N-Glycoproteomic Landscape of Human Lymphoid Cancers Reveals Novel Biomarkers and Potential Therapeutic Targets

Background: Protein glycosylation is one of the most common posttranslational modifications and plays important roles in many biological processes. N-glycosylation facilitates protein trafficking to membranes and their secretion into the extracellular environment. N-glycoproteins are already used as routine biomarkers in clinical practice. These include cluster of differentiation (CD) proteins that define hematopoietic sub-populations and hematologic neoplasms. These proteins also represent important targets for therapeutic antibodies and immunotoxins. Although N-glycoproteins represent an appealing reservoir of potential biomarkers and therapeutic targets, their expression signatures in lymphoid neoplasia has not been established.

Hypothesis: Unbiased characterization of N-glycoproteins of lymphoma cells from different subtypes of lymphoid neoplasia would constitute a valuable resource for identifying rational candidate biomarkers and therapeutic targets.

Materials and Methods: We performed solid phase extraction of glycoproteins followed by identification of deglycosylated peptides using LC-MS/MS on 36 well-characterized cell lines representing 14 different human lymphoid neoplasm subtypes. Label-free spectral counting was used to quantify the relative abundance of N-glycoproteins. In addition, the N-glycoproteome profiles of 8 patient-derived lymphoma samples were compared in a blinded fashion to the profiles of the 36 cell lines. We employed orthogonal methods such as western blotting, flow cytometry and immunohistochemistry on cell lines and primary human lymphoma biopsies to validate several candidates. Pharmacological inhibition and RNAi-mediated silencing were used to functionally validate glycoproteins specific to ALK-positive anaplastic large cell lymphoma (ALCL, ALK+). Tumorigenicity was assessed in xenografts in SCID-BEIGE mice.

Results: In all, 1,155 unique N-glycoproteins were identified from the analysis of the 36 cell lines. Of note, we identified 188 out of the 363 CD proteins listed by the Human Cell Differentiation Molecules consortium. Importantly, we detected virtually all CD proteins currently used for diagnostic evaluation of lymphoid neoplasia. Unsupervised hierarchical clustering of N-glycoproteomic data accurately designated the 36 cell lines according to their respective lineages, cells of origin and appropriate World Health Organization subtypes. Pearson's correlation between the 36 cell lines and the 8 primary biopsies demonstrated that N-glycoproteomic profiles accurately segregated all primary lymphoma samples within their original diagnostic categories. Using tissue microarrays of 644 primary biopsies of mantle cell lymphoma (n=34), follicular lymphoma (n=183), diffuse large B-cell lymphoma (n=245), classical Hodgkin lymphoma (n=150) and ALCL, ALK+ (n=32) we validated the differential expression of CD44, CD276 and paraoxonase 2 (PON2) in mantle cell lymphoma. In addition, our analyses revealed a distinctive cytokine network signature associated with ALCL, ALK+. The evaluation of CD25 expression in a well-characterized cohort of primary lymphoma biopsies (n=28) consisting of ALCL, ALK+ and ALCL, ALK- demonstrated statistically significant correlation between ALK and CD25 expression (p=0.002). In vitro, functional targeting of CD25 using a toxin-conjugated ligand abrogated ALCL, ALK+ cell growth and promoted cell death by apoptosis. This cytokine network signature also included the selective expression of oncostatin M receptor (OSMR) in ALCL, ALK+. We validated OSMR selective expression in ALCL, ALK+ cells by western blot and flow cytometry. Immunohistochemistry on 56 ALCL biopsies (ALK+ and ALK-) revealed a significant correlation between ALK and OSMR expression (p<0.001). Mechanistic studies revealed that OSMR expression was dependent on ALK activity, and mediated via STAT3. RNAi-mediated silencing of OSMR in ALCL, ALK+ cells resulted in decreased colony formation and abrogation of tumor growth in an in vivo xenograft model.

Conclusion: Our data reveal that several glycoproteins demonstrate characteristic and unique expression signatures in distinct forms of human malignant lymphoid neoplasia. Our results indicate that comprehensive profiling of N-glycoproteomes of cells offers opportunities for discovery of novel lymphoma biomarkers and therapeutic targets.