Pancreatic cancer (PC) ranks as the sixth leading cause of cancer-related mortality worldwide, with 511,000 new cases and 467,000 deaths reported in 2022 (1). Over the last decades, the significance of PC as a public health concern has increased, while mortality rates for other common cancers, including gastric, colorectal, prostate, and breast cancers, have concurrently declined (1). Despite significant advances in cancer therapy, PC remains a challenging disease characterized by late-stage diagnosis and limited therapeutic options, resulting in a 5-year survival rate of only 12% (2). Currently, surgical resection followed by adjuvant chemotherapy is the sole potentially curative therapy available (3, 4). However, because PC is typically asymptomatic and there is a lack of biomarkers with sufficient specificity and sensitivity, most patients present with locally advanced or metastatic disease, rendering them ineligible for surgical intervention (3, 4). Emerging biomarkers and novel immunotherapeutic approaches, such as antibody-drug conjugates (ADCs) and Chimeric Antigen Receptor (CAR) T cells, hold promise for the future of PC treatment.

Glycans, carbohydrate structures attached to a variety of carrier molecules predominantly found on the surface of mammalian cells, play crucial cellular roles and have profound implications in both health and disease (5). Alterations in the glycan profiles of cancer cells have been extensively documented, often correlating with the aggressiveness and progression of cancer (6, 7). In the context of PC, glycan structures such as sialylated and fucosylated Lewis blood group antigens, including Sialyl Lewis a (SLe^a) and Sialyl Lewis x (SLe^x), and short O-glycans like the Tn and Sialyl-Tn (STn) antigens, are known to be upregulated and have been associated with increased invasion and metastasis (8–13). SLe^a, detectable by the CA-19.9 monoclonal antibody (mAb), is the only FDA-approved blood-based tumor biomarker for monitoring PC (14). However, the presence of SLe^a in benign conditions limits its diagnostic application, primarily confining it to a prognostic tool for managing late-stage disease (14). In contrast, STn (NeuAcα2–6-GalNAcα1-O-Ser/Thr), a truncated sialylated O-glycan structure, exhibits selective expression in epithelial tumors, including PC, while being notably absent in normal tissues (10–13, 15–21). Additionally, STn has been shown to play an important role in the malignant phenotype of cancer cells (12, 13, 22, 23). This distinctive profile positions STn as a promising candidate for both cancer biomarker and therapeutic target (24–42). Despite extensive research, accurately mapping and targeting of STn has been hindered by variations in the epitope binding features of available mAbs (39, 43, 44). Most studies on STn in PC have focused on its detection using HB-STn1 (clone 3F1) and TKH2 mAbs, with limited reports employing the B72.3 mAb (10, 11, 17, 45–48). The B72.3 mAb, generated by mice immunization with human breast carcinoma cells, was the first anti-STn mAb described and has been widely used in both tissue-based and serological assays, as well as in preclinical studies (49–53).In this work, we address this research gap by performing a comprehensive study on the expression patterns of STn in normal, premalignant, and malignant lesions of the pancreas using the B72.3 mAb. We further analyzed the association between STn expression and clinicopathological features of PC patients to understand its prognostic and therapeutic potential. Additionally, STn expression was assessed in pancreatic ductal adenocarcinoma (PDAC) patient-derived xenografts (PDXs), a potential tool for testing new anti-STn therapeutic approaches for PC.

Pancreatic cancer remains a significant clinical burden worldwide, characterized by a poor survival rate (1, 2). This dismal outlook can be attributed to several factors, including the silent clinical progression of PC, the lack of early diagnostic tools, and the limited efficacy of currently available treatments (3, 4). Recent advances in immunotherapeutic approaches, such as ADCs and CAR T cell therapy, have shown promise in the cancer field (55, 56). Tumor-associated glycan antigens have emerged as potential molecular targets for immunotherapy due to their distinctive cell surface presentation and restricted expression patterns in malignant cells (57). In particular, the truncated O-glycan STn antigen stands out due to its high expression in epithelial-origin tumors, including PC, while being mostly absent in normal tissues (10–13, 15–21). This cancer-associated expression profile has positioned STn as a compelling therapeutic target for various therapeutic strategies, including radiolabeled antibodies, ADCs, and CAR T cells ( Supplementary Table 2 ). Nonetheless, the challenge in accurately mapping and targeting STn arises from the variability in binding properties of available mAbs, which has impeded the approval of STn-targeted therapies. Current research on STn in PC has primarily used the HB-STn and TKH2 mAbs for detection (10, 17, 45–48, 58). However, there has been limited investigation into one of the most established anti-STn mAbs, B72.3, despite its extensive use in tissue-based, serologic, and preclinical assays (49, 50) (53, 59, 60).

In the present study, we performed a comprehensive mapping of STn expression in normal pancreatic tissues, as well as in premalignant and malignant pancreatic lesions, using the B72.3 mAb. Our results demonstrated a notable absence of STn in normal pancreatic tissues, consistent with previous studies (10, 17, 45, 46, 48, 59, 60). In premalignant pancreatic lesions, STn expression was observed in PanIN lesions but not in chronic pancreatitis. Previous studies have reported diverse findings regarding STn expression within premalignant lesions of the pancreas, ranging from absence to moderate expression (10, 47, 53, 59). The present study reinforces the need for further validation using a larger independent cohort to clarify the role of STn in pancreatic lesions and assess its potential as a biomarker for the early detection of PC.

In the context of PDAC, the most prevalent type of PC, accounting for more than 90% of all pancreatic malignancies, our study revealed the presence of STn in 61.5% of the cases. Our results align with previous studies reporting that STn expression in PDAC is highly variable (10, 17, 45–48, 53, 59, 60). This variability highlights the challenges involved in precisely detecting short O-glycans, potentially arising from different antibody specificities, each recognizing particular epitopes of the STn antigen, influencing detection outcomes across studies. Moreover, disparities in sample size, demographic characteristics of patients, and stages of disease progression within the studied populations may further contribute to the observed differences in STn levels. Heterogeneity in STn expression was observed not only among different tumors but also within individual tumor lesions. Intra-tumor heterogeneity suggests that subsets of tumor cells may express different STn levels. This is consistent with previous studies documenting the heterogenous nature of STn expression in PDAC (48, 59). Considering the poor clinical prognosis linked to STn-positive tumors and the oncogenic features induced by O-glycan truncation (11, 12, 22, 61, 62), these observations suggest that short O-glycans may play a role in a subpopulation of tumor cells, potentially contributing to cancer progression. Regarding the staining pattern of STn within PDAC, we observed the presence of this glycan both in the cytoplasm and on the membrane of cancer cells. The expression of STn on the surface of cancer cells is crucial for effective targeting, emphasizing its potential as a promising therapeutic target in PC.

Additionally, our study revealed the presence of STn in rare PC subtypes, such as pancreatic acinar cell carcinoma and adenosquamous carcinoma. Although STn detection in these rare pathological subtypes of PC suggests its potential as a biomarker and indicates possible avenues for exploring immunotherapeutic approaches, further validation in larger, more comprehensive cohorts is warranted to validate STn expression across rare PC subtypes and to assess its broader clinical relevance.

We further evaluated the potential association between STn expression and clinicopathological variables, revealing a correlation between STn status and pathological diagnosis, as expected. Notably, higher STn expression was observed in patients with stage II PDAC compared to those with stage I tumors. To our knowledge, this is the first evidence of an association between STn expression levels and PDAC staging. These findings suggest that STn may serve as a marker of tumor progression and could potentially hold prognostic implications in PDAC. However, our study did not reveal a correlation between STn status and OS in an exploratory cohort. Given the relatively small size of this exploratory cohort, future research involving larger cohorts with comprehensive follow-up data is warranted to assess STn’s potential as a prognostic marker for survival in PDAC patients.

To validate our findings and explore the potential of STn for therapeutic applications, STn expression was assessed within a preclinical PDX model of PDAC. Among the 12 PDXs evaluated, 10 exhibited positive STn expression (83.3%). Notably, most PDXs consistently sustained STn expression across passages, both in terms of the extend of stained tumor area and the subcellular staining location. This suggests that STn-expressing cancer cells persist following engraftment. However, some variations were observed: one case showed a decrease in STn with successive engraftment cycles, while two PDXs exhibited increasing STn expression in subsequent passages. The dynamic interplay with the tumor microenvironment may contribute to the different STn levels found in tumor cells, highlighting the importance of this truncated O-glycan structure in the cancer context (63, 64). Similar to PDAC patient tissues, variations in staining intensity and patterns were noted between different PDXs, confirming their representativeness of STn expression within the PDAC landscape. The consistent expression of STn in PDX models highlights their potential as valuable tools for preclinical studies and testing of STn-targeted therapies. To our knowledge, this study presents the first evidence of STn mapping in PDAC PDXs, highlighting their utility as preclinical models.

Overall, our study provides a comprehensive evaluation of STn expression using the B72.3 mAb across pancreatic tissues, from normal to premalignant and malignant lesions, and extending to a preclinical model. Despite the inherent challenges associated with mapping STn due to antibody specificity and variability, our findings offer valuable insights into its expression dynamics throughout different stages of pancreatic carcinogenesis. Moreover, our evaluation of STn in rare PC subtypes expands its potential as both a biomarker and therapeutic target for these specific PC subtypes. Additionally, our analysis of STn expression in PDAC PDX models underscores their reliability for preclinical studies and the development of STn-targeted therapies.