LpMab-12 Established by CasMab Technology Specifically Detects Sialylated O-Glycan on Thr52 of Platelet Aggregation-Stimulating Domain of Human Podoplanin

Podoplanin (PDPN), also known as Aggrus, possesses three tandem repeat of platelet aggregation-stimulating (PLAG) domains in its N-terminus. Among the PLAG domains, sialylated O-glycan on Thr52 of PLAG3 is essential for the binding to C-type lectin-like receptor-2 (CLEC-2) and the platelet-aggregating activity of human PDPN (hPDPN). Although various anti-hPDPN monoclonal antibodies (mAbs) have been generated, no specific mAb has been reported to target the epitope containing glycosylated Thr52. We recently established CasMab technology to develop mAbs against glycosylated membrane proteins. Herein, we report the development of a novel anti-glycopeptide mAb (GpMab), LpMab-12. LpMab-12 detected endogenous hPDPN by flow cytometry. Immunohistochemical analyses also showed that hPDPN-expressing lymphatic endothelial and cancer cells were clearly labeled by LpMab-12. The minimal epitope of LpMab-12 was identified as Asp49–Pro53 of hPDPN. Furthermore, LpMab-12 reacted with the synthetic glycopeptide of hPDPN, corresponding to 38–54 amino acids (hpp3854: 38-EGGVAMPGAEDDVVTPG-54), which carries α2–6 sialylated N-acetyl-D-galactosamine (GalNAc) on Thr52. LpMab-12 did not recognize non-sialylated GalNAc-attached glycopeptide, indicating that sialylated GalNAc on Thr52 is necessary for the binding of LpMab-12 to hPDPN. Thus, LpMab-12 could serve as a new diagnostic tool for determining whether hPDPN possesses the sialylation on Thr52, a site-specific post-translational modification critical for the hPDPN association with CLEC-2.


Introduction
Podoplanin (PDPN), the endogenous ligand of C-type lectin-like receptor-2 (CLEC-2) [1,2], is highly expressed not only in various tumors including oral cancer, lung cancer, esophageal Three female BALB/c mice (four-weeks old) were purchased from CLEA Japan (Tokyo, Japan). Animals were housed under pathogen-free conditions. "The Animal Care and Use Committee of Tohoku University" approved the animal experiments described herein. The use of one oral cancer tissue was reviewed and approved by "Tokyo Medical and Dental University Institutional Review Board" [23]. Written informed consent was obtained for the human cancer tissue samples used in this study. The use of human heart tissue sections for immunohistochemical analysis was reviewed and approved by the "Partners Institutional Review Board".

Hybridoma production
Three BALB/c mice were immunized by intraperitoneal (i.p.) injection of 1 × 10 8 LN229/ hPDPN cells together with Imject Alum (Thermo Fisher Scientific Inc.), as previously described [22]. After several additional immunizations, a booster injection was given i.p. two days before mice were euthanized by cervical dislocation and spleen cells were harvested. The spleen cells were fused with P3U1 cells using PEG1500 (Roche Diagnostics, Indianapolis, IN). The fused cells were grown in RPMI medium with hypoxanthine, aminopterin, and thymidine selection medium supplement (Thermo Fisher Scientific Inc.). The culture supernatants were screened using enzyme-linked immunosorbent assay (ELISA) for binding to recombinant hPDPN purified from LN229/hPDPN cells.

Flow cytometry
Cell lines were harvested by brief exposure to 0.25% Trypsin/1 mM EDTA (Nacalai Tesque, Inc.). After washing with PBS, the cells were incubated with LpMab-12 (1 μg/ml) for 30 min at 4°C, followed by the incubation with Oregon Green 488 goat anti-mouse IgG (Thermo Fisher Scientific Inc.). Fluorescence data were collected using a Cell Analyzer EC800 (Sony Corp., Tokyo, Japan).

Immunohistochemical analyses of oral cancer
Four-μm-thick histologic sections were deparaffinized in xylene and rehydrated. Without antigen retrieval procedure, sections were incubated with 1 μg/ml of LpMab-12 or LpMab-7 for 1 h at room temperature followed by treatment with Envision+ kit (Dako) for 30 min. Color was developed using 3, 3-diaminobenzidine tetrahydrochloride (DAB; Dako), and then the sections were counterstained with hematoxylin (Wako Pure Chemical Industries Ltd.).

Production of hPDPN glycopeptide
The hPDPN glycopeptide (hpp3854) with a GalNAc residue was purchased from Peptide Institute (Osaka, Japan), and used as an acceptor substrate. For synthesis of the sialylated GalNAc on hpp3854, 25 mM HEPES (pH 7.0) containing 30 μM of acceptor substrate, 10 mM MnCl 2 (Nacalai Tesque, Inc.), and 250 μM CMP-Neu5Ac (Sigma-Aldrich Corp.) was used. A half volume of purified ST6GalNAcT-I enzyme was added to the reaction mixture and incubated at 37°C for 24 h. The recombinant ST6GalNAcT-I enzyme was bound to anti-FLAG M2 affinity gel (Sigma-Aldrich Corp.). After enzymatic reaction, the resin was removed by filtration using an Ultrafree-MC column (EMD-Millipore). Then, the glycopeptides were purified using a reversed-phase SPE cartridge (ZipTip C18; EMD-Millipore). The other glycopeptides were produced sequentially as previously described [2].
As shown in Fig 1B and 1C, lymphatic endothelial cells, identified by a lymphatic marker lymphatic vessel endothelial hyaluronan receptor-1 (LYVE-1), were clearly stained by LpMab-12 in myocardial samples from human hearts, indicating that LpMab-12 is applicable for reliably detecting hPDPN by immunohistochemistry.
Using flow cytometry analysis, apparent dissociation constant of LpMab-12 was determined to be 1.2 × 10 −8 M for LN319 and 1.8 × 10 −8 M for LEC, suggesting that the binding affinity of LpMab-12 is comparable with previously established anti-hPDPN mAbs [22] for hPDPNexpressing cancer cells and normal cells (Fig 1D).

Epitope mapping
To determine the critical epitope for the LpMab-12 interaction with hPDPN, we compared the mAb binding to the hPDPN carrying different point mutations. Using Western blot, we found that LpMab-12 did not detect protein sequences with the following amino acid substitutions: D49A, V51A, T52A, and P53A (Fig 3A).
In agreement, flow cytometry analysis demonstrated that LpMab-12 did not react with D49A, T52A, and P53A mutant proteins (Fig 3B) Thus, our results indicate that the epitope of LpMab-12 is Asp49-Pro53. In our previous study we established that the sialylated O-glycan on Thr52 is critical for platelet aggregating activity of hPDPN [21]. Therefore, the data point that the epitope of LpMab-12 contains the sialylated O-glycan on Thr52 in the Asp49-Pro53 sequence of hPDPN.
In this study, we successfully developed LpMab-12 (mouse IgG 1 , kappa), which specifically detects the glycosylation on Thr52 of hPDPN by flow cytometry (Figs 1 and 3), Western blot (Fig 3), and immunohistochemical analysis (Figs 1 and 2). Because this modification was previously shown to be of critical importance for hPDPN-CLEC-2 interaction [2,19], we hypothesized that LpMab-12 might interfere with the hPDPN-binding to CLEC-2. We found that LpMab-12 only partially and weakly reduced the hPDPN binding to hCLEC-2, yet with a higher efficiency than the other anti-hPDPN glycopeptide mAbs (GpMabs), such as LpMab-3 and LpMab-9 (Fig 6). These results indicate that hCLEC-2 might interact with several sialic acids attached to Ser/Thr of hPDPN. Indeed, a novel platelet aggregation-stimulating domain-4 (PLAG4) of hPDPN (Fig 5) was recently suggested [32], further supporting the notion that complex interactions might be required for an optimal association of hPDPN with hCLEC-2.

Conclusion
Our study suggests that LpMab-12 is useful for determining whether hPDPN possesses the site-specific sialylation on Thr52, an important post-translational modification for the association of hPDPN with CLEC-2 and activation of platelet aggregation. Furthermore, the combination of different epitope-specific mAbs, especially GpMabs, might be advantageous for the PDPN-targeting therapies or disease diagnosis.