next performed flow cytometry, western blotting, and immunohistochemistry to characterize them. One of the established clones (C44Mab-108; IgG1, kappa) reacted with CD44v3-10-overexpressed Chinese hamster ovary-K1 cells (CHO/CD44v3-10). The KD of C44Mab-108 for CHO/CD44 v3-10 was 3.4 × 10−7 M. In western blot analysis, C44Mab-108 detected CD44v3-10 in the lysate of CHO/CD44v3-10 cells. Furthermore, C44Mab-108 stained formalin-fixed paraffin-embedded (FFPE) oral squamous carcinoma tissues in immunohistochemistry. These results indicated that C44Mab-108 is useful to detect CD44v4 in immunohistochemistry using FFPE tissues.


Introduction
CD44 is a type I transmembrane glycoprotein, which is widely distributed in normal tissues [1]. CD44 consists of several exons. CD44 standard isoform (CD44s) is the shortest isoform of CD44 (85-95 kDa) which is produced by assembling the first five and the last five constant region exons [2]. The middle variant exons (v1-v10) can be alternatively spliced and assembled with the first five and the last five exons of CD44s. They are defined as CD44 variant isoforms (CD44v) [3]. It attracted considerable interest when it was discovered that CD44v induced metastatic properties in tumor cells [4]. A growing body of evidence suggests the importance of CD44v in the malignant progression of tumors [5], as well as cancer-initiating properties [6].
Both CD44s and CD44v can bind to hyaluronic acid (HA) and are involved in cell adhesion, proliferation, and migration [7,8]. CD44s is widely expressed in normal tissues and plays important roles in hematopoiesis, the immune system, and organogenesis [9]. CD44v plays critical roles in the promotion of tumor invasion, metastasis, stemness, and resistance to chemo-and radiotherapy [10,11]. CD44v3-10 can bind to heparin-binding epidermal growth factor-like growth factor (HB-EGF) and fibroblast growth factors (FGFs) via v3-encoded region, and functions as a co-receptor of receptor tyrosine kinases [12]. Moreover, the v6-encoded region is essential for the recruitment of hepatocyte growth factor (HGF) to its receptor, c-MET [13]. In addition, the v8-10-encoded region confers oxidative stress resistance [14]. Therefore, understanding the function of each variant is essential to identify the properties of carcinomas. However, the function of the variant 4-encoded region has not been elucidated. Therefore, specific antibodies against CD44v4 are indispensable for basic research, tumor diagnosis, and therapy.
In this study, we developed a novel anti-CD44v4 mAb, C 44 Mab-108 (IgG 1 , kappa) by peptide immunization of the v4-encoded region, and evaluated its applications, including flow cytometry, western blotting, and immunohistochemical analyses.

Hybridoma Production
The female BALB/c mice were purchased from CLEA Japan (Tokyo, Japan). All animal experiments were conducted according to the guidelines and regulations to minimize animal suffering and distress in the laboratory. The Animal Care and Use Committee of Tohoku University (Permit number: 2019NiA-001) approved animal experiments. The mice were housed under specific pathogen-free conditions and monitored daily for health during the duration of the experiment. To develop mAbs against CD44v4, we intraperitoneally immunized mice with the KLH-conjugated CD44v4 peptide (100 µg) plus an adjuvant (Imject Alum; Thermo Fisher Scientific Inc.). The procedure included three additional weekly immunization (100 µg/mouse). A final booster injection (100 µg/mouse) was performed two days before harvesting spleen cells. The harvested spleen cells were subsequently fused with P3U1 cells, using PEG1500 (Roche Diagnostics, Indianapolis, IN, USA). The hybridomas were selected in the presence of hypoxanthine, aminopterin, and thymidine (HAT; Thermo Fisher Scientific Inc.). The supernatants were screened using enzyme-linked immunosorbent assay (ELISA) with the CD44v4 peptide [1 µg/mL in phosphate-buffered saline (PBS)], followed by flow cytometry using CHO/CD44v3-10 and CHO-K1 cells. CHO/CD44v3-10-reactive and parental CHO-K1-non-reactive supernatants were determined to be positive for CD44v3-10. To establish single clones, limiting dilution was performed.

ELISA
The synthesized peptide (DHTKQNQDWTQWNPSHSNP), which is included in CD44v4 plus C-terminal cysteine, was immobilized at a concentration of 1 µg/mL in PBS on Nunc Maxisorp 96 well immunoplates (Thermo Fisher Scientific Inc.) for 30 min at 37 • C. This peptide might be dimerized using the disulfide bond. The plates were washed with 0.05% Tween 20 in PBS (PBST; Nacalai Tesque, Inc.), and the wells were blocked with 1% bovine serum albumin (BSA) in PBST for 30 min at 37 • C. Then, the plates were incubated with the supernatants of hybridomas, followed by peroxidase-conjugated antimouse immunoglobulins (1:2000 diluted; Agilent Technologies Inc., Santa Clara, CA, USA). ELISA POD Substrate TMB Kit (Nacalai Tesque, Inc.) was used for enzymatic reactions. The optical density at 655 nm was measured using an iMark microplate reader (Bio-Rad Laboratories, Inc., Berkeley, CA, USA).
In Supplementary Table S1, CD44v3−10 ectodomain [16] was immobilized at a concentration of 1 µg/mL for 30 min at 37 • C. After blocking with 1% BSA in PBST for 30 min at 37 • C, the plates were incubated with the supernatant of C 44 Mab-108, followed by incubation with peroxidase-conjugated anti-mouse immunoglobulins, anti-mouse heavy chains (IgG 1 , IgG 2a , IgG 2b , IgG 3 , and IgM; SouthernBiotech, Birmingham, AL, USA), or antimouse light chains (kappa and lambda; SouthernBiotech). The enzymatic reactions were conducted and the optical density was measured. C 44 Mab-108 isotype was determined by the reactivity of secondary antibodies.
In Supplementary Figure S2, fifty-eight synthesized peptides [17] were immobilized at a concentration of 10 µg/mL for 30 min at 37 • C. After blocking with 1% BSA in PBST for 30 min at 37 • C, the plates were incubated with C 44 Mab-108 (1 µg/mL), followed by incubation with peroxidase-conjugated anti-mouse immunoglobulins. The enzymatic reactions were conducted and the optical density was measured.

Flow Cytometry
The cells were isolated using 0.25% trypsin and 1 mM ethylenediaminetetraacetic acid (Nacalai Tesque, Inc.). The cells were washed with blocking buffer (0.1% BSA in PBS) and treated with C 44 Mab-108 or C 44 Mab-46 for 30 min at 4 • C. Subsequently, the cells were treated with Alexa Fluor 488-conjugated anti-mouse IgG (1:2000; Cell Signaling Technology, Inc., Danvers, MA, USA) for 30 min at 4 • C. Fluorescence data were collected using the SA3800 Cell Analyzer and analyzed using SA3800 software ver. 2.05 (Sony Corporation, Tokyo, Japan) and FlowJo (BD Biosciences, Franklin Lakes, NJ, USA).

Determination of Dissociation Constant (K D ) via Flow Cytometry
Serially diluted C 44 Mab-108 was suspended with CHO/CD44v3-10 cells. The cells were further treated with Alexa Fluor 488-conjugated anti-mouse IgG (1:200; Cell Signaling Technology, Inc.). Fluorescence data were collected using BD FACSLyric and analyzed using BD FACSuite software version 1.3 (BD Biosciences). To determine the dissociation constant (K D ), GraphPad Prism 8 (the fitting binding isotherms to built-in one-site binding models; GraphPad Software, Inc., San Diego, CA, USA) was used.

Immunohistochemical Analysis
The formalin-fixed paraffin-embedded (FFPE) OSCC tissue microarray (Product Code: OR601c, US Biomax Inc., Rockville, MD, USA) and the esophageal tissue microarray (Product Code: BC02011, US Biomax Inc.) were deparaffinized in xylene (Sigma-Aldrich Corp.) and rehydrated. The FFPE OSCC tissue for peptide blocking assay was obtained from Tokyo Medical and Dental University [36]. The tissues were autoclaved in citrate buffer (pH 6.0; Nichirei Biosciences, Inc., Tokyo, Japan) for 20 min for antigen retrieval. After blocking with SuperBlock T20 (Thermo Fisher Scientific, Inc.), the sections were incubated with C 44 Mab-108 (10 µg/mL) and C 44 Mab-46 (1 µg/mL), or without the primary antibody (control) for 1 h at room temperature and then treated with the EnVision+ Kit for mouse (Agilent Technologies Inc.) for 30 min. The color was developed using 3,3 -diaminobenzidine tetrahydrochloride (DAB; Agilent Technologies Inc.) for 2 min. Counterstaining was performed with hematoxylin (FUJIFILM Wako Pure Chemical Corporation). Hematoxylin and eosin (HE) staining (FUJIFILM Wako Pure Chemical Corporation) was performed using consecutive tissue sections. Leica DMD108 (Leica Microsystems GmbH, Wetzlar, Germany) was used to examine the sections and obtain images.

Development of Anti-CD44v4 mAbs by Peptide Immunization
To develop anti-CD44v4 mAbs, mice were immunized with the KLH-conjugated CD44v4 peptide ( Figure 1A). The splenocytes were fused with myeloma P3U1 cells ( Figure 1B). The developed hybridomas were subsequently seeded into 96-well plates and cultured for six days. The positive wells for the naked CD44v4 peptide were selected using ELISA, followed by the selection of CHO/CD44v3-10-reactive and parental CHO-K1non-reactive supernatants using flow cytometry ( Figure 1C). After the limiting dilution, 12 clones were established. After several additional screenings, including flow cytometry and immunohistochemistry, an anti-CD44v4 mAb (clone C 44 Mab-108; mouse IgG 1 , kappa) was finally selected and investigated in this study ( Figure 1D and Supplementary Table S1). We confirmed that C 44 Mab-108 recognized only a CD44v4-containing peptide (aa 271-290) among peptides, which cover the extracellular domain of CD44v3-10 [17] (Supplementary Figure S1).

Immunohistochemical Analysis Using C 44 Mab-108 against OSCC Tissues
To investigate whether C 44 Mab-108 can be used for immunohistochemical analyses using paraffin-embedded tumor sections, we used sequential sections of an OSCC tissue microarray. In a well-differentiated OSCC section ( Figure 4A-H), a clear membrane-staining in OSCC was observed by C 44 Mab-46 ( Figure 4C,D), but hardly detected by C 44 Mab-108 ( Figure 4A,B). In an OSCC section with the stromal invaded phenotype ( Figure 4I-P), C 44 Mab-108 ( Figure 4I,J) strongly stained stromal invaded OSCC and could clearly distinguish tumor cells from stromal tissues. In contrast, C 44 Mab-46 ( Figure 4K,L) stained both. The reactivity of C 44 Mab-108 was eliminated completely by a CD44 peptide (aa 271-290), which contains the C 44 Mab-108 epitope ( Figure 4Q,R). We summarized the data of immunohistochemical analysis in Table 1; C 44 Mab-108 stained 20 out of 50 (40%) cases of OSCC. Similar staining patterns were also observed in esophageal SCC tissues. C 44 Mab-108 also stained tumors selectively, and could clearly distinguish tumor cells from stroma (Supplementary Figure S2). These results indicated that C 44 Mab-108 is useful for immunohistochemical analysis of paraffin-embedded tumor sections.  44 Mab-108 reactivity to OSCC tissue by the CD44 peptide (aa 271-290) containing the C 44 Mab-108 epitope. After antigen retrieval, sections were incubated with C 44 Mab-108 (10 µg/mL) or C 44 Mab-108 (10 µg/mL) plus human CD44 peptide (aa 271-290, 10 µg/mL) followed by treatment with the Envision+ kit. The color was developed using DAB, and sections were counterstained with hematoxylin. Scale bar = 100 µm.

Discussion
CD44v have been implicated as a marker of cancer-initiating cells and plays protumorigenic functions in many carcinomas [37]. The v3-encoded region possesses heparan sulfate moieties that recruit to HB-EGF and FGFs [12]. The v6-encoded region forms a ternary complex with MET and HGF, which is essential for the c-MET activation [13]. Furthermore, the CD44 intracellular domain is required for the recruitment of the ezrin, radixin, and moesin complex, which potentiates the c-MET signal transduction [38]. These functions of CD44v are important for the malignant progression of tumors to potentiate proliferation, invasiveness, and metastatic spread. Furthermore, CD44v interacts with a cystine-glutamate transporter (xCT) subunit through a v8-10-encoded region [14], which mediates oxidative stress and antitumor drug resistance in several carcinomas [39]. Therefore, the establishment and characterization of mAbs, which recognize CD44v are thought to be essential for the development of CD44-targeting tumor diagnosis and therapy.
In contrast, the function of the v4-encoded region has not been elucidated. Therefore, specific antibodies against CD44v4 have been desired. In this study, we developed C 44 Mab-108, which can recognize the v4 region. We showed the usefulness of flow cytometry ( Figure 2), western blotting (Figure 3), and immunohistochemistry ( Figure 4 and Supplementary Figure S2). In Supplementary Figure S3, we showed the homology of the v4 region between human, mouse, rat, and Chinese hamster sequences. In our preliminary epitope mapping analysis, the alanine-substitution of each amino acid in the human v4 region, which is identical to the mouse sequence, reduced the reactivity of C 44 Mab-108 (manuscript preparation). The result suggests that C 44 Mab-108 recognizes both human and mouse CD44v4.
As shown in Figure 2C, the affinity of C 44 Mab-108 for CHO/CD44v3-10 was not high. Although high affinity is the general goal for antibody generation, recent reports showed that low rather than high affinity exhibits elevated activity through inducing clustering of receptors. For instance, low-affinity variants of an anti-PD-1 mAb nivolumab mediated more potent signaling and affected T-cell activation. These findings reveal a new paradigm for antibody-mediated receptor signaling [40]. Since CD44 is involved in intracellular signaling, the relationship between the antibody affinity and the effect of cellular signaling should be investigated in future studies.
Most CD44v mAbs have been developed by the immunization of recombinant or cell surface expressed CD44v, which received the glycosylation. CD44 is known to be heavily glycosylated [41], and the glycosylation pattern is thought to depend on the host cells. Moreover, it is not clear whether the available anti-CD44 mAbs recognize the peptide or glycopeptide structures of CD44v. Since C 44 Mab-108 was established by the peptide immunization, C 44 Mab-108 can recognize the definite peptide structure of the variant 4-encoded region. Since the region is also expected to receive the glycosylation [41], further studies are required to reveal whether the glycosylation affects the recognition by C 44 Mab-108. In the immunohistochemical analysis ( Figure 4, Table 1), C 44 Mab-108 could detect endogenous CD44v4 in some OSCC tissues. Although an anti-pan-CD44 mAb, C 44 Mab-46 recognized not only OSCC tissues, but also stromal tissues, C 44 Mab-108 stained the tumor tissues selectively. Furthermore, C 44 Mab-108 recognized a limited population of OSCC tissues compared to C 44 Mab-46 ( Figure 4, Table 1). In the future, we should evaluate the clinical significance of the expression pattern and histology of C 44 Mab-108-positive OSCC. The information could help the future diagnostic and therapeutic applications of C 44 Mab-108 in OSCC.
Clinical trials of anti-CD44 mAbs have been conducted [42]. A humanized anti-CD44v6 mAb BIWA4 (bivatuzumab)−mertansine drug conjugate was evaluated but discontinued due to severe skin toxicities [43,44]. An anti-pan CD44 mAb, RG7356, exhibited an acceptable safety profile in patients with advanced CD44-expressing solid tumors. However, the study was terminated due to no evidence of a clinical and pharmacodynamic dose-response relationship with RG7356 [45]. Therefore, the development of anti-CD44 mAbs with more potent and fewer side effects is required.