Single-cell RNA sequencing reveals tumor heterogeneity in small cell neuroendocrine cervical carcinoma

Xiang, Xuesong; Tao, Xiang; Hua, Keqin; Jiang, Hua; Ding, Jingxin

doi:10.1038/s42003-025-07605-y

Download PDF

Article
Open access
Published: 05 February 2025

Single-cell RNA sequencing reveals tumor heterogeneity in small cell neuroendocrine cervical carcinoma

Communications Biology volume 8, Article number: 184 (2025) Cite this article

5362 Accesses
8 Citations
1 Altmetric
Metrics details

Subjects

Abstract

Small cell neuroendocrine cervical carcinoma (SCNECC) is an aggressive gynecological malignancy with poor prognosis. The precision therapeutic strategies for SCNECC are severely limited by the complex tumor microenvironment. Here, we mapped the single-cell landscape of a total of six samples from matched SCNECC cancerous foci and normal adjacent cervical tissues. Through analysis of 68,455 high-quality cells, malignant epithelial cells were identified with increased neuroendocrine differentiation and reduced keratinization. Within four epithelial cell clusters, the key transcription factors ASCL1, NEUROD1, POU2F3, and YAP1 defined molecular subtypes. Transitional trajectory among subtypes characterized two distinct carcinogenesis pathways in SCNECC. The P-type SCNECC showed potentially enhanced immune infiltration over other subtypes. Intercellular communication analysis identified several immune checkpoints and differentially expressed signaling pathways among subtypes. Through western blotting, the TC-YIK cell line was identified as an N-type SCNECC cell with high expression of SLFN11 and mTOR. Based on immunohistochemical staining of malignant subtyping markers, a cohort of 66 SCNECC patients from our hospital were divided into five subtypes. We further combined YAP1 expression with other clinicopathological factors (Cox p < 0.05) to establish a prognostic nomogram. Overall, these findings provide clues for tumorigenesis, precision treatments and prognostic prediction in SCNECC.

Single-cell discovery of the scene and potential immunotherapeutic target in hypopharyngeal tumor environment

Article Open access 02 December 2022

Heterogeneity of neuroendocrine transcriptional states in metastatic small cell lung cancers and patient-derived models

Article Open access 19 April 2022

Single-cell and spatial proteo-transcriptomic profiling reveals immune infiltration heterogeneity associated with neuroendocrine features in small cell lung cancer

Article Open access 04 September 2024

Introduction

Small cell neuroendocrine cervical carcinoma (SCNECC) is the most common neuroendocrine tumor in female genital tract, but also a rare malignancy comprising less than 5% of cervical cancers^1,2,3,4. Compared to cervical squamous cell carcinoma or adenocarcinoma, a proportion of SCNECC patients show earlier development of distant metastases and higher recurrence rate after primary therapy, leading to much lower five-year survival rate^5,6. The precision treatments for SCNECC are severely limited by our poor understanding of its heterogeneity.

Previous next-generation sequencing (NGS) showed an aggregation of SCNECC somatic mutations in several signaling pathways, but no driver gene was found^7,8,9. Although similarities were found in all neuroendocrine tumors, regardless of primary organs¹⁰, more than 90% of SCNECC were positive for the human papillomavirus 18 (HPV 18)¹¹. As an unusual HPV-associated neuroendocrine tumor, SCNECC may comprise heterogeneous cells in the tumor microenvironment (TME)^{12,13,14,15,16,17}; however, RNA sequencing data provide only the average gene expression for all cells rather than for specific cell types.

In contrast, single-cell RNA sequencing (scRNA-seq) for other cervical cancers and neuroendocrine carcinomas from other organs has identified molecular subtypes associated with these pathologies^15,18. The high-throughput sequencing of mRNA from single cells reveals heterogeneity in the TME that is masked by bulk RNA-seq¹⁹. For the classification of tumor subtypes for precise immune and targeted therapy, scRNA-seq can identify the key molecules associated with tumor heterogeneity. Because SCNECC is rare, scRNA-seq data for fresh tissues of SCNECC patients are not available.

In this study, we pioneered the use of scRNA-seq to map the TME of SCNECC, analyze tumor heterogeneity, and establish tumor subtypes marked by key transcription factors ASCL1, NEUROD1, POU2F3, and YAP1. The transitional trajectory of subtypes simulated the tumorigenesis of SCNECC, and the communication networks of the TME elucidated future personalized treatments for SCNECC patients. Through cell line assays and immunohistochemical (IHC) staining, our study verified the molecular subtypes at the protein level, identified SLFN11 and mTOR upregulated in N-type SCNECC cell, constructed a prognostic model based on YAP1 expression and clinicopathological factors.

Results

The single-cell atlas reveals cellular diversity in SCNECC

After quality control (Supplementary Fig. 1), 68,455 high-quality cells were clustered into five groups (I–V) and annotated as eight cell types (Fig. 1a–c and Supplementary Fig. 2a–c). Large heterogeneity was observed in the distribution of cell types across different tissue sources (Fig. 1d and Supplementary Fig. 2d). The number and proportion of epithelial cells increased from normal adjacent tissues (NATs) to primary tumors (PTs), consistent with SCNECC being an epithelial cell-derived tumor. Immune cell infiltration varied across NAT-PT pairs, indicating the heterogeneity in the immune microenvironment which may influence immunotherapy sensitivity. Stromal cells were largely reduced in PTs compared to NATs, suggesting that SCNECC is a stromal-deficient tumor.

**Fig. 1: Diverse cell types in microenvironment of SCNECC.**

Cell-type-specific upregulated genes were identified through differential gene expression analysis (Fig. 1e). Notably, within the epithelial cell population, the top-upregulated genes were neuroendocrine-related molecules, indicating the presence of both cancer and normal cell subgroups with heterogeneous gene expression patterns. This is exemplified by TAGLN3, KIF1A, and several potassium channel genes that regulate nerve signals, and neural differentiation markers IGSF9 and DLL3, as well as the neurotransmission regulator DDC. The malignant nature of most epithelial cells from PTs was supported by the higher levels of copy number variations (CNVs) compared to other cell types and epithelial cells from NATs (Fig. 1f, g and Supplementary Fig. 2e). Additionally, a heat map of CNVs on each chromosome across all cells (Supplementary Fig. 2f) revealed remarkable variations between PTs or even within the same PT, highlighting the need to define epithelial cell subpopulations to better characterize intra-tumoral and inter-tumoral heterogeneity.

Heterogeneity of SCNECC epithelial cells

A total of 35,920 epithelial cells were clustered into three groups (i–iii, Fig. 2a and Supplementary Fig. 3a) through uniform manifold approximation and projection (UMAP) algorithm, among which group i came mostly from PTs with higher CNV levels and group ii originated primarily from the normal epithelium, whereas group iii constituted both malignant and normal epithelial cells. By employing the shared nearest neighbor (SNN) algorithm based on the UMAP results, these cells were subdivided into 12 subpopulations (Ep1–12) that had distinct sources and CNV levels (Fig. 2b, c and Supplementary Fig. 3b–d), of which Ep1–7 were considered malignant cells and Ep8–12 were normal epithelial cells.

**Fig. 2: Heterogeneity analysis of epithelial cells.**

Analyzing markers of epithelial-derived tumor at the single-cell level confirmed the heterogeneity of epithelial cells that caused the absence of cell type-defining molecules in top-upregulated genes (Fig. 2d)^13,20,21. Markers KRT5, SFN, and KRT6A were highly expressed only in some normal epithelial cells, indicating decreased epithelial differentiation during carcinogenesis. The top differentially expressed genes (DEGs) were compared at the cell group level between malignant and normal epithelia (Fig. 2e). We found a decrease in classical epithelial markers in malignant cells, including the KRT family members, CRNN and SBSN, and an increase in neuroendocrine-related transcription factors, including NEUROD1 and ASCL1, as well as their upstream and downstream genes, including RFX6, INSM1, ST18, MYC and SOX2^11,22. Functional enrichment analysis using DEGs revealed that malignant cells in SCNECC have more complex functions related to oxidative stress, DNA damage repair, immune regulation, and neuroendocrine differentiation compared to normal epithelial cells, providing insights into the malignant transformation process (Fig. 2f and Supplementary Fig. 3e).

According to the similarity in the DEGs profiles of 12 epithelial subpopulations (Fig. 2g), these cells could be classified as four clusters (α, β, γ, δ). The α (NECC_1–4) and β (NECC_5) clusters, especially for NECC_2, NECC_4 and NECC_5, all had an association with neuroendocrine markers ZNF536, DNER, CHGA, RFX6, and VIP which demonstrated their roles of maintaining neuroendocrine differentiation. Besides, the NECC_1 and NECC_3 subpopulations highly expressed genes related with cell cycle regulation and DNA damage repair, including PIF1, PLK1, KIF20A, and DPF1^23,24, indicating their potential key role during tumorigenesis of α cluster. The γ cluster (NECC_6–7) markers showed relationship with terminal epithelial differentiation, represented by ELF5²⁵, and lack of neuroendocrine features. These results demonstrate that malignant cells of SCNECC showed decreased epithelialization and increased neuroendocrine differentiation. Based on the markers expression, 12 subpopulations of epithelial cells were refined as four clusters, among which the key transcriptomic features and tumorigenesis related cell subpopulations were identified waiting further validation.

Molecular subtypes of SCNECC based on epithelial heterogeneity

The gene set variation analysis (GSVA) of Ep1–12 revealed distinct functional profiles among the clusters (Fig. 3a and Supplementary Fig. 4a). The α cluster comprised NECC_1&3 and NECC_2&4 subclusters, which were closely associated with cell cycle regulation and neuroendocrine functions, respectively. The functions of the β cluster mainly involved pathways for immune, endocrine regulation and anti-tumor resistance. The γ cluster, lacking neuroendocrine characteristics, had initial differentiation and cell death pathways, showing an evolutionary association with poorly differentiated epithelial tumors.

**Fig. 3: Establishing molecular subtyping based on epithelial heterogeneity.**

The Single-Cell rEgulatory Network Inference and Clustering (SCENIC) analysis uncovered the regulons formed by transcription factors and their downstream target genes that drive functional differences (Fig. 3b and Supplementary Fig. 4b). There were many neuroendocrine regulons in the α and β clusters, including the DLX and SOX families and their downstream target genes ASCL1 and ZNF536. NECC_6 and 7 in the γ cluster shared a POU2F3-dominated regulon (Fig. 3b), which controls neuroreceptor differentiation in keratinocytes and cell proliferation, supporting the proposed evolutionary origin of this cluster. Through hierarchical clustering horizontal axis of the regulon activity scores heatmap (Fig. 3c), the similarities in the transcriptomic regulation network of subpopulations from the same cluster were confirmed.

The markers of the four clusters were characterized by DEGs analysis (Fig. 3d). The top-upregulated genes of the α cluster, including DLX2, SOX2, and ASCL1, were closely linked to their regulons in the regulatory network (Supplementary Fig. 4b). Many neuroendocrine markers were overexpressed in the β cluster, such as NEUROD1, CHGA, and RFX6. Several markers were co-expressed in the α and β clusters, indicating their differentiation similarities. The γ cluster markers included POU2F3, NCAM1, PDGFRA and CD24, suggesting the origin of differentiation and the potential therapeutic targets.

Molecular subtypes have been identified in neuroendocrine carcinomas that originated from other organs^26,27,28. Drawing on their research, we found that the subtyping markers ASCL1, NEUROD1, POU2F3, and YAP1 showed distinct expression patterns in SCNECC epithelial cells, consistent with the distribution of α, β, γ, δ clusters, respectively (Fig. 3e, f). Therefore, ASCL1, NEUROD1, and POU2F3 could serve as potential molecular markers for subclassifying malignant epithelial cells in SCNECC, whereas YAP1 is expressed predominantly in normal epithelial cells, which is similar to the results of a scRNA-seq study of small cell lung neuroendocrine carcinoma¹⁵. Using molecular subtyping based on these four markers, abbreviated as “A-N-P(Y)” subtyping, we modified the nomenclature of the 12 epithelial subpopulations (Fig. 3f and Supplementary Fig. 4c–d). The subtyping markers expression clustering on the four cell clusters (Supplementary Fig. 4e) suggested similar neuroendocrine differentiation features in A and N cell types²⁸. The three tissue pairs showed distinct expression pattern of A-N-P(Y) subtyping markers (Fig. 3g), confirming that the tumor tissues of the three patients belong to A-type, N-type, and P-type, respectively. Interestingly, according to the proportion of four epithelial clusters in each cancer tissues (Supplementary Fig. 4f), SCNECC_A presented in both N-type and P-type cancer tissues, suggesting potential evolutionary relationships between N-type and P-type with A-type, respectively.

The tumorigenesis model of SCNECC

Monocle2 was used to visualize the evolution trajectory of all epithelial cells, in which normal cells clustered at one end and transitioned gradually into two distinct malignant branches (Fig. 4a). The distribution of Ep1–12 on the evolution trajectory characterized their heterogeneous differentiated states (Fig. 4b and Supplementary Fig. 5a), such as SCNECC_A1&3 that were both relatively far from normal cells, indicating that they were more terminally evolved in α cluster as we guessed. Furthermore, the subpopulations of the same A-N-P(Y) type showed similar evolutionary stages. Based on the expression of subtyping markers along the trajectory (Fig. 4c) and distribution of four epithelial cell clusters (Supplementary Fig. 5b–c), we found that 1) ASCL1 expression increased in both malignant directions, 2) NEUROD1 and POU2F3 were each elevated in one, different malignant direction, and 3) YAP1 was expressed mainly in normal cells and decreased during carcinogenesis. The pseudotime analysis separately for the tissues of three patients (Supplementary Fig. 5d–f) found that N- and P-type cancer cells are interspersed with A-type cells at various evolutionary stages respectively, which hinted at possible transformation relationships and was consistent with the comprehensive analysis of all three samples combined. All epithelial cells were further categorized into five evolutionary states (Fig. 4d), with state 1 corresponding to normal cells. Based on the relationship between A-N-P(Y) subtyping and five states (Supplementary Fig. 5g–h), the transition from state 1 to states 2/3/4 was designated as the “Tu P/A” direction, and from state 1 to state 5 was the “Tu A/N” direction (Fig. 4e).

**Fig. 4: The evolutionary process of SCNECC tumorigenesis.**

Regardless of the normal or cancerous properties of epithelial cells, CytoTRACE independently analyzed their differentiation status (Fig. 4f), confirming that malignant cells showed a lower differentiated state characterized by higher gene count scores (Supplementary Fig. 5i). Among different molecular subtypes (Fig. 4g and Supplementary Fig. 5j–k), P-type cells had the lowest differentiation, followed by the A-type and N-type cells, and finally, the well-differentiated normal cells. Informed by comparisons of 12 subpopulations (Fig. 4h and Supplementary Fig. 5l–m), we identified two subclusters within A-type cells, in which SCNECC_A2&4 cells were highly differentiated and closer to N-type cells, whereas SCNECC_A1&3 cells were poorly differentiated.

The expression patterns of ASCL1, NEUROD1, POU2F3, and YAP1 along two evolutionary directions revealed the position of the corresponding types on the trajectory (Fig. 4i). ASCL1 expression increased initially and then decreased along Tu A/N but increased continuously along Tu P/A, peaking at its end. NEUROD1 upregulation occurred primarily in the second half of Tu A/N among N-type cells. POU2F3 expression increased and then decreased along Tu P/A, suggesting that P-type cells dominated the mid-trajectory of this branch. YAP1 downregulation was associated with a decrease in normal epithelial cells during tumorigenesis. Along with the functional characteristics shown in Fig. 3a and Supplementary Fig. 4a, the Tu P/A path was characterized by the initial dedifferentiation of P-type cells and the eventual neuroendocrine differentiation into partial A-type malignant cells. The Tu A/N path was characterized by high neuroendocrine differentiation throughout the A-type to N-type trajectory. The entire transcriptional network of the two SCNECC evolutionary paths was clustered into seven modules (Fig. 4j), and different stages of the two paths each had their own characteristics in differentiation, oncogenic effect of HPV, cell proliferation, immunosuppression, cell adhesion, autophagy, and apoptosis (Fig. 4k).

The immune landscape of SCNECC

Given the central roles in adaptive or innate immunity, we characterized 7818 T cells and 1645 mononuclear macrophages, which were the most prevalent immune cells in the TME of SCNECC (Supplementary Fig. 6). The distribution of immune cells lacked clear boundaries between NATs and PTs or between different pairs of tissues, suggesting immune heterogeneity driven by changes in the ratios of subpopulation infiltration. The SNN algorithm found T1 − 8 and M1 − 7 subpopulations, identified by upregulated markers and classical markers (Fig. 5a, b and Supplementary Fig. 7)^29,30,31.

**Fig. 5: The heterogeneity analysis of immune microenvironment.**

The immune subpopulations differed between NATs and PTs in proportion rather than in species (Fig. 5c, d). Compared to NATs, PTs showed increased CD8⁺ exhausted T cells, CD4⁺ regulatory T cells, CD4⁺ cytotoxic T cells, CD4⁺ naive T cells, and SPP1⁺ macrophages, whereas natural killer (NK) cells, CD8⁺ cytotoxic T cells, and monocytes decreased. For the enrolled tissue pairs, the P-type was the only pair that immune cell infiltration increased from NAT to PT, its cancerous tissue also had the highest immune infiltration compared to other subtypes (Fig. 5e–h and Supplementary Fig. 8a–b). It was suggested that P-type SCNECC was possibly an immune-rich phenotype, which was consistent with small cell lung neuroendocrine carcinoma^28,32.

Within myeloid cells, two macrophage subpopulations had notable intertumoral heterogeneity (Fig. 5g, h). In the A-type and N-type tissue pairs, SPP1⁺ macrophages, associated with carcinogenesis (Supplementary Fig. 8c–d)^33,34, increased from NATs to PTs, whereas FOLR2⁺ macrophages, characterized as tumor inhibitory³⁵, showed reduced infiltration. In contrast, the P-type tissue pair, with the total infiltration of myeloid cells added from NAT to PT, showed a reduced SPP1⁺ macrophage proportion and increased FOLR2⁺ macrophages. Functional analysis revealed distinct characteristics among myeloid subgroups (Fig. 5i, j), especially the enrichment of FOLR2⁺ macrophages in active immune functions and metabolic reprogramming of SPP1⁺ macrophages that suggested possible polarization. The transcriptional regulatory network identified similarities and differences between two macrophage subpopulations (Supplementary Fig. 9); e.g., the regulons dominated by HES1 and EGR2 were both important in two macrophage subpopulations, whereas EGR1 specifically affected FOLR2⁺ macrophages, and MAFG regulated SPP1⁺ macrophages. More experiments are needed to verify their effects on the functions of macrophage subsets in the microenvironment of SCNECC. Additionally, other myeloid cells, which increased from NAT to PT in P-type tissue pair (Fig. 5g, h), retained immune activation functions (Supplementary Fig. 8c–d and Supplementary Fig. 9), supporting the P-type as a possible immune-rich SCNECC.

Intercellular communication in the TME of SCNECC

We defined the subpopulations of fibroblasts using well-annotated markers (Supplementary Fig. 10)^36,37,38,39. Then, malignant cells, CD8⁺ T cells, macrophages, and fibroblasts from PTs were used to map the ligand-receptor communication network in the TME of SCNECC (Fig. 5k–m and Supplementary Fig. 11). The interactions between CD8⁺ T cells and other cell types mainly comprised cell adhesion, immune response, immunosuppression, and chemotaxis (Fig. 5k and Supplementary Fig. 11b–c). We identified multiple ligand receptors associated with co-stimulation and immune killing, including CRTAM-CADM1, ALCAM-CD6, CD2-CD58, CD74-APP, and CXCR4-CXCL12. Several immunosuppressive interactions were also highly expressed, particularly between exhausted CD8⁺ T cells and cancer cells, including the TIGIT-NECTIN family, LAG3-FGL1, and PPIA-BSG, but without the classical immune checkpoints, such as CD274-PDCD1 and CD86-CTLA4. In addition to adaptive immunity, immunosuppressive pairs such as MERTK-LGALS3 were highly expressed in communication of myeloid cells with other cells (Fig. 5l), indicating the potential role of innate immunity in SCNECC tumor escape⁴⁰. In addition to the immune response, differentially expressed signaling pathways were observed among the molecular subtypes of SCNECC, such as the highly expressed PDGF family in P-type cancer cells, and enhanced WNT signaling in A and N types (Fig. 5l, m), indicating potential subtype-specific targeted therapies. Thus, the complex intercellular crosstalk within the TME of SCNECC should be explored further for therapeutic targets.

Subtyping cell lines and possible clinical value of molecular subtyping

TC-YIK, the only human SCNECC cell line, is characterized by a semi-suspended and semi-adherent growth pattern. In comparing ASCL1, NEUROD1, POU2F3, and YAP1 expression in TC-YIK, keratinocyte (Hacat), cervical epithelial cells (Ect-1), and cervical cancer (Hela and SiHa) cell lines (Fig. 6a–b and Supplementary Fig. 12a–b), we found that NEUROD1 expression was much higher and ASCL1 expression was slightly higher in TC-YIK cells. POU2F3 expression was lower in TC-YIK than SiHa cells. YAP1 expression was weakest in TC-YIK cells and highest in Ect-1 cells, which was consistent with its upregulation in normal epithelial cells, as seen previously. These findings suggested that TC-YIK cells belong to the N-type SCNECC cell line. Additionally, scRNA-seq data revealed heterogeneous expression of genes related to drug treatment among the four SCNECC epithelial cell clusters (Fig. 6c). In all malignant subtypes, mTOR expressed especially highly in type P, whereas TOP2A was upregulated in type A, PDGFRA was expressed specifically in type P, and SLFN11 was expressed predominantly in types P and N. Protein expression analysis (Fig. 6d and Supplementary Fig. 12c) confirmed high expression of SLFN11 in TC-YIK cells, indicating potential sensitivity to PARP inhibitors in N-type patients⁴¹. Expression of mTOR was also detected in TC-YIK cells, supporting the ongoing clinical trials of mTOR inhibitors in SCNECC patients^42,43,44.

**Fig. 6: Application of molecular subtyping and exploration of clinical values.**

In a large cohort of 66 SCNECC patients (Supplementary Table 1), we performed IHC staining for protein expression of ASCL1, NEUROD1, POU2F3, and YAP1 (Table 1). ASCL1 and NEUROD1 were expressed in all patients (H-scores > 1), with ASCL1 showing the highest percentage (65%) of high positive scores (H-scores > 50). POU2F3 was expressed in 35% of patients, and the overall expression of NEUROD1 was higher than POU2F3 (91% vs. 26% for H-scores > 10 in those cases with expression), and some patients showed extremely high expression of POU2F3 (two cases with H-scores > 100). Based on these three markers of malignant cells, patients were classified into five subtypes (Fig. 6e): A⁺ (54%), A⁺N⁺ (26%), N⁺ (3%), A⁺P⁺ (5%), and A⁻N⁻P⁻ (12%). Representative positive cases are shown in Supplementary Fig. 13a. Although several patients expressed two subtyping markers, aligning with the complex evolutionary relationship, 67% (44/66) of patients had one dominant malignant marker (H-scores differences > 50, Fig. 6f). A comparison of the clinicopathological characteristics of each type (Supplementary Table 1) indicated that molecular subtyping may provide an independent grouping system for SCNECC patients (p > 0.05). Further, YAP1 was expressed in all cases (H-scores > 1, Table 1) but predominantly in stromal cells of cancer foci, which was consistent with scRNA-seq data (Fig. 6g). YAP1 expression did not differ significantly among the subtypes of tissues (Fig. 6h), suggesting its limited utility as a marker for malignant cells despite its importance in subtyping.

Table 1 H-scores of ASCL1, NEUROD1, POU2F3, and YAP1 in 66 SCNECC patients

Full size table

Based on the H-scores of the four subtyping markers and the clinicopathological data for the 66 patients, we conducted prognostic analyses with a median follow-up time of 30 months (range 4–59), focusing on recurrence and death (Supplementary Fig. 13b). Using univariate Cox regression analysis (p < 0.05), we found that risk factors for disease-free survival (DFS) included older age, larger tumors, vaginal invasion, deeper myometrial infiltration, and lower YAP1 H-scores (≤50), whereas only older age affected overall survival (OS), and no independent risk factors (p < 0.05) were found by multivariate analysis. Kaplan-Meier analysis (Fig. 6i–j) found a better prognosis for patients with higher expression of YAP1 (H-scores > 50, p = 0.025 for DFS and p = 0.021 for OS). YAP1 high positivity could serve as an independent prognostic variable by comparing clinicopathological characteristics between groups (p > 0.05, Supplementary Table 2). A nomogram using clinicopathological features and prognostic markers was used to predict the three-year recurrence in SCNECC patients (Fig. 6k). For example, a 49-year-old patient with a 2.5 cm tumor, no vaginal involvement, medium myometrial infiltration, and a YAP1 H-score of 40.51 was predicted to have a 46.5% recurrence risk and recurrence at 10 months after surgery. The calibration curve of the nomogram further confirmed its accuracy (Fig. 6l).

Discussion

SCNECC, as the most common gynecological neuroendocrine malignancy, not only shares tissue morphology and biological behavior with neuroendocrine carcinomas across multiple organs¹⁰, but also aroused much interest for its high positive rate of HPV18¹¹. These carcinogenic factors contribute to the complex heterogeneity of SCNECC, making interpretation of NGS difficult. Here, we present the single-cell atlas of SCNECC, identifying the molecular subtypes by the key transcriptional factors in epithelial cell clusters. The subtypes were applied to tumorigenesis modeling. Further, the different characteristics of immune microenvironment and intercellular communication networks among subtypes gave insights into SCNECC biology and possible therapeutics. The subtyping was validated using SCNECC cell line and samples from 66 patients, and provided clues for precision therapy and prognostic prediction.

Recently, molecular subtyping established by lineage-specific transcription factors in multi-organ small cell neuroendocrine carcinoma^26,28,45, especially in the lung, contributed to a better understanding of tumorigenesis and precision therapy^28,46,47. In this study, the heterogeneity of epithelial cells was investigated by their expression profile, functional enrichment, and transcriptional regulatory network. The molecular subtypes in SCNECC were then established by key transcriptional factors ASCL1, NEUROD1, POU2F3, and YAP1 in four epithelial clusters.

The etiology of neuroendocrine carcinoma is uncertain, studies in multiple organs supposed that it may originate from amine precursor uptake and decarboxylation (APUD) cells or by neuroendocrine differentiation within other histological tumor types^{48,49,50,51,52,53}. SCNECC often co-exists with other epithelial tumors in pathological examinations⁵⁴. Recently, a case of organoid experiment with mixed NECC suggested that a complex evolutionary relationship exists between different tumor components^55,56,57. However, the sample rarity precluded the sampling and sequencing of human tissues, and the neuroendocrine characteristics was difficult to maintain during long-term culture of the organoid model after xenotransplantation. We propose an evolutionary model of SCNECC through scRNA-seq of human tissue, including one pathway of dedifferentiated carcinoma cells developing neuroendocrine differentiation and the other pathway leading directly from high neuroendocrine differentiated cells, possibly APUD cells. Malignant cells at different evolutionary stages corresponded to molecular subtypes, which also explained the A⁺N⁺ and A⁺P⁺ types in our cohort of 66 SCNECC patients. It was found that HPV18 integrates into the genome of SCNECC patients at oncogene sites, including MYC, SOX, and NR4A¹¹. This integration not only fuels carcinogenesis in poorly differentiated tumors, but also induces transcription factors such as ASCL1 and INSM1, promoting neuroendocrine differentiation. In our study, along with a rapid decrease in immune activity, carcinogenesis of HPV was enriched mainly in the whole process of dedifferentiated carcinoma developing, as well as in the initiation of the high neuroendocrine differentiated tumor. The malignant transformation of the epithelial cells that comprise the different SCNECC subtypes was largely mediated by two carcinogenic factors: abnormal expression of multiple transcription factors and sustained infection with HPV.

With low PD-L1 expression, tumor mutation burden, and stable expression of mismatch repair protein compared to cervical squamous cell carcinoma, SCNECC has been considered to be resistant to immunotherapy^58,59. However, there were case reports of effective immunotherapy in SCNECC^60,61, and several related clinical trials are in progress^62,63,64. Furthermore, based on the success of clinical trials in small cell lung cancer^65,66, PD-L1 inhibitors Atezolizumab and Durvalumab have been recommended for recurrent or metastatic SCNECC in the 2023 cervical cancer guidelines of the National Comprehensive Cancer Network. Our understanding of the SCNECC immune microenvironment lags behind clinical practice due to the rarity of the disease and the limitations of bulk RNA sequencing. We analyzed the heterogeneity of the major immune cell subsets in SCNECC at the single-cell level. Among three tissue pairs, the P-type tumor had immune-inflamed or “hot tumor” traits, whereas the A and N types had “cold tumor” traits, and prior research on neuroendocrine carcinomas from other organs also showed this pattern^28,32. A higher degree of neuroendocrine differentiation corresponded to lower immune infiltration, sometimes reaching an “immune desert” state. Despite less differentiation, the P-type tumor had richer immune infiltration, potentially enhancing its sensitivity to immunotherapy. Besides, some terminated clinical trials have shown that SCNECC did not respond well to PD-1 inhibitors like pembrolizumab⁶². In our cell communication network, it was also the emerging immune checkpoints, such as TIGIT and MERTK, showed higher expression than classical immune escape receptors, indicating a direction for developing immunotherapeutics in SCNECC.

The morphology of neuroendocrine cancer cell lines is associated with the degree of neuroendocrine differentiation²⁸; weaker adhesion indicates a greater degree of differentiation, whereas tighter adhesion indicates primitive mesenchymal differentiation. We identified semi-suspended and semi-adherent cultured TC-YIK cells. By comparing the expression of subtyping markers with cervical cancer and normal epithelial cell lines, we identified TC-YIK as an N-type SCNECC cell line with both neuroendocrine and primary differentiation characteristics. Coincidentally, TC-YIK has been used as a pancreatic endocrine model due to the similarities in function²², and NEUROD1 was identified as the main transcription factor that regulated downstream target genes using chromatin immunoprecipitation and knockdown experiments.

To reduce the bias of small numbers, we extended our study to a large cohort of 66 SCNECC patients. These patients were divided into five subtypes based on the malignant markers ASCL1, NEUROD1, and POU2F3 in IHC staining. Interestingly, in the course of this study, the A-N-P(Y) subtyping groups were also identified through IHC in another study of 45 SCNECC patients from South Korea⁶⁷. We found higher levels of ASCL1 compared to the South Korean study, and identified A⁺N⁺ and A⁺P⁺ types, which suggested a complex evolutionary relationship between different malignant subtypes and the normal epithelium of SCNECC. Based on our scRNA-seq data and cell line assay, YAP1 was not a marker of malignant epithelial cells; therefore, unlike the Korean study, we did not use YAP1 in subtyping patients. YAP1 showed no differential expression across five subtypes in our study, was highly expressed in stromal cells, and was downregulated in malignant cells, similar to the results of the scRNA-seq analysis. Recently, YAP1 silencing via histone deacetylase (HDAC) was shown to promote the metastasis of small cell lung neuroendocrine carcinoma⁶⁸, and HDAC inhibitors could stimulate YAP1 expression and inhibit the metastasis. Through our prognostic analysis, we found that low positive YAP1 was associated with faster postoperative recurrence and death (p = 0.025 for DFS and p = 0.021 for OS), suggesting a similar tumor-promoting role in SCNECC. Therefore, it can be speculated that small molecular drugs, such as HDAC inhibitors, may also be effective in the treatment of SCNECC, which is waiting for further experimental validation. Additionally, the positivity rate for POU2F3 was the lowest (5%) in the combined Korean cohort and ours (111 cases). Since the P-type SCNECC is so rare, it is easy to understand that although PDGFRA was highly expressed in P-type cells, a recent study found no activation mutations of PDGFRA in 12 SCNECC patients⁶⁹. Improving molecular subtyping of SCNECC will provide a better understanding of tumor evolution, carcinogenesis mechanism, and precision therapy.

The advances in single-cell technology have improved cell capture rates and characterized the TME of SCNECC with unprecedented clarity. Nevertheless, the rarity of SCNECC poses sampling challenges, and the absence of relevant sequencing and follow-up data in public databases hinders deeper bioinformatic analysis. TC-YIK is the only SCNECC cell line, thus limiting in vitro experiments. For future studies, expanding sample sizes and follow-up cohorts, as well as establishing more SCNECC cell lines will be essential for validating and expanding our findings.

In summary, here we present the single-cell landscape of SCNECC. Through analyzing key transcription factors of epithelial cell clusters, we established tumor subtypes and enhanced our understanding of tumor etiology. In the immunosuppressive TME of SCNECC, P-type tumor was possibly with richer immune infiltration during evolution. Diverse cell communication pathways provide clues for developing precision immune and targeted therapies among subtypes of SCNECC. Furthermore, the protein expression of cell lines and a cohort of 66 patients validated the A-N-P(Y) subtypes, laying the foundation for the development of precision therapeutics and prognostic prediction.

Methods

Collection and preparation of single-cell suspensions from SCNECC samples

This study was approved by the Ethics Committee of the Obstetrics and Gynecology Hospital of Fudan University (license no. 2023-41). Informed consent was obtained from all enrolled patients. All ethical regulations relevant to human research participants were followed. Cervical samples from treatment-naive SCNECC patients were collected between March and May 2023, including three pairs of PTs and NATs, which were numbered E1–3 and C1–3, respectively. The fresh samples were immersed in phosphate-buffered saline, and transported to the laboratory at 4 °C within 30 min. The samples were cut into 1 mm³ pieces and incubated in a collagenase solution at 37 °C for an hour¹⁸. Undissolved tissue was removed with a 40 μm filter, and the sample was treated with red blood cell lysis buffer. An acridine orange/propyl iodide assay was used to measure active cells, and only samples with >85% active cells were used for 10× Genomics scRNA-seq.

Construction and quality control of the 10× Genomics scRNA-seq library

A suspension of 20,000 cells from each tissue sample was loaded onto the 10× Genomics platform to generate single-cell Gel Beads-in-Emulsion (GEMs). Cells in GEMs were broken to connect their mRNA with the barcodes in the Gel Beads. After reverse transcription, the cDNA library for each single cell carries a unique barcode. Then, all single-cell cDNA libraries were combined and sequenced, and barcodes were identified through data analysis to distinguish the specific cell source.

The read alignments in fastq format and primary quality control data were processed with CellRanger software (v7.0.1) using the GRCh38 human reference genome. Through the Seurat package (v4.0.0) in R⁷⁰, unique molecular identifiers (UMIs) representing mRNAs of different genes were identified, and low-quality cells were discarded, including cells with UMI number <1000, gene number <200, percentage of mitochondrial-derived UMI counts >15%, or percentage of hemoglobin RNA UMI counts >5%. DoubletFinder (version 2.0.2) was used to identify doublets. Gene expression for each cell was normalized to total expression through the NormalizeData function of the Seurat package, and normalized counts were multiplied by a scaling factor (1000) and log-transformed.

scRNA-seq data analysis

To remove the batch effects in scRNA-seq data, the mutual nearest neighbor (MNN) was performed with the R package batchelor⁷¹. Then, cells were visualized using the UMAP algorithm through the RunUMAP function in Seurat. Besides, to optimize clustering results, we utilized the SNN algorithm, which leveraged the UMAP results as a preprocessing step rather than re-normalization or re-dimensionality reduction of the data, to assess the similarity between single cells and obtain a more precise delineation of cell subpopulations. The annotation of cell clusters was performed through the SingleR package (v1.4.1) and checked manually for curated markers⁷². The DEGs between cell clusters were identified using the FindMarkers function (presto test, p value < 0.05, |log₂ fold change | > 1) in Seurat. All the heatmaps in this study were normalized using the Z-score algorithm, specifically implemented through the ScaleData function from the Seurat package. Based on the hypergeometric distribution algorithm, GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) analysis was performed using the clusterProfiler package (version 4.1). Functional comparisons between cell clusters were performed by GSVA. Through the inferCNV package, the CNV density was compared to distinguish malignant and normal cells⁷³. Through SCENIC algorithm in pySCENIC (version 0.10.2) software, gene regulatory networks for various cell types or clusters were identified⁷⁴. Monocle2 (v2) and CytoTRACE packages^75,76 were used for pseudotime analysis to construct the evolutionary trajectory of cell subpopulations. The correlated ligand-receptor pairs were identified from the scRNA-seq data by the CellPhoneDB⁷⁷.

Cell culture and western blotting

Human SCNECC cell line TC-YIK was from Typical Culture Collection Center in Wuhan, whereas human cervical epithelial cell line Ect-1, human keratinocytes cell line Hacat, and human cervical cancer cell lines Hela and SiHa were from Authenticated Cell Cultures Collection in Shanghai. All cell lines were mycoplasma-free. The cell pellet for each 10-cm culture dish or T25 flask was lysed in 100 μL RIPA buffer (Beyotime), the protein was quantified by the BCA protein assay kit (New Cell & Molecular Biotech) and denatured at 100 °C for 10 min. Equal amounts of samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a polyvinylidene fluoride membrane (Millipore) for immunoblotting. The primary antibodies (listed in Supplementary Table 3) were diluted with buffer (P0023A, Beyotime), then used for incubation overnight at 4 °C, followed by corresponding secondary antibody for 1 hour at room temperature. Chemiluminescence was detected by the ECL Detection Reagent (New Cell & Molecular Biotech).

Immunohistochemistry of formalin-fixed, paraffin-embedded (FFPE) sections

From Jan 2017 to Dec 2021, tissues from 66 SCNECC patients were obtained to make FFPE sections for IHC staining in the Obstetrics and Gynecology Hospital of Fudan University. The diagnoses were confirmed by two pathologists who specialize in gynecological tumors. The consensus was that SCNECC was characterized by hyperchromatic nuclei and sparse cytoplasm, with common nuclear molding and crush artifacts. FFPE sections were immersed into dewax solution (G1128, Servicebio) three times for 10 min each, then in ethanol three times for 5 min each, and rinsed with deionized H₂O. Samples were treated with heat for antigen retrieval, and the endogenous peroxidase activity was quenched with 3% H₂O₂. The sections were blocked with 3% bovine serum albumin for 30 min at room temperature, then incubated with primary antibodies (listed in Supplementary Table 3 and diluted with phosphate buffered saline) overnight at 4 °C, followed by a secondary horseradish peroxidase-labeled antibody (GB23303, Servicebio) for 50 min at room temperature. Color was developed with 3, 3’-diaminobenzidine solution, and hematoxylin was used to counterstain. The stained tissues were dehydrated, mounted, and examined by light microscopy by two researchers who independently confirmed positive results as a brownish-yellow color.

Analysis of immunohistochemistry data

The IHC profiler plugin of Image J software was used to process the image of the SCNECC site in each section. Based on the percentage of high positive (3 points), positive (2 points), low positive (1 point), and negative (0 point) regions, H-scores were calculated. Any marker with an H-score > 1 was considered expressed, whereas a highly positive marker had an H-score > 50.

Clinicopathological characteristics and prognosis analysis

Medical records of 66 SCNECC patients were used to collect clinicopathological information. Preoperative HPV infection was determined by Roche cobas 4800 HPV system (Roche Molecular Systems Inc, Pleasanton, CA). DFS and OS were measured from the date of surgery to the recurrence or death due to SCNECC, or censored at the last follow-up. Prognostic relevance was determined by Cox regression and Kaplan-Meier analysis using R packages survival and survminer. The R package regplot was used to construct the nomogram model and draw the calibration curve.

Statistics and reproducibility

For scRNA-seq data, the experimental samples were divided into two groups: tumor tissues (NECC_E1–3) and normal adjacent tissues (NECC_C1–3). The single-cell data underwent quality control using CellRanger and Seurat, resulting in a range of 9648 to 13,143 high-quality cells per sample, with sequencing quality exceeding 90%. The gene expression profile of 68,455 cells was constructed for downstream bioinformatics analysis, which were performed using R (version 4.2.1, R Foundation for Statistical Computing, Vienna, Austria). The statistical methods outlined in the specific packages were utilized. For western blotting, the TC-YIK cell line was used to validate the subtyping markers at the protein expression level, the Hela and SiHa were served as control group 1, representing cervical cancer cell lines, while Ect-1 and Hacat served as control group 2, representing epithelial cell lines. The expression of each marker was verified through more than three independent replicates. To further validate the subtypes, FFPE sections from 66 patients were tested through IHC staining. The expression level of each marker in each patient was obtained by averaging the results from three sections of the same patient.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The raw sequence data have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in National Genomics Data Center (Nucleic Acids Res 2024), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA-Human: HRA009572) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa-human^78,79. The source data for the graphs has been uploaded on GitHub (https://github.com/Cedar98/SCNECC_single_cell_analysis.git).

Code availability

The code used to analyze our single-cell sequencing data have been made available on GitHub (https://github.com/Cedar98/SCNECC_single_cell_analysis.git). Any additional information will be available from the corresponding author, J.D. (djxdd@sina.com), upon reasonable request.

References

Salvo, G., Gonzalez Martin, A., Gonzales, N. R. & Frumovitz, M. Updates and management algorithm for neuroendocrine tumors of the uterine cervix. Int. J. Gynecol. Cancer 6, 986–995 (2019).
Article Google Scholar
Chen, J. et al. Prognostic factors and treatment of neuroendocrine tumors of the uterine cervix based on the FIGO 2018 staging system: a single-institution study of 172 patients. PeerJ. 9, e11563 (2021).
Ishikawa, M. et al. Prognostic factors and optimal therapy for stages I-II neuroendocrine carcinomas of the uterine cervix: a multi-center retrospective study. Gynecol. Oncol. 1, 139–146 (2018).
Article Google Scholar
Chu, T. et al. The prognosis of patients with small cell carcinoma of the cervix: a retrospective study of the SEER database and a Chinese multicentre registry. Lancet Oncol. 6, 701–708 (2023).
Article Google Scholar
Yuan, L. et al. Prognostic factors of surgically treated early-stage small cell neuroendocrine carcinoma of the cervix. Int. J. Gynecol. Cancer 7, 1315–1321 (2015).
Article Google Scholar
Satoh, T. et al. Gynecologic cancer InterGroup (GCIG) consensus review for small cell carcinoma of the cervix. Int. J. Gynecol. Cancer 3, S102–S108 (2014).
Article Google Scholar
Pei, X. et al. The next generation sequencing of cancer-related genes in small cell neuroendocrine carcinoma of the cervix. Gynecol. Oncol. 3, 779–786 (2021).
Article Google Scholar
Frumovitz, M. et al. Sequencing of mutational hotspots in cancer-related genes in small cell neuroendocrine cervical cancer. Gynecol. Oncol. 3, 588–591 (2016).
Article Google Scholar
Eskander, R. N. et al. Unique genomic landscape of high-grade neuroendocrine cervical carcinoma: implications for rethinking current treatment paradigms. JCO Precis. Oncol. 4, PO.19.00248 (2020).
Stelwagen, J., de Vries, E. G. E. & Walenkamp, A. M. E. current treatment strategies and future directions for extrapulmonary neuroendocrine carcinomas: a review. JAMA Oncol. 5, 759–770 (2021).
Article Google Scholar
Wang, X. et al. Human papillomavirus integration perspective in small cell cervical carcinoma. Nat. Commun. 1, 5968 (2022).
Article Google Scholar
Li, C., Hua, K. Dissecting the single-cell transcriptome network of immune environment underlying cervical premalignant lesion, cervical cancer and metastatic lymph nodes. Front. Immunol. 13, 897366 (2022).
Li, C. et al. Diverse intratumoral heterogeneity and immune microenvironment of two HPV-related cervical cancer types revealed by single-cell RNA sequencing. J. Med. Virol. 6, e28857 (2023).
Article Google Scholar
Li, C., Guo, L., Li, S. & Hua, K. Single-cell transcriptomics reveals the landscape of intra-tumoral heterogeneity and transcriptional activities of ECs in CC. Mol. Ther. Nucleic Acids. 24, 682–694 (2021).
Tian, Y. et al. Single-cell transcriptomic profiling reveals the tumor heterogeneity of small-cell lung cancer. Signal Transduct. Target Ther. 1, 346 (2022).
Article Google Scholar
Zhou, Y. et al. Single-cell RNA sequencing reveals spatiotemporal heterogeneity and malignant progression in pancreatic neuroendocrine tumor. Int. J. Biol. Sci. 14, 3760–3775 (2021).
Article Google Scholar
Dong, B. et al. Single-cell analysis supports a luminal-neuroendocrine transdifferentiation in human prostate cancer. Commun. Biol. 1, 778 (2020).
Article Google Scholar
Li, C. et al. Single-cell transcriptomics reveals cellular heterogeneity and molecular stratification of cervical cancer. Commun. Biol. 1, 1208 (2022).
Article Google Scholar
Lei, Y. et al. Applications of single-cell sequencing in cancer research: progress and perspectives. J. Hematol. Oncol. 1, 91 (2021).
Article Google Scholar
Puram, S. V. et al. Single-cell transcriptomic analysis of primary and metastatic tumor ecosystems in head and neck cancer. Cell 7, 1611–1624.e24 (2017).
Article Google Scholar
Kim, N. et al. Single-cell RNA sequencing demonstrates the molecular and cellular reprogramming of metastatic lung adenocarcinoma. Nat. Commun. 1, 2285 (2020).
Article Google Scholar
Lizio, M. et al. Mapping mammalian cell-type-specific transcriptional regulatory networks using KD-CAGE and ChIP-seq data in the TC-YIK cell line. Front. Genet. 6, 331 (2015).
Moore, X. T. R., Gheghiani, L. & Fu, Z. The role of polo-like kinase 1 in regulating the forkhead box family transcription factors. Cells. 12, 1344 (2023).
Jin, Z. et al. Expression, regulating mechanism and therapeutic target of KIF20A in multiple cancer. Heliyon 2, e13195 (2023).
Article Google Scholar
Luk, I. Y., Reehorst, C. M. & Mariadason, J. M. ELF3, ELF5, EHF and SPDEF transcription factors in tissue homeostasis and cancer. Molecules. 23, 2191 (2018).
Cejas, P. et al. Subtype heterogeneity and epigenetic convergence in neuroendocrine prostate cancer. Nat. Commun. 1, 5775 (2021).
Article Google Scholar
Stone, L. Subtypes in small cell bladder cancer. Nat. Rev. Urol. 9, 521 (2023).
Article Google Scholar
Rudin, C. M. et al. Molecular subtypes of small cell lung cancer: a synthesis of human and mouse model data. Nat. Rev. Cancer 5, 289–297 (2019).
Article Google Scholar
Yu, X. et al. Unravelling the heterogeneity and dynamic relationships of tumor-infiltrating T cells by single-cell RNA sequencing analysis. J. Leukoc. Biol. 6, 917–932 (2020).
Article Google Scholar
Zheng, L. et al. Pan-cancer single-cell landscape of tumor-infiltrating T cells. Science 6574, abe6474 (2021).
Article Google Scholar
Zhang, Q. et al. Landscape and dynamics of single immune cells in hepatocellular carcinoma. Cell 4, 829–845.e20 (2019).
Article Google Scholar
Wang, W. Z. et al. Small cell lung cancer: subtypes and therapeutic implications. Semin. Cancer Biol.2, 543–554 (2022).
Article CAS Google Scholar
Zhang, Y. et al. SLC2As as diagnostic markers and therapeutic targets in LUAD patients through bioinformatic analysis. Front. Pharmacol. 13, 1045179 (2022).
Qi, J. et al. Single-cell and spatial analysis reveal interaction of FAP(+) fibroblasts and SPP1(+) macrophages in colorectal cancer. Nat. Commun. 1, 1742 (2022).
Article Google Scholar
Nalio Ramos, R. et al. Tissue-resident FOLR2(+) macrophages associate with CD8(+) T cell infiltration in human breast cancer. Cell 7, 1189–1207.e25 (2022).
Article Google Scholar
Buechler, M. B. et al. Cross-tissue organization of the fibroblast lineage. Nature 7860, 575–579 (2021).
Article Google Scholar
Geldhof, V. et al. Single cell atlas identifies lipid-processing and immunomodulatory endothelial cells in healthy and malignant breast. Nat. Commun. 1, 5511 (2022).
Article Google Scholar
Kalucka, J. et al. Single-cell transcriptome atlas of murine endothelial cells. Cell 4, 764–779.e20 (2020).
Article Google Scholar
Luzina, I. G. et al. Regulator of cell cycle protein (RGCC/RGC-32) protects against pulmonary fibrosis. Am. J. Respir. Cell Mol. Biol. 2, 146–157 (2022).
Article Google Scholar
Zhou, Y. et al. Blockade of the phagocytic receptor MerTK on tumor-associated macrophages enhances P2X7R-dependent STING activation by tumor-derived cGAMP. Immunity 2, 357–373.e9 (2020).
Article Google Scholar
Xiong, J., Barayan, R., Louie, A. V. & Lok, B. H. Novel therapeutic combinations with PARP inhibitors for small cell lung cancer: a bench-to-bedside review. Semin. Cancer Biol. 2, 521–542 (2022).
Article Google Scholar
Zhang, Z. & Wang, M. Advanced research of mTOR and lung carcinoid tumors. Zhongguo Fei Ai Za Zhi 1, 43–47 (2013).
Google Scholar
Yao, J. C. et al. Everolimus for advanced pancreatic neuroendocrine tumors. N. Engl. J. Med. 6, 514–523 (2011).
Article Google Scholar
Levy, S. et al. First-line everolimus and cisplatin in patients with advanced extrapulmonary neuroendocrine carcinoma: a nationwide phase 2 single-arm clinical trial. Ther. Adv. Med. Oncol. 14, 17588359221077088 (2022).
Feng, M. et al. Identification of lineage-specific transcriptional factor-defined molecular subtypes in small cell bladder cancer. Eur. Urol. 85, 523–526 (2023).
Baine, M. K. et al. SCLC Subtypes Defined by ASCL1, NEUROD1, POU2F3, and YAP1: A Comprehensive Immunohistochemical and Histopathologic Characterization. J. Thorac. Oncol. 12, 1823–1835 (2020).
Article Google Scholar
Lissa, D. et al. Heterogeneity of neuroendocrine transcriptional states in metastatic small cell lung cancers and patient-derived models. Nat. Commun. 1, 2023 (2022).
Article Google Scholar
Taube, E. T. et al. Prognostic impact of neuroendocrine differentiation in high-grade serous ovarian carcinoma. Virchows Arch. 3, 333–342 (2015).
Article Google Scholar
Kleist, B. & Poetsch, M. Neuroendocrine differentiation: The mysterious fellow of colorectal cancer. World J. Gastroenterol. 41, 11740–11747 (2015).
Article Google Scholar
Bogina, G. et al. Neuroendocrine differentiation in breast carcinoma: clinicopathological features and outcome. Histopathology 3, 422–432 (2016).
Article Google Scholar
Santoni, M. et al. Neuroendocrine differentiation in prostate cancer: novel morphological insights and future therapeutic perspectives. Biochim. Biophys. Acta 2, 630–637 (2014).
Google Scholar
Beltran, H. et al. Divergent clonal evolution of castration-resistant neuroendocrine prostate cancer. Nat. Med. 3, 298–305 (2016).
Article Google Scholar
Frazier, S. R., Kaplan, P. A. & Loy, T. S. The pathology of extrapulmonary small cell carcinoma. Semin. Oncol. 1, 30–38 (2007).
Article Google Scholar
Gibbs, J. et al. Clinicopathologic features, incidence, and survival trends of gynecologic neuroendocrine tumors: a SEER database analysis. Am. J. Obstet. Gynecol. 1, 53.e1–53.e6 (2019).
Article Google Scholar
Voutsadakis, I. A. Large cell neuroendocrine carcinoma of the ovary: A pathologic entity in search of clinical identity. World J. Clin. Oncol. 2, 36–38 (2014).
Article Google Scholar
McCluggage, W. G., Kennedy, K. & Busam, K. J. An immunohistochemical study of cervical neuroendocrine carcinomas: Neoplasms that are commonly TTF1 positive and which may express CK20 and P63. Am. J. Surg. Pathol. 4, 525–532 (2010).
Article Google Scholar
Masuda, M. et al. Clonal origin and lineage ambiguity in mixed neuroendocrine carcinoma of the uterine cervix. Am. J. Pathol. 194, 415–429 (2023).
Hillman, R. T. et al. Comparative genomics of high grade neuroendocrine carcinoma of the cervix. PLoS One 6, e0234505 (2020).
Article Google Scholar
Carroll, M. R. et al. Evaluation of PARP and PDL-1 as potential therapeutic targets for women with high-grade neuroendocrine carcinomas of the cervix. Int. J. Gynecol. Cancer 9, 1303–1307 (2020).
Article Google Scholar
Paraghamian, S. E., Longoria, T. C. & Eskander, R. N. Metastatic small cell neuroendocrine carcinoma of the cervix treated with the PD-1 inhibitor, nivolumab: a case report. Gynecol. Oncol. Res. Pract. 4, 3 (2017).
Paterniti, T. A. et al. Complete response to combination nivolumab and ipilimumab in recurrent neuroendocrine carcinoma of the cervix. Obstet. Gynecol. 5, 813–816 (2021).
Google Scholar
Frumovitz, M. et al. Phase II study of pembrolizumab efficacy and safety in women with recurrent small cell neuroendocrine carcinoma of the lower genital tract. Gynecol. Oncol. 3, 570–575 (2020).
Article Google Scholar
Patel, S. P. et al. A phase II basket trial of dual anti-CTLA-4 and anti-PD-1 blockade in rare tumors (DART SWOG 1609) in patients with nonpancreatic neuroendocrine tumors. Clin. Cancer Res. 10, 2290–2296 (2020).
Article Google Scholar
Patel, S. P. et al. A phase II basket trial of dual anti-CTLA-4 and anti-PD-1 blockade in rare tumors (DART) SWOG S1609: high-grade neuroendocrine neoplasm cohort. Cancer 17, 3194–3201 (2021).
Article Google Scholar
Horn, L. et al. First-line atezolizumab plus chemotherapy in extensive-stage small-cell lung cancer. N. Engl. J. Med. 23, 2220–2229 (2018).
Article Google Scholar
Paz-Ares, L. et al. Durvalumab plus platinum-etoposide versus platinum-etoposide in first-line treatment of extensive-stage small-cell lung cancer (CASPIAN): a randomised, controlled, open-label, phase 3 trial. Lancet 10212, 1929–1939 (2019).
Article Google Scholar
Kim, G. et al. Application of small cell lung cancer molecular subtyping markers to small cell neuroendocrine carcinoma of the cervix: NEUROD1 as a poor prognostic factor. Am. J. Surg. Pathol. 48, 364–372 (2023).
Wu, Z. et al. YAP silencing by RB1 mutation is essential for small-cell lung cancer metastasis. Nat. Commun. 1, 5916 (2023).
Article Google Scholar
Chang, K. L. et al. Mutational analysis of PDGFRA oncogene in high-grade neuroendocrine carcinoma of the uterine cervix in twelve Taiwanese women. Taiwan J. Obstet. Gynecol. 1, 125–126 (2024).
Article Google Scholar
Butler, A. et al. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol. 5, 411–420 (2018).
Article Google Scholar
Haghverdi, L., Lun, A. T. L., Morgan, M. D. & Marioni, J. C. Batch effects in single-cell RNA-sequencing data are corrected by matching mutual nearest neighbors. Nat. Biotechnol. 5, 421–427 (2018).
Article Google Scholar
Aran, D. et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat. Immunol. 2, 163–172 (2019).
Article Google Scholar
Wang, R. et al. Single-cell dissection of intratumoral heterogeneity and lineage diversity in metastatic gastric adenocarcinoma. Nat. Med. 1, 141–151 (2021).
Article Google Scholar
Van de Sande, B. et al. A scalable SCENIC workflow for single-cell gene regulatory network analysis. Nat. Protoc. 7, 2247–2276 (2020).
Article Google Scholar
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 4, 381–386 (2014).
Article Google Scholar
Gulati, G. S. et al. Single-cell transcriptional diversity is a hallmark of developmental potential. Science 6476, 405–411 (2020).
Article Google Scholar
Vento-Tormo, R. et al. Single-cell reconstruction of the early maternal-fetal interface in humans. Nature 7731, 347–353 (2018).
Article Google Scholar
Chen, T. et al. The genome sequence archive family: toward explosive data growth and diverse data types. Genom. Proteom. Bioinform. 19, 578–583 (2021).
CNCB-NGDC Members and Partners. Database Resources of the National Genomics Data Center, China National Center for Bioinformation in 2024. Nucleic Acids Res. D1, D18–d32 (2024).

Download references

Acknowledgements

We thank OE Biotech Co., Ltd (Shanghai, China) for assistance with single-cell RNA-seq. This study was supported by the Natural Science Foundation of Shanghai, China (21ZR1410400).

Author information

These authors contributed equally: Xuesong Xiang, Xiang Tao.

Authors and Affiliations

Department of Gynecological Oncology, The Obstetrics and Gynecology Hospital of Fudan University, Shanghai, P. R. China
Xuesong Xiang, Keqin Hua, Hua Jiang & Jingxin Ding
Shanghai Key Laboratory of Female Reproductive Endocrine-related Diseases, The Obstetrics and Gynecology Hospital of Fudan University, Shanghai, P. R. China
Xuesong Xiang, Keqin Hua, Hua Jiang & Jingxin Ding
Department of Pathology, The Obstetrics and Gynecology Hospital of Fudan University, Shanghai, P.R. China
Xiang Tao

Authors

Xuesong Xiang
View author publications
Search author on:PubMed Google Scholar
Xiang Tao
View author publications
Search author on:PubMed Google Scholar
Keqin Hua
View author publications
Search author on:PubMed Google Scholar
Hua Jiang
View author publications
Search author on:PubMed Google Scholar
Jingxin Ding
View author publications
Search author on:PubMed Google Scholar

Contributions

J.D., H.J., and K.H. designed and supervised the project. X.X. and X.T. performed most of the experiments. X.X. and J.D. analyzed the data. X.X. wrote the manuscript and generated the figures. X.T., H.J., and K.H. helped draft. J.D. and H.J. revised the manuscript. All authors have read and approved the final manuscript.

Corresponding authors

Correspondence to Keqin Hua, Hua Jiang or Jingxin Ding.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Communications Biology thanks Xiaowei Xie and Yuxuan Zheng for their contribution to the peer review of this work. Primary Handling Editors: Kuangyu Yen and Kaliya Georgieva.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Reporting summary

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Xiang, X., Tao, X., Hua, K. et al. Single-cell RNA sequencing reveals tumor heterogeneity in small cell neuroendocrine cervical carcinoma. Commun Biol 8, 184 (2025). https://doi.org/10.1038/s42003-025-07605-y

Download citation

Received: 17 June 2024
Accepted: 26 January 2025
Published: 05 February 2025
Version of record: 05 February 2025
DOI: https://doi.org/10.1038/s42003-025-07605-y

This article is cited by

Development and external validation of a machine learning-based prognostic model for small cell neuroendocrine cervical carcinoma: a multi-center study
- Yaxin Kang
- Lele Chang
- Qin Xu
BMC Cancer (2025)
Advancements in single-cell sequencing for cervical cancer research
- Congli Pu
- Biyuan Xing
- Yingchao Zhao
Molecular and Cellular Biochemistry (2025)

Subjects

Abstract

Similar content being viewed by others

Introduction

Results

The single-cell atlas reveals cellular diversity in SCNECC

Heterogeneity of SCNECC epithelial cells

Molecular subtypes of SCNECC based on epithelial heterogeneity

The tumorigenesis model of SCNECC

The immune landscape of SCNECC

Intercellular communication in the TME of SCNECC

Subtyping cell lines and possible clinical value of molecular subtyping

Discussion

Methods

Collection and preparation of single-cell suspensions from SCNECC samples

Construction and quality control of the 10× Genomics scRNA-seq library

scRNA-seq data analysis

Cell culture and western blotting

Immunohistochemistry of formalin-fixed, paraffin-embedded (FFPE) sections

Analysis of immunohistochemistry data

Clinicopathological characteristics and prognosis analysis

Statistics and reproducibility

Reporting summary

Data availability

Code availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding authors

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Supplementary information

Rights and permissions

About this article

Cite this article

Share this article

This article is cited by

Search

Quick links