Molecular basis of antibody binding to mucin glycopeptides in lung cancer

Abstract

Glycopeptides bearing Tn epitopes are emerging targets for cancer diagnosis and immunotherapy. In this study, we analyzed membrane proteins containing O-glycosylated tandem repeat (TR) sequences in lung cancer patients of different types and stages, using gene microarray data in public domain. The expression of Tn and glycopeptide epitopes on the surface of lung cancer cell lines were studied by monoclonal IgG antibodies 14A, 16A, and B72.3. The binding of mAbs to synthetic glycopeptides were studied by surface plasmon resonance. Nine mucin mRNAs were found to be expressed in lung cancer patients but at similar level to healthy individuals. At protein level, a glycopeptide epitope on cancer cell surface is preferably recognized by mAb 16A, as compared to peptide-alone (14A) or sugar-alone epitopes (B72.3). 14A and 16A favor clustered TR containing more than three TR sequences, with 10-fold lower Kd than two consecutive TR. B72.3 preferrably recognized clustered sialyl-Tn displayed on MUC1 but not other O-glycoproteins, with 100-fold stronger binding when MUC1 is transfected as a sugar carrier, while the total sugar epitopes remain unchanged. These findings indicate that clusters of both TR backbones and sugars are essential for mAb binding to mucin glycopeptides. Three rules of antibody binding to mucin glycopeptides at molecular level are presented here: first, the peptide backbone of a glycopeptide is preferentially recognized by B cells through mutations in complementarity determining regions (CDRs) of B cell receptor, and the sugar-binding specificity is acquired through mutations in frame work of heavy chain; secondly, consecutive tandem repeats (TR) of peptides and glycopeptides are preferentially recognized by B cells, which favor clustered TR containing more than three TR sequences; thirdly, certain sugar-specific B cells recognize and accommodate clustered Tn and sialyl-Tn displayed on the surface of a mucin but not other membrane proteins.

Figure 1

Expression of mucin mRNA in four subtypes of lung cancer. mRNA array data for all lung cancer types were acquired as described in the text. The 212 array data include the following patients: 150 adenocarcinoma, 20 bronchial carcinoid, 5 small cell lung cancer, 21 squamous lung cancer, and 16 healthy individuals. Array data were processed by Robust Multiarray Average normalization. ^*p<0.05 when the expression level in a specific subtype of lung cancer is compared to healthy controls. Error bars mean standard error.

Figure 2

Expression of mucin mRNA in different stages of lung adenocarcinoma. mRNA array data in different stages of lung adenocarcinoma were acquired as described in the text. 358 array dataset were collected (132 for T1 stage, 188 for T2 stage, 26 for T3 stage, and 12 for T4 stage in stage category; 241 for N0 stage, 64 for N1 stage, 53 for N2 stage in metastasis stage category). (A) T1 to T4 stages of lung adenocarcinoma. (B) N0 to N2 stages (metastasis) of lung adenocarcinoma. No statistically significant differences were found in different stages.

Figure 3

A glycopeptide epitope on lung cancer cell surface is preferably recognized by mAb 16A. Lung adenocarcinoma cell lines, NCI-H1395, HCC4019, H838, H1573, H1703, H2030, and H3255 were studied by flow cytometry staining. Monoclonal antibodies 14A, which binds to MUC1 peptide part only (RPAPGSTAPPAHG); 16A, which binds to MUC1 glycopeptide RPAPGS(GalNAc)TAPPAHG; and B72.3, which binds to sugars only (clustered Tn antigen), were used as primary antibodies. Goat anti-mouse IgG (Allophycocyanin-conjugated), and mouse IgG isotype control were from Southern Biotech (Birmingham, AL, USA).

Figure 4

16A mAb binds to both sugar and peptide parts of a MUC1 glycopeptide. The biotinylated glycopeptide, RPAPGS(GalNAc)TAPPAHG-dPEG™11-Biotin, (1 μg/ml) was bound to streptavidin-coated plates (2 μg/ml) and incubated with 16A monoclonal Ab (mAb) for 2 h. Binding of 16A was visualized by a secondary Ab (goat anti-mouse IgG) followed by colorimetric detection. To measure the inhibitory effects of competing ligands, ligands (GalNAc-BSA, GalNAc, RPAPGS(GalNAc)TAPPAHG, and RPAPGSTAPPAHG) were mixed with the 16A mAb at 0 to 500 μM for 1 h, before incubation with plate-bound glycopeptide RPAPGS(GalNAc)TAPPAHG-dPEG. Data were representative of 3 independent experiments.

Figure 5

MUC1 TR is the preferable backbone for B72.3. Ag104 cells, which express only Tn and sialyl-Tn O-linked glycans, were transfected by a MUC1 gene using a pcDNA3.1-hMUC1-IRES-eGFP plasmid. MUC1-expressing cells were selected by sorting of GFP-positive cells. Stably established MUC1-expressing cells and Ag104 cells transfected by mock pcDNA3.1-IRES-eGFP plasmid were stained by B72.3, 14A, 16A mAb and Sambucus Nigra Lectin specific to α-2,6 linked sialyl acid. Red line indicates the mock transfected Ag104 cells. Blue line indicates the MUC1 transfected Ag104 cells.