Single-Cell Atlas Reveals DLL3 as a Drug Target in Aggressive Cervical Cancer

Neuroendocrine carcinoma of the cervix, or NECC, is among the rarest and most feared forms of cervical cancer. Unlike the adenocarcinomas and squamous cell carcinomas that account for the vast majority of cases, NECC progresses with startling speed, metastasizes early, and carries a grim prognosis. Clinicians currently manage the disease with strategies borrowed from other neuroendocrine malignancies and from conventional cervical cancer protocols, but those borrowed approaches have delivered disappointing results. Now, a team of researchers from Fudan University and the Shanghai Key Laboratory of Female Reproductive Endocrine-Related Diseases has produced one of the most detailed molecular portraits of this tumor type to date, and their work points to a specific, druggable molecule that may finally offer patients a targeted option.

The study, published in the journal Genes & Diseases, combined single-cell RNA sequencing with T-cell receptor sequencing to dissect the cellular landscape of NECC at unprecedented resolution. The researchers analyzed patient-derived surgical tissues, deliberately comparing NECC samples against cervical adenocarcinoma and squamous cell carcinoma controls so that the features unique to the neuroendocrine subtype could be isolated. Tumor tissues were enzymatically dissociated, and viable single cells were identified using acridine orange and propidium iodide staining assessed on a Countstar Fluorescence Cell Analyzer before being isolated for high-resolution transcriptomic profiling. This rigorous quality-control step ensured that the resulting atlas reflected living cells rather than debris or dying tissue, a critical consideration when working with surgically resected tumors.

The resulting single-cell map revealed the full cast of cellular players within the NECC microenvironment, but one population stood out: neuronal progenitor cells, abbreviated NPCs, which appear to drive the malignant character of the tumor. Using trajectory and pseudotime analyses, computational techniques that reconstruct the developmental path of cells from snapshots of thousands of individual cells, the team identified three distinct differentiation states within this progenitor compartment. Each state carried its own functional signature, and together they traced an arc of malignant progression from early differentiation to aggressive proliferation.

The earliest state, termed early differentiated NPCs or edNPCs, was associated with remodeling of the tumor microenvironment, effectively preparing the surrounding tissue to support tumor growth. The intermediate state, transitional differentiated NPCs or tdNPCs, showed hallmark features of glycolytic metabolic reprogramming, the metabolic shift that many cancers use to sustain rapid growth even in oxygen-poor conditions. The final state, late differentiated NPCs or ldNPCs, exhibited overactivation of the cell cycle and the rapid cellular proliferation that defines the aggressive clinical behavior of NECC. Pseudotime analysis captured the dynamic expression of genes including ENTPD2, IFITM3, FAM162A, PFKFB3, CDKN3, and MKI67 across these transitions, and single-cell gene set enrichment analysis, or ssGSEA, defined the active molecular pathways operating within each subpopulation.

But the most consequential discovery came from mapping how these malignant cells communicate with their surroundings. Cellular communication profiling identified delta-like ligand 3, known as DLL3, as a central signaling hub that simultaneously bridges malignant progression and local immunosuppression. DLL3 is a member of the Notch ligand family, and its expression pattern in NECC proved strikingly specific. Immunohistochemical validation confirmed that DLL3 is highly expressed in NECC specimens relative to cervical adenocarcinoma and squamous cell carcinoma tissues, making it a molecular flag that distinguishes this aggressive subtype from its more common counterparts.

The mechanistic story that emerged is one of immune sabotage. DLL3-mediated crosstalk between malignant cells and surrounding tumor-infiltrating lymphocytes, operating primarily through DLL3–NOTCH1 and DLL3–NOTCH2 pathways, promotes T-cell exhaustion, the dysfunctional state in which cytotoxic T cells lose their ability to kill tumor cells. In effect, the tumor uses DLL3 signaling not only to advance its own progression but also to disarm the immune cells that would otherwise destroy it. The researchers further identified a CD38–PECAM1 signaling axis between plasma cells and T cells that reinforces this local immunosuppressive network, adding a second layer to the tumor’s defensive architecture.

Identifying a target is only the first step; demonstrating that it can be exploited therapeutically is what transforms an atlas into a treatment strategy. To that end, the team established a patient-derived NECC organoid and autologous TIL co-culture system, a laboratory model in which fragments of a patient’s tumor are grown as three-dimensional organoids alongside the patient’s own tumor-infiltrating lymphocytes. This system preserves much of the biological complexity of the original tumor and its immune context, making it a powerful platform for testing therapies in a personalized setting before they are ever given to a patient.

When the researchers treated this co-culture system with tarlatamab, also known as AMG757, a DLL3-targeted bispecific T-cell engager, the results were encouraging. Bispecific T-cell engagers are engineered antibodies designed to physically bridge T cells and tumor cells, forcing the immune cells into direct contact with their targets. In the NECC organoid model, tarlatamab induced T-cell activation, elevated the expression of perforin, CD137, and CD107a in CD8+ T cells, all markers of cytotoxic engagement, and decreased the viability of organoid cells. In other words, the drug reawakened the exhausted immune cells and directed them to kill the tumor cells that had been hiding behind DLL3-mediated immunosuppression.

The significance of this work extends beyond a single rare cancer. NECC has long been an orphan of cancer research, too uncommon to attract the large genomic studies that have transformed the treatment of more prevalent tumors. By defining the spatial and transcriptomic architecture of the disease, this study provides the foundational map that targeted therapy development requires. The identification of the DLL3–NOTCH axis as a driver of immune evasion reframes NECC not merely as an aggressive variant of cervical cancer but as a biologically distinct entity whose vulnerability lies in the very signaling it uses to evade the immune system.

Tarlatamab itself has attracted attention in the broader oncology landscape for its activity in small cell lung cancer, another neuroendocrine malignancy characterized by DLL3 expression, and the new findings suggest that this therapeutic class may be repurposed for NECC with a strong biological rationale. Clinical translation will require further validation, but the combination of a specific biomarker, a mechanistic explanation of immune evasion, and a translational model demonstrating drug activity represents the complete chain of evidence needed to justify moving toward patient trials. For patients facing a disease with rapid progression, early metastasis, and poor survival, that chain of evidence offers something that has been in short supply: a precisely aimed therapeutic option grounded in a molecular understanding of their tumor.

Subject of Research: Single-cell multi-omic profiling of neuroendocrine carcinoma of the cervix identifying DLL3 as a therapeutic target

Article Title: Single-cell atlas identifies DLL3 as a therapeutic target in neuroendocrine cervical cancer

Article References: Single-cell atlas identifies DLL3 as a therapeutic target in neuroendocrine cervical cancer. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: neuroendocrine carcinoma of the cervix, single-cell RNA sequencing, DLL3, tarlatamab, tumor immune microenvironment, T-cell exhaustion, NOTCH signaling, organoids, tumor-infiltrating lymphocytes, pseudotime analysis, immunosuppression, precision oncology