Single cell sequencing deciphering the heterogeneous landscape of blastic plasmacytoid dendritic cell neoplasm with novel MYB-ZFAT fusion gene

Abstract

Background

Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare, aggressive blood cancer from plasmacytoid dendritic cell precursors. It’s marked by CD4, CD56, CD123, and CD303/CD304 expression and involves molecular disruptions like chromatin deletions, mutations, and chromosomal translocations.

Methods

The current study employed a comprehensive method with clinical samples, histology, FACS immunophenotyping, karyotype analysis, transcriptome and protein structure analysis, and single-cell sequencing to explore BPDCN’s molecular basis.

Results

The study discovered a new MYB-ZFAT gene fusion in a BPDCN patient and showed a diverse cell population, contradicting a single cell type theory. It found four major clusters (Cluster 1,2,3,8 ) and one cluster (clulster 12) with unique profiles and roles in disease progression. The research noted Key pathways include T cell receptor signaling, NK cell cytotoxicity, and hematopoiesis are involved in pathogenesis. The study emphasized MYB activation’s role in BPDCN’s cellular clustering and identity.

Conclusion

The study indicates BPDCN’s complexity with varied cellular origins and a significant role for MYB activation in its development. This research deepens our comprehension of BPDCN’s pathogenesis and cell populations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12935-025-03899-4.

Introduction

Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a very rare and aggressive blood cancer originating from plasmacytoid dendritic cell precursors, characterized by abnormal CD4, CD56, and CD123,CD303/CD304 expression [1]. While BPDCN’s exact etiology is not fully understood, it’s likely linked to various molecular disruptions such as chromatin deletions, genetic mutations, and chromosomal translocations, all of which are considered part of its pathogenetic mechanisms [2]. Renosi et al. reviewed the genetic and epigenetic features of neoplasms involving plasmacytoid dendritic cells (pDC). BPDCN is characterized by complex karyotypes, frequent deletions involving immune and tumor suppressor genes, and MYC/MYB rearrangements, while pDC-AML is marked by RUNX1 mutations [3].

Suzuki’s study found that all five pediatric and four adult BPDCN patients had MYB gene rearrangements. This mutation is common and enhances MYB’s transcriptional activity, possibly by shortening its regulatory domain, which may contribute to the disease’s development [4]. Booth et al. investigated the role of MYB fusions in BPDCN. They showed that MYB fusions, found in about 20% of BPDCN cases, aberrantly regulated G2/M cell cycle genes, impaired dendritic cell differentiation, and induced leukemia in vivo. Using CUT&RUN chromatin profiling and in vivo models, they demonstrated that MYB fusions increased binding at cell cycle gene promoters, leading to higher expression of these genes [5]. Erica A. K. DePasquale analyzed 52,803 single-cell transcriptomes, including 18,779 T-cells, to understand BPDCN’s impact on the tumor environment. The study found increased interferon alpha (IFNA) response and decreased tumor necrosis factor-alpha (TNFA) signaling in BPDCN patients. It also revealed significant T-cell exhaustion linked to T-cell clonotype expansion, showing a complex tumor-immune interaction in BPDCN [6].

In this study, we identified a new MYB-ZFAT gene fusion in a BPDCN patient. We analyzed the chimeric protein’s structure with the SMART database, created 3D models with I-TASSER and PyMOL, and used single-cell sequencing to explore the fusion gene’s role in BPDCN’s pathophysiology. This approach offers a detailed view of the molecular mechanisms behind this rare cancer.

Case report

In April 2021, a 12-year-old girl was admitted to Hunan Children’s Hospital with a month-long illness. She had firm, non-attached masses all over her body and non-itchy, non-painful red, peeling rashes on her head, neck, face, chest, back, and arms. A mass near her right wrist grew to egg-sized. The rash turned into purpuric petechiae, showing tiny skin hemorrhages. She also had subcutaneous nodules and occasional knee joint swelling and discomfort (Fig. 1A -E).

Fig. 1.

(A) At the initial phase of the disease, the patient developed rashes on the wrists and back. The rash turned into purpuric petechiae, showing tiny skin hemorrhages on the wrists and Chest before chemotherapy. (B) BPDCN cells stained with Giemsa may appear as large blasts with irregular nuclear contours and scant cytoplasm. (C) The typical morphological features observed in the patient’s bone marrow include: large cell size, high nuclear-to-cytoplasmic ratio, finely granular chromatin, prominent nucleoli, abundant cytoplasm, and the presence of scant azurophilic granules. (D) After treatment with VDLD (Vincristine, Dexamethasone, Daunorubicin, Pemetrexed )for 4 days, the patient’s skin lesions on the wrists and back have shown improvement. (E) After a 20-day treatment with VDLD, there has been significant improvement in the patient’s skin lesions on the wrists and back, with the rashes on the back being almost unnoticeable. (F) After treatment with CAM(Cyclophosphamide, Cytarabine, Mercaptopurine), the patient’s skin lesions on the wrist have been alleviated

Based on the complete blood count (CBC) report, the patient exhibited a white blood cell (WBC) count of 3.24 × 10^9 per liter, a hemoglobin (HGB) concentration of 102.00 g per liter, and a platelet (PLT) count of 164.00 × 10^9 per liter. Furthermore, the absolute neutrophil count (ANC), denoted as NE#, was recorded at 1.39 × 10^9 per liter.

HE staining showed diffuse infiltration of tumor cells in the dermis and subcutaneous tissue(Figure 2A). Immunohistochemical analysis yielded the following results: Ki-67 with a strong positive index of 50%, indicative of high proliferative activity, and terminal deoxynucleotidyl transferase (TdT) with a score of 3+. The panel of markers demonstrated negativity for CD3, CD20, PAX-5, CD30, CD35, CD21, CD23, and CD38, while positivity was observed for CD4 and CD43. Additionally, myeloperoxidase (MPO) showed a negative reaction, whereas CD99 and CD68 exhibited positive staining. CD117 and CD34 were both negative (Fig. 2B -G).

Fig. 2.

(A) HE staining of affected skin biopsy sample. Red arrow showed diffuse infiltration of tumor cells in the dermis and subcutaneous tissue. (B) IHC analysis of affected skin biopsy sample. Brown precipitate stained with DAB indicated strongly Positive of Bcl. (C) IHC analysis of affected skin biopsy sample showed weakly Positive of CD4. (D) IHC analysis of affected skin biopsy sample showed strongly Positive of CD43. (E) IHC analysis of affected skin biopsy sample showed para-nuclear punctate of CD68. (F) IHC analysis of affected skin biopsy sample showed Moderately Positive of Ki-67. (G) IHC analysis of affected skin biopsy sample showed Moderately Positive of TdT3

Flow cytometric analysis (FACS) of the bone marrow aspirate uncovered the presence of aberrant cell populations characterized by diminished CD45 expression relative to lymphocytes and an elevated side scatter (SSC) that was modestly greater than that of lymphocytes. These atypical cells constituted approximately 65% of the nucleated cellularity and were predominantly positive for a spectrum of antigens, including human leukocyte antigen-DR (HLA-DR), CD4, CD7, CD56, CD123, and CD303. The expression of this panel of markers is consistent with the immunophenotypic profile associated with blastic plasmacytoid dendritic cell neoplasm (BPDCN). (Figure 3A - G, Figure S1). The specimen from this patient was analyzed for 20 metaphase cells, and 9 cells showed a complex karyotype with chromosomal abnormalities. Karyotype analysis results showed 43–45,XX,-C, del(6)(q21), del(7)(q32),+mar, inc[cp9]/46,XX [11] (Figure. 3 H). These findings suggest BPDCN.

Fig. 3.

(A) The FSC and SSC profiles show a typical distribution pattern of cells. (B) The flow cytometry analysis showed that cells were divided into 4 populations using CD45 and SSC. Cells with high CD45 expression are typically indicative of lymphocytes (green).CD45- cells are typically non-hematopoietic cells (purple).Low CD45 expression may indicate monocytes or other non-lymphoid cells (Red & Blue). (C) The flow cytometry analysis demonstrated the cell population with HLA-DR positive and CD15 negative. (D) The flow cytometry analysis demonstrated the cell population with CD123 positive and CD15 negative. (E) The flow cytometry analysis demonstrated the cell population with CD4 positive and CD3 negative. (F) The flow cytometry analysis demonstrated the cell population with CD16 negetive and CD56 positive. (G) The flow cytometry analysis demonstrated the cell population with CD303 positive and CD304 negative. (H) The karyotype analysis showed abnormal chromosomal arrangement: 43–45,XX,-C, del(6)(q21),del(7)(q32),+mar, inc[cp9]/46,XX[11]

Methods

Clinical sample collection

The patient visited the Hunan Children’s Hospital and received a diagnosis of BPDCN based on the 2008 WHO Classification of Myeloid Neoplasms and Acute Leukemia [7]. To conduct further research, bone marrow, peripheral blood, and skin biopsies were collected from the patient with informed consent, which was obtained in accordance with the Declaration of Helsinki. The Human Research Ethics Committee of the Second Xiangya Hospital approved the studies.

HE staining

Samples of tissue were taken from the affected skin lesions. These samples were then placed in a 10% formalin solution and sent to the Pathological Laboratory of the Second Xiangya Hospital. Once labeled, the samples were embedded in paraffin, sliced into sections that were 5–7 mm thick, and finally stained with HE (artificial hematoxylin and eosin).

Immunohistochemistry

Samples of tissue were obtained from the affected skin lesions and sliced into sections that were 4 μm thick. The sections were dewaxed and washed using heated citrate buffer. Afterward, they were incubated with specific antibodies (anti-CD4, anti-Bcl, anti-CD43, anti-CD68, anti-Ki67, and anti-TdT3). After Washing with PBS for 3 times, they were incubated HRP-conjugated secondary antibody. DAB solution and hematoxylin were added to the slices, and the staining results were evaluated by the pathologists of the Second Xiangya Hospital.

FACS

For immunophenotyping, bone marrow cells collected from the patient were incubated with IgG1 FITC, IgG1 PE, IgG1 PerCP, IgG1 APC, IgG1 PE/Cy7, IgG1 V450, IgG1 APC Cy7, CD45 V500, cIgG1 FITC, cIgG1 PE, cIgG1 PerCP, cIgG1 APC, cIgG1 PE Cy7, cIgG1 V450, cIgG1 APC Cy7, CD45 V500, CD33 FITC, CD117 PE, CD34 PerCP, CD56 APC, CD13 PE Cy7, CD11b Pacific blue, HLA-DR APC Cy7, CD45 V500, CD14 FITC, CD304 PE, CD34 PerCP, CD123 APC, CD4 PE/Cy7, HLA-DR APC Cy7, CD45 V500, CD56 Pacific blue, CD61 FITC, CD36 PE, CD45 PerCP, CD34 APC, D16 FITC, CD13 PE, CD45 PerCP, CD24 APC, CD15 FITC, CD371 PE, CD34 PerCP, CD123 PE Cy7, CD33 APC, CD45 V500, CD56 Pacific blue, HLA-DR APC Cy7, CD4 FITC, CD13 PE, CD123 PE Cy7, CD56 APC, CD45 V500, CD7 Pacific blue, and HLA-DR APC Cy7 antibodies (all antibodies were purchased from Becton, Dickinson and Company (BD)), New York, USA). Fluorescence intensity and positive cell percentages were determined with a BD FACS Canto (Becton, Dickinson and Company (BD)), and data were analyzed using BD FACS Diva (Becton, Dickinson and Company (BD)). (Table 1)

Table 1.

FACS antibodies

BD Biosciences catalog	Cell Surface Markers
664,934	CD45 PerCP
662,854	CD10 PE
337,899	CD22PE
335,828	CD20PE-CY7
663,494	CD38 PE-Cy7
	CD58APC
665,750	CD14 APC
555,527	CD64FITC
566,919	CD123PE
557,758	CD16APC-CY7
663,487	CD56 PE-Cy7
664,995	CD11b PE
665,003	CD15 FITC
665,330	HLA-DR APC
662,910	CD13 PE-CY7
	CD71APC-CY7
665,336	CD5 APC
555,327	CD2PE
663,493	CD4 PE-Cy7
349,201	CD3 FITC
663,521	CD8 APC-Cy7
664,936	CD117 PE
664,935	CD7 FITC
664,937	CD33 APC
652,804	CD19 APC
348,801	CD34PE-CY7
566,427	CD303BV421
354,508	CD304PE-CY7
665,342	CD79a PE
665,337	MPO FITC
663,526	CD3 PE
332,791	TDT APC

Chromosome karyotype analysis

Chromosome Prep: Short-term Culture involves counting bone marrow cells, culturing in 1640 medium at 37 °C for 24–48 h with colchicine, and CPG stimulant for 72 h. Post centrifugation, hypotonic treatment, and fixation, samples are stored at 4 °C. G-banding involves centrifuging for a cell suspension, air-drying on slides, baking at 75 °C, trypsin digestion, saline rinse, Giemsa stain, and final rinse and dry. Chromosome Analysis uses the Metafer workstation for automatic scanning and capturing of 40 metaphase images per slide for detecting abnormalities.

Transcriptome sequencing and MYB-ZFAT verifying

Total RNA was extracted from cell pellets with an RNeasy Mini Kit, treated with DNase I, and purified on Qiagen columns. RNA quality and quantity were measured by spectrophotometry. PolyA + RNA was isolated using an Oligotex kit. First-strand cDNA was synthesized from 1 µg PolyA + RNA with random primers and Superscript II. Second-strand synthesis was done at 16 °C, followed by purification. Illumina GAII sequencing was performed for short reads, with the RefSeq gene set and human genome build 36 as references. CLC Bio Genomic Work Bench was used to map reads with a strict cutoff for unique mapping and mismatch allowance. Expression levels were calculated in RPKM. The existence of MYB-ZFAT fusion gene was verified by PCR combined with Sanger sequencing.

Protein structure analysis

Using SMART (Simple Modular, Architecture Research Tool), the domains of the wild-type MYB and MYB-ZFAT fusion proteins were identified. The prabi database analyzed the MYB-ZFAT’s secondary structure (https://prabi.ibcp.fr/htm/site/web/aboutUs/teaching). I-TASSER Suite modeled their 3D structures (http://zhanglab.ccmb.med.umich.edu/I-TASSER/download/), and PyMOL 1.8.6 visualized the PDB Models (https://sourceforge.net/projects/pymol/files/pymol/1.8/).

Single cell sequencing

The document outlines a single-cell sequencing experiment with several key steps. Initially, human bone marrow mononuclear cells were extracted and stored in an EDTA anticoagulant tube for sample preparation. For library construction, the 10x Genomics platform was used, leveraging microfluidic technology to encapsulate cells and barcoded beads into droplets. Within these droplets, cells were lysed, allowing mRNA to bind to the barcodes, forming GEMs. Reverse transcription was then conducted to create cDNA libraries, which were indexed for sequencing. The cDNA libraries were sequenced to generate raw data. Quality control was performed using Cell Ranger software, which aligned reads to the reference genome with STAR. Metrics such as the number of high-quality cells, mean reads per cell, median genes per cell, sequencing saturation, and mapping rates were assessed. In the data analysis phase, low-quality cells and artifacts were filtered out. High-quality cells were subjected to dimensionality reduction and clustering using algorithms like MNN and t-SNE to identify distinct cell populations. Marker genes were identified, and cell types were annotated based on reference datasets. Differential gene expression analysis was conducted to find significantly upregulated or downregulated genes between different cell types or conditions. Additionally, functional enrichment analysis (GO and KEGG) and protein-protein interaction network analysis were performed to explore the biological functions and pathways associated with the identified genes.

Results

Fusion protein

The existence of MYB-ZFAT fusion gene was verified by PCR combined with Sanger sequencing (Fig. S2, Figure S2 & S3). We explore the fusion protein’s complexities, clarifying its biological function and structural analysis. This detailed molecular characterization is key to grasping the protein’s impact on the disease’s pathophysiology. The MYB-ZFAT fusion protein’s structure was closely studied and compared to the normal MYB protein. This fusion arises from the chromosomal translocation t(6;8)(q23;q24), the cause of its abnormal activation (Fig. 4A -C). A premature stop codon at the 347th amino acid in the MYB sequence results in a shortened MYB-ZFAT fusion protein with three DNA-binding domains (R1: AA 40–86, R2: AA 92–138, R3: AA 144–189) and a transcriptional activation domain (AA 267–313). The fusion omits the essential negative regulatory domain (NR: AA 518–677), which may lead to uncontrolled MYB activation and drive neoplasia.(Figure 4D and E). Rearrangements of the MYB oncogene have been identified in pediatric patients with recurrent chronic myelomonocytic leukemia (CMML), in addition to those afflicted with blastic plasmacytoid dendritic cell neoplasm (BPDCN) [8]. These genetic alterations are considered significant as they may play a pivotal role in disease pathogenesis and could have implications for therapeutic strategies [9].

Fig. 4.

A. The fusion gene is formed by the juxtaposition of two distinct genes. MYB in chromosome 6 and ZFAT in chromosome 8. The breakpoint in MYB is chr6:135515598. The breakpoint in ZFAT is chr8:135529901. B. This image showed how MYB-ZFAT fusion gene organized in chromsome. C. This image showed retained protein domains in MYB-ZFAT fusion protein: Myb-like-DNA-binding domain (green) and LMSTEN motif (purple). D. The diagram identifieds the structure of MYB WT composed with DNA-binding domain consists of three imperfect repeats labeled (R1, R2, R3); the transcriptional activation domain (TA) and the C-terminal negative regulatory domain (NR). E. MYB-ZFAT only composed withDNA-binding domain consists of three imperfect repeats labeled (R1, R2, R3) and the transcriptional activation domain (TA). F. The 3D model of the WT MYB protein generated with I-TASSER. G. The 3D model of the MYB-ZFAT fusion generated with I-TASSER

The 3D models of the WT MYB protein and the MYB-ZFAT fusion were generated with I-TASSER. The 730-AA WT MYB has three ligand-binding sites: Zinc ions at AA 119, 122, 135, 139, 147, 150, 163, 167, 175, 178, 191, 195; chloride ions at AA 203, 206, 208; and glycerol at AA 237, 240, 244, 245, 261, 262, 263. The 347-AA MYB-ZFAT fusion protein has five predicted binding sites: cholesterol at AA 37 and 64; extra Zinc ions at AA 40 and 58; n-decyl-beta-D-thiomaltoside at AA 11, 17, 18, 21, 76, 77; phosphate groups at AA 15, 18, 24; and calcium ions at AA 34, 45, 48. These structural details are key to understanding the MYB-ZFAT fusion’s molecular effects. (Fig. 4G).

Single cell sequencing

Single-cell sequencing reveals greater uniformity in BPDCN’s neoplastic cells compared to the diverse normal bone marrow cells. It shows distinct BPDCN cell populations with little overlap with the varied cells in healthy bone marrow (Fig. 5A). The cell populations in BPDCN patients are predominantly composed of four major clusters, namely Cluster 1, Cluster 2, Cluster 3, and Cluster 8, which together form the most significant proportion of the cellular landscape (Fig. 5B). The heatmap analysis shows strong links between Clusters 11, 5, and 7, and also among Clusters 8, 12, 3, 1, and 2, indicating shared cellular traits. Clusters 10, 9, 4, and 6 also correlate closely. Single-cell sequencing offers a detailed view of BPDCN’s cellular diversity and interactions (Figure 5C).

Fig. 5.

(A) The dimensionality reduction results based on MNN (Mutual Nearest Neighbors) are visualized for single-cell clustering through t-SNE, with the clustering algorithm using SNN, ultimately achieving the optimal cell grouping.The horizontal and vertical axes represent the first and second principal components of the dimensionality reduction, respectively. Each point in the graph represents a cell, with cells from different groups distinguished by different colors. (B) Bar chart of the proportion of cell clusters among samples. The horizontal axis represents different samples, and the vertical axis represents the percentage of cell counts in different groups. (C) Heatmap of the correlation between cell clusters. The horizontal and vertical axes represent different cell groups, and the numbers in the graph are Pearson correlation coefficients. The higher the value, the redder the color of the heatmap, indicating a higher degree of correlation between the cell groups

Given BPDCN’s characteristic markers CD4, CD56, CD123, and CD303/CD304, our patient’s bone marrow showed positivity for CD4, CD123, CD56, and CD303. We identified a new MYB-ZFAT fusion gene linked to MYB oncogene activation. We then analyzed expression of CD4, CD123, CD303, CD56, MYB, and lineage markers in cellular clusters. This thorough evaluation helped define the disease’s immunophenotype and molecular basis in the patient.

Cluster 1

In controls, Cluster 1 matched normal pDCs with CD4, CD123, CD303 expression but lacked CD56 and MYB. In BPDCN, Cluster 1 showed an altered profile with CD4, CD123, CD56, MYB expression and lost CD303, a late pDC marker. Several genes including LCNL1, MPPED2, SMIM5, VIPR2, PCDH9, PLXNA4, KIF17, and SLITRK5 (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Fig. 6.

(A) Violin plot of CD4 gene expression levels.The horizontal axis represents cell group numbers, and the vertical axis represents the normalized gene expression values. Green represents the BPDCN group, and purple represents the CTRL group. (B) Violin plot of CD123 gene expression levels.The horizontal axis represents cell group numbers, and the vertical axis represents the normalized gene expression values. Green represents the BPDCN group, and purple represents the CTRL group. (C) Violin plot of CD303 gene expression levels.The horizontal axis represents cell group numbers, and the vertical axis represents the normalized gene expression values. Green represents the BPDCN group, and purple represents the CTRL group. (D) Violin plot of CD56 gene expression levels.The horizontal axis represents cell group numbers, and the vertical axis represents the normalized gene expression values. Green represents the BPDCN group, and purple represents the CTRL group. (E) Violin plot of MYB gene expression levels.The horizontal axis represents cell group numbers, and the vertical axis represents the normalized gene expression values. Green represents the BPDCN group, and purple represents the CTRL group

Table 2.

Lineage markers, BPDCN markers and MYB expression of each Cluster

CLuster	Early stage markers	Myeloid markers	B cell lineage markers	T cell lineage markers	Erythroblast markers	Megakayocyte lineage markers	BPDCN markers	MYB
Cluster1 (Ctrl)	CD34- HLA-DR+	CD36+	Kappa+ Lamda-	CD4+ CD7+ CD8-	CD71+ CD36+	CD36+	CD4+ CD123+ CD303+ CD56-	Negative
Cluster1 (BPDCN)	CD34- HLA-DR+	Negative	Kappa+ Lamda+	CD4+ CD7+ CD8-	CD71+ CD36-	Negative	CD4+ CD123+ CD303- CD56+	Positive
Cluster2 (Ctrl)	CD34- HLA-DR+	Negative	Negative	Negative	Negative	Negative	Negative	Negative
Cluster2 (BPDCN)	CD34- HLA-DR+	Negative	Kappa+ Lamda+	CD4+ CD7+ CD8-	CD71+ CD36-	Negative	CD4+ CD123+	Positive
Cluster3 (Ctrl)	CD34- HLA-DR+	CD36+	Kappa+ Lamda-	CD4+ CD7- CD8-	CD71+ CD36+	CD36+	CD4+ CD123+ CD303+ CD56-	Negative
Cluster3 (BPDCN)	CD34- HLA-DR+	Negative	Kappa+ Lamda+	CD4+ CD7- CD8-	CD71+ CD36-	Negative	CD4+ CD123+ CD303+ CD56+	Positive
Cluster 4 (Ctrl)	CD34- HLA-DR-	Negative	Kappa- Lamda-	CD8+ CD2+ CD3E+ CD3D+ CD3G+ CD7+ CD1A- TCR+	Negative	Negative	Negative	Negative
Cluster 4 (BPDCN)	CD34- HLA-DR-	Negative	Kappa- Lamda-	CD8+ CD2+ CD3E+ CD3D+ CD3G+ CD7+ CD1A- TCR-	Negative	Negative	Negative	Negative
CLuster 5 (Ctrl)	CD34- HLA-DR+	CD13+ CD33+ CD11b+ CD14+ CD64+ CD36+ CD11c+	Kappa- Lamda-	CD4+	CD36+	CD36+	CD4+ CD123- CD303- CD56-	Negative
Cluster 5 (BPDCN)	CD34- HLA-DR+	CD13+ CD33+ CD11b+ CD14+ CD64+ CD36+ CD11c+	Kappa+ Lamda+	CD4+	CD36+	CD36+	CD4+ CD123- CD303- CD56-	Negative
Cluster 6 (Ctrl)	CD34- HLA-DR+	Negative	Kappa- Lamda-	CD4+ CD2+ CD3E+ CD3D+ CD3G+ CD7+ TCR+	Negative	Negative	CD4+ CD123- CD303- CD56-	Negative
Cluster 6(BPDCN)	CD34- HLA-DR+	CD36+	Kappa+ Lamda+	CD4- CD2+ CD3E+ CD3D+ CD3G+ CD7- TCR-	CD71+ CD36+	CD36+	CD4- CD123- CD303- CD56-	Negative
Cluster 7 (Ctrl)	CD34- HLA-DR+	CD13+ CD33+ CD11b+ CD14+ CD64+ CD36+ CD11c+	Kappa+ Lamda-	CD4+	CD36+ CD71-	CD36+	CD4+ CD123- CD303- CD56-	Negative
Cluster 7 (BPDCN)	CD34- HLA-DR+	CD13- CD33+ CD11b+ CD14+ CD64- CD36- CD11c+	Kappa+ Lamda+	CD4+	CD36- CD71+	CD36-	CD4+ CD123- CD303- CD56-	Negative
Cluster 8 (Ctrl)	Negative	Negative	Negative	Negative	Negative	Negative	Negative	Negative
Cluster 8 (BPDCN)	CD34- HLA-DR+	Negative	Kappa+ Lamda+	CD4+ CD7+	CD71+	Negative	CD4+ CD123+ CD303+ CD56+	Positive
Cluster 9 (Ctrl)	CD34- HLA-DR+	CD11b+	Negative	CD2+ CD3E+ CD3D+ CD3G+ CD7+ TCR+	Negative	Negative	Negatvie	Negatvie
Cluster 9 (BPDCN)	CD34- HLA-DR+	Negative	Kappa+ Lamda+	CD2+ CD3E+ CD3D+ CD3G+ CD7+ TCR+	Negative	Negative	Negatvie	Negatvie
Cluster 10 (Ctrl)	CD34- HLA-DR+	Negative	CD79a+ CD19+ CD22+ CD20+ Kappa+ Lamda-	Negative	Negative	Negative	Negatvie	Negatvie
Cluster 10 (BPDCN)	CD34- HLA-DR+	Negative	CD79a+ CD19- CD22+ CD20+ Kappa+ Lamda+	CD3E+ CD7+	Negative	Negative	Negatvie	Negatvie
Cluster 11 (Ctrl)	CD34- HLA-DR+	CD13+ CD33+ CD11b+ CD14+ CD64+ CD36+ CD11c+	Kappa+ Lamda-	CD4+ CD8A+ CD8B+ CD2+ CD3E+ CD3D+ CD3G+ CD7+ TCR+	CD36+	CD36+ CD42b-	CD4+ CD123- CD303- CD56-	Negative
Cluster 11 (BPDCN)	CD34- HLA-DR+	CD13+ CD33- CD11b- CD14- CD64- CD36- CD11c-	Kappa- Lamda+	CD4+ CD8A- CD8B- CD2- CD3E- CD3D- CD3G- CD7- TCR-	Negative	CD36- CD42b+	CD4+ CD123- CD303- CD56-	Negative
Cluster 12 (Ctrl)	CD34- HLA-DR+	CD13+ MPO+ CD33+ CD15+ CD11b+ CD14+ CD64+ CD36+ CD11c+	Kappa+ Lamda-	CD4+ CD2+ CD3G+ TCR+	CD71+ CD36+	CD36+	CD4+ CD123+ CD303+ CD56-	Negative
Cluster 12 (BPDCN)	CD34- HLA-DR+	CD11b+	Kappa+ Lamda+	CD4+ CD7+	CD71+ CD36-	CD36-	CD4+ CD123+ CD303+ CD56+	Positive

Table 3.

Top genes and Functions of Each Cluster

Cluster	Definition	Top genes	Function	Reference
Cluster 1	Primitive pDC population in BPDCN, at an early stage of differentiation, and related to tumor migration, microenvironment.	LCNL1	Lung cancer	PMID: 38,180,474
			Bladder cancer	PMID: 34,059,019
		MPPED2	Gliblastoma	PMID: 33,734,003
			Neuroblastoma	PMID: 22,262,177
			Breast cancer	PMID: 31,181,813
			Colorectal cancer	PMID: 30,846,004
			Squamous cell Cacinoma	PMID: 27,698,780
		VIPR2	tumor cell migration	PMID: 36,237,322
			Glioblastoma	PMID: 35,456,975
			Lung cancer	PMID: 34,329,340
		PCDH9	Breast caner	PMID: 36,936,167
			Prostate cancer	PMID: 36,936,167
			Glioma	PMID: 35,084,653
		PLXNA4	Cytotoxic T cell trafficking in cancer	PMID: 34815265P
			Endometrial cancer	MID: 33,012,523
			Thyroid carcinoma	PMID: 32,065,787
			Osteocarsoma	PMID: 32,614,239
		KIF17	Breast cancer	PMID: 32,322,170
		SLITRK5	Gastric cancer	PMID: 37,772,256
		SMIM5	Renal cell carcinoma	PMID: 24318988P
			Oral squamous cell carcinoma	MID: 30,275,705
		PTCRA	T cell development	PMID: 16,754,716
Cluster 2	A group of cells originally did not have clear lineage marker expression, but after MYB activation, they acquired part of the T-, B-, and erythroid lineages as well as the plasma cell DC phenotype, and are a mixed lineage tumor-associated cell population.	CLLU1	Chronic lymphocytic leukemia	PMID: 17,284,524
				PMID: 16,529,606
				PMID: 22,899,580
				PMID: 22,738,889
				PMID: 16,339,396
				PMID: 19,212,335
		PAX2	Target of tumor suppressor gene WT1	PMID: 14,603,255
				PMID: 21,778,682
		CDCA7	C-myc target gene	PMID: 37,248,764
				PMID: 11598121P
				MID: 16,580,749
		BCAT1	C-myc target gene	PMID: 33760210P
				MID: 23,758,864
				PMID: 8,692,959
		SIX1	Regulates C-myc, CCND1	PMID: 16,488,997
				PMID: 30,379,332
		IGLC2	Trigger clonal expansion and differentiation of B lymphocytes into immunoglobulins-secreting plasma cells	PMID: 22,158,414 PMID: 20,176,268
		IGLC3	Membrane-bound or secreted glycoproteins produced by B lymphocytes	PMID: 27,994,173
		FCGR2B	Fc Gamma Receptor IIb, B cell receptor signalling pathway	PMID: 17,522,256
				PMID: 15,506,939
Cluster 3	Compared to the Ctrl group, the BPDCN group Cluster3 acquired Lamda, CD7, CD56, MYB, and lost CD36 This is a group of cells originally present in normal cells, with a mixed phenotype of partial myeloid lineage, B lineage, T lineage, erythroid lineage and plasma cell DC, and after MYB activation in BPDCN, it lost the myeloid phenotype, acquired Lamda, and CD56. A tumor-associated cell population.	CLEC4C (CD303)	BPDCN dendritic cell development	PMID: 11,748,283
			Kinase related genes	PMID: 21,880,719
		IL1R2	ERK signaling	PMID: 31,744,444
			Esophageal cancer	PMID: 33,277,616
			Gastric cancer	PMID: 31,921,558
			Breast cancer	PMID: 36,029,680
			Renal cell carcinoma
		TIFAB	P53, Malignant Hematopoiesis	PMID: 32,101,751
			HOXA9, AM	PMID: 34,877,491
			TRAF6,MDS	PMID: 26,458,771
		HTR3A	5 H tryptophan	PMID: 37,795,441
			prognosis of breast	PMID: 37,231,904
			Gastric cancer	PMID: 35,757,399
			Lung cancer	PMID: 33,042,464
		SLC18A2	Dopamine	PMID: 19,228,741
			Norepinephrine	PMID: 26,905,753
			Serotonin	PMID: 30,954,278
			Histamine	PMID: 31,240,161
			prostate cancer	PMID: 32,084,491
		TDRD1	prostate cancer	PMID: 36,497,036
			Non small cell lung cancer	PMID: 33,381,926
			Ovarian cancer	PMID: 36,865,141
				PMID: 25,788,493
				PMID: 37,609,819
				PMID: 33,374,923
Cluster 4	This population of cells normally mainly expressed markers of the T lineage, expressing Kappa across the lines. Kappa and TCR were lost in the BPDCN. It can be defined as a population of T cells that did not develop into a tumor cell population.	CCR7	Dendritic cell maturation,	PMID: 38,267,413
			Tumor related Dendritic cells	PMID: 24,691,220
				PMID: 34,780,080
		CD8B	Surface glycoprotein on cytotoxic T lymphocytes	PMID: 1,672,332
				PMID: 37,781,365
		NELL2	Oncogenesis	PMID: 34,365,093
				PMID: 10,548,494
		LRRN3	Extracellular matrix and extracellular space, Vascular Skin Disease, melanoma, phechromocytomas	PMID: 34,552,928
				PMID: 19,661,783
		APBA2	Breast carcinogenesis	PMID: 34,773,240
			AML	PMID: 16,849,538
		EPHX2	Prostate cancer	PMID: 32,687,069
			Colon cancer	PMID: 32,687,069
			Hepatocellular carcinoma	PMID: 34,278,494
		SCGB3A1	Lung cancer	PMID: 38,505,085
			breast cancer	PMID: 30,405,052
			Testicular cancer	PMID: 17,029,216
Cluster 5	This population of cells mainly expressed myeloid markers and CD4 across lines. In the BPDCN, the expression of Lamda was obtained. The remaining phenotypes remained unchanged. It can be considered that it is mainly a myeloid cell population that does not develop into a tumor cell population	CYP1B1	Colorectal cancer	PMID: 37,059,712
			Triple nagative Breast cancer	PMID: 36,077,068
			Lung cancer	PMID: 24,989,113
			AML	PMID: 29,793,312
			ALL	PMID: 23,757,320
		CES1	Prostate cancer	PMID: 34,185,414
			Hepatocellular carcinoma	PMID: 36,472,914
			Liver cancer	PMID: 33,206,527
		THBS1	Platelet aggregation	PMID: 33,538,796
			Angiogenesis	PMID: 36,196,150
			GliomaTumorigenesis	PMID: 33,664,245
			Tumor Microenvironment	PMID: 33,925,464
		MCEMP1	Mast cell differentiation	PMID: 37,041,174
			Immune responses
		QPCT	Breast cancer	PMID: 38,417,050
			AML	PMID: 37,381,175
			Renal cell carcinoma	PMID: 34,036,385
			Thyroid cancer	PMID: 37,060,824
		STEAP4	Prostate cancer	PMID: 36,011,027
			Hepatocellular carcinoma	PMID: 35,390,632
			Colorectal cancer	PMID: 36,187,390
		CRISPLD2	Neutrophil extracellular traps	PMID: 37,178,049
		CLEC4D	clearance of the antigen, processing and further presentation to T-cells	PMID:14,971,047
Cluster 6	This population of cells, under normal conditions, mainly expressed T lineage markers, expressed Kappa across the lines, and in the case of MYB activation, lost Kappa gained Lamda, acquired CD36, and lost CD4, CD7, TCR. It can be defined as a group of T cell populations that vary in phenotype due to the tumor.	TRAT1	Ovarian cancer	PMID: 34,868,239
			Non-small cell lung cancer
		MAF	Induce monocytic differentiation and apoptosis inbipotent myeloid progenitors	PMID: 10,477,683
			Interact with c-Myb to down regulate Bcl−2 expression and increase apoptosis in peripheral CD4 cells	PMID: 17,823,980
		CD40LG	T-B Cell-Activating Molecule	PMID: 8,617,933
				PMID: 19,426,221
		GATA 3	Regulator of T-cell development.	PMID: 23,287,858
		TNFRSF25	Stimulates NF-kappa B activity and regulate cell apoptosis.	PMID: 9,430,227
			Lymphocyte proliferation induced by T-cell activation.	PMID:22,956,587
		TNFRSF4	Activate NF-kappaB, CD4 + T cell response, T cell-dependent B cell proliferation and differentiation	PMID:22,956,587
				PMID: 10,754,303
		KLRB1	Natural killer (NK) cell cytotoxicity	PMID: 17,295,095
		CD28	enhances CD40L-mediated activation of NF-kappa-B and kinases MAPK8 and PAK2 in T-cells	PMID: 15,067,037
		NPDC1	Colon cancer	PMID: 37,822,345
			Gastric Cancer	PMID: 36,578,802
			AML	PMID: 31,235,852
		SLAMF1	T cell-B cell cross-talk	PMID: 30,058,241
			Stimulation Toll-like receptor	PMID: 29,440,514
			Dendritic Cells Developmental	PMID: 16,317,102
Cluster 7	Under normal conditions, this population of cells mainly expressed the markers of the myeloid lineage, expressing both CD4 and Kappa across the lines. In BPDCN, this population lost a fraction of myeloid Markers: CD13, CD64, CD36; Lamda, and erythroid marker CD71. It can be considered a population of myeloid cell populations whose phenotypes change due to the tumor.	CXCL10	Monocytes, natural killer and T-cell migration	PMID: 12,750,173
				PMID: 12,663,757
		TCF7L2	Wnt signaling pathway/ MYC	PMID: 32,248,976
				PMID: 32,068,780
		HES4	promotes early T-cell development	PMID: 31,919,081
			Bladder cancer	PMID: 38,370,367
			Colorectal cancer	PMID: 36,672,336
		CLEC10A	Immunological role and prognostic potential of CLEC10A in pan-cancer	PMID: 35,702,069
		CTSL	lysosomal cysteine protease tumor progression	PMID: 26,299,995
		BATF3	Development of classical dendritic cells (cDCs) and CD103(+) dendritic cells	PMID: 37,392,735
				PMID: 19,008,445
				PMID: 28,486,109
		TPPP3	Glioblastoma	PMID: 37,863,960
			Oral squamous cell carcinoma	PMID: 36,058,091
			Clear cell sarcomas	PMID: 31,488,818
			Nasopharyngeal	PMID: 34,668,461
		PTGIR	Ovarin cancer microenviroment	PMID: 36,551,640
			Oropharyngeal cancer	PMID: 33,003,642
Cluster 8	This group of cells, under normal conditions and did not express any lineage-related markers, obtained HLA-DR in BPDCN, some T lines B, some erythroid lines, and typical BPDCN markers, which can be regarded as the emergence of this tumor-associated cell population due to MYB activation.	KLF4	G-S cell cycle, P53, maintaining ES, Dendritic cells development	PMID: 34,680,943
				PMID: 30,578,299
				PMID: 16,904,174
				PMID: 36,827,419
		SPC25	Lung cancer	PMID: 34,746,885
			Breast cancer	PMID: 35,965,791
			Hepatocellualr carcinoma	PMID: 32,742,802
			Prostate cancer	PMID: 29,552,205
				PMID: 29,552,205
		MCM10	Preventing DNA damage	PMID: 34,645,815
			NK cell deficiency	PMID: 32,865,517
		PBK	MAPK family, DNA damage, P27 expression, phosphotylates P38, TP53, activation of lymphoid cells	PMID: 17,160,018
				PMID: 17,160,018
				PMID: 25,466,965
		E2F8	Lung cancer	PMID: 37,863,324
			Colon cancer	PMID: 26,089,541
			Prostate cancer	PMID: 27,683,099
			Ovarian	PMID: 32,823,614
			G1/s phase transition	PMID: 26,992,224
			Wnt/beta-catenin	PMID: 37,635,486
		PKMYT1	ALKBH5, gastic cancer	PMID: 35,114,989
			Promising therapeutic strategy for CCNE1-amplified cancers	PMID: 35,444,283
		TMSB15A	Breast cancer	PMID: 23,079,573
			Neck squamous cell carcinoma	PMID: 37,161,060
			Gastic cancer	PMID: 38,577,455
		DIAPH3	Diagnostic cancer biomarker	PMID: 36,750,200
			Lung adenocarcinoma	PMID: 31,586,548
			Cervical cancer	PMID: 34,659,408
			Beta-catenin/TCF, hepatocellular carcinoma	PMID: 28,795,316
Cluster 9	This is a group of cells mainly expressing T lines, expressing CD11b, acquiring B lines Kappa and Lamda, in BPDCN, losing myeloid marker CD11b, and can be considered a population of NK cells which phenotype changes due to the tumor.	GZMH	Apoptosis	PMID: 17,765,974
			Natural Killer cell	PMID: 17,409,270
		KLRD1	NK cells and some cytotoxic T-cells	PMID: 12,480,700
			Oral squamous cell carcinoma	PMID: 37,491,735
			Esophageal cancer	PMID: 36,588,538
		FGFBP2	Cytotoxic lymphocytes	PMID: 30,993,913
		KLRF1	NK cells and stimulates their cytoxicity and cytokine release	PMID:18,922,855
		PTGDR	Allergic inflammation	PMID: 30,986,261
		XCL2	Chemotactic activity for lymphocytes	PMID: 30,986,261
		FCRL6	Mature cytotxcic lymphocytes, B cell chronic lymphocytic leukemia	PMID: 18,991,291
		S1PR5	NK cells	PMID: 19,808,259
			Homestasis of patrolling monocytes	PMID: 23,519,784
		EOMES	Necessary for the differentiation of effector CD8 + T cells	PMID: 36,580,919
Cluster 10	This population of cells expressed only the markers of the B line in normal control, and in BPDCN, lost CD19, while obtaining the markers of two T lines, CD3E and CD7, can be considered as a population of cells that cause phenotypic changes due to tumor.	VPREB3	Representative of all stages of B-cell differentiation	PMID: 22,210,545
			MYC translocated lymphomas	PMID: 20,823,132
			B-cell lymphomas	PMID: 24,493,312
		CD22	The localization of B-cells in lymphoid tissues	PMID: 30,323,814
		PAX5	Early stages of B-cell differentiation	PMID: 22,127,921
			Acute lymphoblastic leukemia	PMID: 9,442,394
				PMID: 30,643,249
		FCRL1	B-cells activation and differentiation.	PMID: 37,822,931
			B-cell lymphoma	PMID: 36,409,389
			Immunotherapy target of CLL	PMID: 17,895,404
			Hairy cell lymphoma
			B-cell non hodgkin lymphoma
		FCRL2	Regulatory role in normal and neoplastic B cell development.	PMID: 19,682,311
				PMID: 21,068,405
			Prognostic marker for chronic lymphocytic leukemia	PMID: 18,314,442
				PMID: 31,092,813
		POU2AF1	Nodular lymphocyte predominant hodgkin lymphoma	PMID: 7,859,290
			Essential for response of B-cells to antigens	PMID: 7,779,176
		CD19	ell surface protein is restricted to B cell lymphocytes	PMID:2,463,100
				PMID:1,373,518
		COL19A1	Maintain the integrity of the extracellular matrix	PMID: 35,346,795
		BACE2	Aspartic protease	PMID: 10,683,441
Cluster 11	It can be argued that in tumors, this group of cells has lost almost all lineage marker and retained only a few myeloid and T lineage marker, but transformed to megakaryocytes, and the significance of whether this transition is due to tumorigenesis is unclear.	CYP1B1	NADPH-dependent electron transport pathway	PMID: 22,210,049
		PTAFR	Melanoma metastasis	PMID: 32,168,105
			Breast cancer	PMID: 30,008,927
			Eosinophilic leukemia	PMID: 30,866,528
		CLEC4E	Activation of plasmacytoid Dendritic cells	PMID: 27,183,629
		TREM1	Activatin of myeloid cells promtes antitumor immunity	PMID: 37,647,386
			Augment antitumor T cell immunity by inhibiting meyloid derived suppressor cells and restraning anti-PD-1 resistance	PMID: 37,651,197
		ALDH1A1	AML	PMID: 37,298,333
			Hematopoietic progenitors	PMID: 35,028,852
			Maintains leukemia stem cell properties	PMID: 37,761,947
				PMID: 18,082,256
				PMID: 31,768,017
		THBS1	A marker for acute myeloid leukemia and plays a role in tumor invasion	PMID: 32,117,788
			THBS1 producing tumor infiltrating monocyte-like cells contribute to immunosuppression and metastasis	PMID: 37,749,092
		CRISPLD2	Gstric cancer Tumor microenviroment	PMID: 37,600,223
			Bladder cancer	PMID: 37,178,049
			Attenuate pro-inflammatory cytokines	PMID: 34,150,000
Cluster 12	This is a group of mixed across lineage markers of cells, in BPDPN, the multilineage phenotype have changed, and obtained the CD7, CD56 tumor related markers, can be regarded as a group of relatively naive cell group, its originally has a variety of lineage marker expression, in tumor, the phenotype changes, but still is mixed with other tumor related cell group.	CRISP3	Prostate cancer	PMID: 32,357,309
			Mammay carcinoma	PMID: 30,609,035
		MMP8	Tumor microenviroment	PMID: 36,941,602
			Invasive of Breast cancer	PMID: 38,017,545
			Inflammatory disorders and Cancer progression	PMID: 21,388,856
		CYP4F3	Mediator of inflammation	PMID: 8,486,631
			Promyelocytic leukemic cell	PMID: 18,577,768
			AML	PMID: 19,029,204
				PMID: 14,715,252
		ANXA3	Triple-negative breast cancer	PMID: 37,139,408
			Colon cancer	PMID: 35,117,344
			Cervical cancer cells	PMID: 37,843,781
		CEACAM8	High expression in acute lymphoblastic leukemia	PMID: 17,909,799
		CD177	Beta-catenin signaling	PMID: 32,042,113
			Modulate the Tumor regulatory T cells	PMID:34,599,187
		TFF3	Lung cancer	PMID: 36,909,373
			Breast cancer	PMID: 28,687,783
			Colorectal cancer progression	PMID: 34,262,017
		TCN1	Colorectal cancer	PMID: 35,686,504
			Gastic cancer	PMID: 34,549,663
			Colon cancer	PMID: 32,686,693
		ARG1	Target for cancer treatment	PMID: 35,316,752
			Tumor-associated macrophages	PMID: 30,613,266
			Immune suppression in pancreatic cancer	PMID: 36,727,849
		CLC	Essential for Treg	PMID: 35,274,950
			Associated with inflammation	PMID: 12,031,912
			Myeloid leukemia	PMID: 12,399,964
			Myeloid-specific genes in childhood ALL

Cluster 2

In controls, the cluster lacked BPDCN markers. In BPDCN, it gained B cell (Kappa, Lambda), T cell (CD4, CD7), and erythroid (CD71) markers, plus CD123 and MYB, but not CD303 or CD56. The gene profile was rich in oncogenesis and c-myc pathway genes (Figure 6 & Figure S5-S9) (Tables 2 and 3).

Cluster 3

In BPDCN, Cluster 3 gained B cell (Lambda), NK cell (CD56), T cell (CD7), and MYB markers, losing CD36, a myeloid marker. It showed a mixed phenotype of myeloid, B, T, erythroid lineages, and pDCs. MYB activation led to a shift from myeloid to a tumor-associated cell population with B and NK cell markers. The top 10 genes expressed are linked to dendritic cell development (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 4

Compared to controls, BPDCN’s Cluster 4 lacked Kappa light chains and TCR complex, suggesting a T cell subset not transformed into tumor cells. Normally, this cluster shows T lineage markers with Kappa co-expression. The top 10 gene expressions relate to T cell oncogenesis, tumor immunology, and dendritic cell functions (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 5

Cluster 5 in BPDCN, unlike the control group, expressed Lambda immunoglobulin light chains alongside myeloid and CD4 markers (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 6

Cluster 6 in BPDCN showed a shift with Lambda gain, Kappa loss, CD36 upregulation, and CD4, CD7, TCR downregulation, plus MYB oncogene acquisition. Normally T lineage marker-expressing with Kappa, these cells change post-MYB activation, losing Kappa, gaining Lambda and CD36, and T cell markers. (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 7

Cluster 7 in BPDCN lost myeloid markers CD13, CD64, CD36, and gained erythroid marker CD71 and Lambda light chain, unlike controls with myeloid markers and CD4, Kappa. This indicates a tumor-influenced phenotype shift. Classified as a myeloid subset. Its top 10 genes are linked to immune regulation, cancer progression, and T and dendritic cell development. (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 8

Cluster 8 in BPDCN showed a shift with new expressions of HLA-DR, Kappa, Lambda, CD4, CD7, CD71, and BPDCN-specific markers CD56, CD123, CD303, not seen in controls. The top genes in this cluster relate to dendritic cell development, chromosomal stability, DNA damage response, TP53, and cell cycle regulation (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 9

Cluster 9 in BPDCN showed a loss of myeloid marker CD11b and gained B cell markers Kappa and Lambda, differing from controls with T lineage markers and CD11b. The top 10 gene expressions linked to NK cell function regulation (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 10

Cluster 10 in BPDCN, unlike controls with only B lineage markers, lost CD19 and gained Lambda, CD3E, and CD7, indicating a T cell-like shift due to the tumor. Gene expression showed top genes crucial for B cell function, development, and lymphocytic leukemia pathogenesis (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 11

Cluster 11 in BPDCN lost myeloid (CD33, CD11b, etc.), T lineage (CD8A, CD3E, etc.), and B (Lambda) markers, and gained CD42b, unlike controls with myeloid, B, and T markers. Top gene expressions link to leukemia pathogenesis, pDC activation, and tumor microenvironment dynamics, (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Cluster 12

Cluster 12 in BPDCN, unlike controls, lost myeloid markers but gained Lambda, CD56, and CD2, and lost CD3G and CD7 but gained CD7 and MYB. Normally expressing myeloid and variably T and pDC markers, in BPDCN it retained only CD11b, and upregulated CD56. The top 10 gene expressions are enriched in cancer progression, ALL, and Wnt/Beta-catenin signaling regulation (Figure 6 & Figure S4-S9) (Tables 2 and 3).

Pseudotime analysis

1. Cluster 1, 2, 3, 5, 6, 7, 8, 11, and 12 expressing CD4 are cell populations that may differentiate into plasma cell DC (pDC).

2. Cluster 1, 2, 3, 7, 8, and 12 expressing both CD4 and CD123 are likely to be the cell populations that begin to differentiate into plasma cell DC (pDC).

3. Cluster 3, 8, and 12 expressing CD4, CD123, and CD303 are late-stage cell populations that have differentiated into plasma cell DC (pDCs).

4. Cluster 3 and 8, which express CD4, CD123, CD303, and CD56 simultaneously, can be defined as a cell population typical of BPDCN.

5. The cell population with high MYB expression can be divided into three types:

Cluster 1 is a group of BPDCN cells that are highly expressed in MYB and express CD4, CD123, and CD56, but not CD303, and should be defined as a group of BPDCN cells with malignant potential that appear earlier.

Cluster 12, which highly expresses MYB and expresses CD4, CD123, and CD303 but does not express CD56, should be defined as a population of plasma cell DC (pDC) cells that lack the characteristics of BPDCN but possess malignant potential.

Cluster 3 and Cluster 8, both of which express high levels of MYB and all of the genes CD4, CD123, C303, and CD56, should be defined as typical BPDCN cell clusters (Figure 7).

Fig. 7.

A. Based on the results calculated by Monocle2, the trajectory is plotted in the graph, where the points represent cells, the colors represent different cell clusters, group information, cell types, sample information, and state status, with the black solid line indicating the cell differentiation trajectory. B-D. The cell trajectory graph based on the cluster, cell type, sample, and group information in Fig. 7A has been split, with each cluster, cell type, sample, and group displayed separately on the trajectory graph. In the graph, the points represent cells, the colors represent different cell clusters, group information, cell types, sample information, and state status, with the black solid line indicating the cell differentiation trajectory. E-F. The cell trajectory graph of the cells from BPDCN group and CTRL group. Green cell population representing the BPDCN group and purple cell population representing the CTRL group in the illustration. G. A pie chart showing the proportion of different states in BPDCN and CTRL samples

KEGG analysis

KEGG analysis found that The main differentially expressed genes between BPDCN and normal control are enriched in the following three pathways: T cell receptor signaling pathway, Natural killer cell mediate cytotoxity, Hematopoietic cell lineage (Fig. 8).

Fig. 8.

(A) Arrange the fold changes of differentially expressed genes from largest to smallest, and select the top 25 up-regulated and down-regulated genes to draw a heatmap: The horizontal axis represents the differential grouping information, and the vertical axis represents the top 25 up-regulated and down-regulated genes (if there are fewer than 25 differentially expressed genes in either up- or down-regulation, all genes will be plotted; mitochondrial genes and ribosomal genes are not plotted by default). Yellow indicates high expression, and purple indicates low expression. (B) KEGG Enrichment Analysis Top 20 (select pathways with corresponding differential gene counts greater than 2, sorted by each item’s -log10Pvalue from highest to lowest) bubble chart. The horizontal axis Enrichment Score represents the enrichment score, the larger the bubble, the more differential protein-encoding genes it contains. The bubble color changes from purple to blue to green to red, and the smaller the enrichment p Value, the more significant it is

Discussion

Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare, aggressive blood cancer originating from plasmacytoid dendritic cell precursors. It can affect tissues like skin, bone marrow, lymph nodes, and blood, and is identified by markers CD4, CD56, CD123, and either CD303 or CD304 [10]. The precise etiology of BPDCN is not fully understood but involves various molecular disruptions. This includes deletions affecting tumor regulators and suppressor genes, and mutations in multiple genes like CDKN2A/B, CDKN1B, RB1, BCL2, CCND1, ATM, TP53, ASXL1, IDH2, NPM1, TET2, KRAS, NRAS, IKZF1/2/3, and ZEB2. BPDCN also features chromosomal rearrangements involving genes such as ALK, ETV6, and MYC. These genetic changes disrupt cell processes and likely drive the disease’s development [11].

The role of the MYB gene in BPDCN

In 2017, Suzuki’s team studied 14 BPDCN patients, including 5 children and 9 adults, through RNA sequencing and transcriptome analysis. They found MYB gene rearrangements in all children and 4 adults. The study identified gene fusions like MYB-ZFAT, MYB-PLEKHO1, MYB-DCPS, and MYB-MIR3134, present at diagnosis and relapse but absent in remission. The frequent MYB rearrangements are linked to increased MYB transcriptional activity in BPDCN [9]. The study also found MYB and MYC gene rearrangements in some BPDCN patients. These rare translocations in myeloid malignancies may cause the MYB protein’s regulatory domain to be shortened, possibly leading to its overactivation and playing a role in BPDCN’s oncogenesis [12–14]. These insights into the molecular pathology of BPDCN have expanded our understanding of the disease’s genetic underpinnings and may have implications for targeted therapeutic strategies [5].

Single cell sequencing of BPDCN in previous study

Erica A. K. DePasquale analyzed 52,803 single-cell transcriptomes, including 18,779 T-cells, to study the effects of malignant BPDCN cell growth on the tumor environment. The study showed increased interferon alpha (IFNA) response and decreased tumor necrosis factor alpha (TNFA) signaling in BPDCN patients. It also found significant T-cell exhaustion, linked to T-cell clonotype expansion, offering insights into BPDCN’s immune interactions [6].

Characterization of each cell cluster in the current study

Cluster 1

Top genes expressed in cluster 1 are correlated with tumor spread, microenvironment, and prognosis. This suggested the molecular signatures’ link to BPDCN’s clinical traits.

Cluster 2

The gene profile of cluster2 was rich in oncogenesis and c-myc pathway genes, indicating a key role in tumorigenesis. It’s suggested that post-MYB activation, this initially unmarked cell group transforms into a mixed lineage phenotype with T-, B-, erythroid, and pDC traits, forming a diverse tumor cell population. This reflects the tumor microenvironment’s plasticity and the complex interaction of lineage and oncogenic factors in BPDCN.

Cluster 3

The top 10 genes expressed in cluster 3 are linked to dendritic cell development, highlighting their role in BPDCN’s pathogenesis and the complexity of cell differentiation and oncogenic signaling.

Cluster 4

The top 10 gene expressions in cluster 4 relate to T cell oncogenesis, tumor immunology, and dendritic cell functions, indicating a complex T cell role in BPDCN’s immunological and neoplastic processes.

Cluster 5

This cluster, mainly myeloid, did not seem to become tumorigenic. However, some of its most highly expressed genes were linked to oncogenesis and the tumor microenvironment, indicating that these cells, though not turning into tumor cells, are affected by BPDCN’s neoplastic processes. This reveals the myeloid compartment’s role in the disease’s tumorigenic landscape.

Cluster 6

This T cell population’s variable phenotype is influenced by BPDCN. Highly expressed genes in this group relate to cancer development, T cell activation, T-B cell communication, dendritic cell maturation, and cancer-associated pathways like C-MYB and NF-kappaB. This suggests a complex role for these T cells in BPDCN’s pathophysiology, potentially in immune response and tumor progression.

Cluster 7

The phenotype of cluster 7 indicates a tumor-influenced phenotype shift. Classified as a myeloid subset, it shows phenotypic changes due to tumorigenesis. Its top 10 genes are linked to immune regulation, cancer progression, and T and dendritic cell development, intersecting with Wnt and MYC pathways critical for cellular transformation and tumor maintenance. This reveals the myeloid compartment’s complex interaction with the BPDCN tumor microenvironment.

Cluster 8

This cell group, once without specific markers, has transformed in BPDCN, adopting a profile with T, B lymphoid, erythroid elements, and BPDCN markers, likely due to MYB oncogene activation. This indicates a new tumor-associated cell population. The top genes in this cluster relate to dendritic cell development, chromosomal stability, DNA damage response, TP53, and cell cycle regulation. These suggest a complex network in BPDCN’s pathogenesis, showing these genes’ roles in tumor-associated cell development and behavior.

Cluster 9

The phenotypic change of cluster 9 in BPDCN may due to the tumor suggests a role in modulating NK cell activity, with the top 10 gene expressions linked to NK cell function regulation, potentially impacting the immune response and BPDCN’s pathophysiology.

Cluster 10

The phenotype and top genes expression in Cluster 10 suggested a role in B cell-T cell interplay and BPDCN’s neoplastic process. This implicates Cluster 10 in disease progression and identifies potential B and T cell regulatory network targets.

Cluster 11

This suggests a shift towards a megakaryocyte-like state, with the role in tumorigenesis to be clarified. Top gene expressions of this cluster offered insights into cellular adaptations in BPDCN’s neoplastic landscape.

Cluster 12

This heterogeneous cluster shows a multilineage shift with tumor-associated markers, suggesting an immature cell population that changes in the tumor microenvironment. The top 10 gene expressions of Cluster 12 indicating a complex network influencing this cell population’s tumorigenic potential and behavior in BPDCN.

Based on the findings above, our study’s single-cell sequencing categorized cells into four main clusters: Cluster 1 as primitive plasmacytoid dendritic cells (pDCs); Cluster 2 and 3 as mixed lineage tumor-associated cells; and Cluster 8 as a tumor-associated cell population due to MYB activation. All clusters are MYB-positive but show varied expression of pDC markers, suggesting diverse cellular states and roles in BPDCN’s pathophysiology.

In the control group, Cluster 1 is marked by CD303, a mature pDC marker. In BPDCN, there’s a loss of CD303 with the gain of CD56 and MYB, indicating a reversion to an immature pDC stage due to the oncogenic MYB gene activation.

Cluster 2 lacks clear lineage markers in controls but shows significant immunophenotypic changes with MYB activation, gaining CD4, CD123, CD7, and Kappa/Lambda light chains, indicating a mixed lineage cell population emergence.

Cluster 3, a minor pDC group in controls, expresses CD4, CD123, CD303. In BPDCN, MYB activation leads to increased CD56 and other markers, indicating a large mixed phenotype tumor cell population.

Cluster 8, like Cluster 2, is unmarked in controls but gains CD4, CD123, CD303, CD56, and immunoglobulin light chains with MYB activation, showing a tumor-associated cell population with an aberrant phenotype. This highlights MYB’s significant role in BPDCN’s cellular differentiation and identity.

The study indicates that BPDCN’s tumor cellularity mainly consists of four unique cell populations, with some possibly evolving from pDCs. The origins of Cluster 2 and 8 from specific precursors are still unclear.

KEGG pathways

T cell receptor signaling pathway

The TCR signaling pathway is essential for activating T lymphocytes, key to the immune system’s ability to respond effectively. It starts with TCR binding to peptide-MHC complexes and is modulated by molecules like CD28, leading to T cell functions including proliferation and cytokine production.

In dendritic cells (DCs), maturation involves cytokine secretion like IL-12, crucial for T cell response initiation. Mature DCs also express chemokine receptors like CCR7, aiding migration to lymphoid tissues for immune response regulation. The interaction between TCR signaling and DC function is fundamental to the immune response’s coordination and effectiveness.

Natural killer cell mediate cytotoxity

NK cells are vital in the immune response to infections and diseases, with the ability to perform antibody-dependent cellular cytotoxicity (ADCC). They target IgG-coated “foreign” cells like virus-infected or cancerous cells by binding to the Fc portion of antibodies that have attached to the cells.

NK cells and dendritic cells (DCs) interact, which is crucial for initiating and coordinating adaptive immunity against cancer. This interaction has been associated with improved survival and responses to anti-PD-1 therapy in metastatic melanoma patients. Immature DCs are vulnerable to NK-cell cytolysis, while mature DCs are resistant, allowing NK cells to regulate DC function by eliminating immature DCs and activating other immune cells.

Hematopoietic cell lineage

Dendritic cells (DCs) are essential for the immune system, facilitating antigen presentation for immunity and tolerance. Originating from multipotent, self-renewing hematopoietic stem cells (HSCs), DCs contribute to the development of all blood cell types, including immune cells and components like erythrocytes and platelets.

DCs are present in various tissues, interacting with T and B cells to regulate innate and adaptive immune responses. Their communication with these cells is crucial for immune activation against pathogens and for maintaining self-tolerance.

Conclusion

1. The MYB-ZFAT fusion gene produces a 347-amino-acid protein, causing the native MYB gene’s inhibitory domain to be removed. This leads to uncontrolled MYB activation, a key abnormality in BPDCN’s development. The loss of this domain may trigger neoplastic change by disrupting the gene’s normal role, thus fueling the disease’s severity.

2. Single-cell sequencing shows BPDCN has a diverse cell population, contradicting the idea of a uniform cell type. The varied origins of tumor cells highlight the need for complex treatment strategies. In BPDCN patients with the MYB-ZFAT fusion, a significant cell subset with high MYB expression likely fuels the disease. Conversely, the smaller fraction without MYB expression mainly consists of non-BPDCN cells, indicating a more complex cellular makeup. Understanding this heterogeneity is crucial for developing targeted, personalized treatments.

3. The KEGG analysis points to key biological pathways in BPDCN tumorigenesis. The T cell receptor (TCR) signaling, vital for T lymphocyte activation, is a significant factor. The NK cell cytotoxicity pathway, important for immune defense against abnormal cells, is also involved in BPDCN. The hematopoietic cell lineage, involving blood cell development from stem cells, is linked to disease progression. Further research into these pathways’ interactions is needed to understand BPDCN’s pathogenesis. Targeting these processes could reveal new therapeutic approaches and improve treatment effectiveness.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Dr. Ruijuan Li and Dr. Wenyong Kuang performed main experiments. Miss Kexin Zhao and Miss Weiyi Fang helped with partial experiments. Dr. Wenyong Kuang provided patients’ sample. Dr. Haixia Yang and Dr. Benshan Zhang helped with diagnosis and treatment of patient. Dr. Zhao Cheng designed the study, analyzed data, and composed manuscript. Dr. Xianming Fu. provided the funding to support this study. Professor Hongling Peng supervised this study. All authors read and approved the final manuscript.

Funding

National Natural Science Foundation of China, 81400093 (Zhao Cheng); Natural Science Foundation of Hunan Province, 2018JJ3757 (Zhao Cheng); Natural Science Foundation of Hunan Province, 2021JJ30937 (Ruijuan Li); Natural Science Foundation of Hunan Province, 2023JJ60415 (Zhao Cheng); Scientifc Program of Health Commission of Hunan Province, 202203042723 (Ruijuan Li); Scientifc Program of Health Commission of Hunan Province, B202303046108 (Zhao Cheng); Scientifc Program of Health Commission of Hunan Province, 202104020628(Xianming Fu); Scientific Program of Health Commission of Hunan Province, 20200639 (Wenyong Kuang); Changsha Municipal Natural Science Foundation, kq2014234 (Ruijuan Li); Changsha Municipal Natural Science Foundation, kq2208319 (Zhao Cheng).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Compliance with ethical standards

The experimental protocol was established according to the ethical guidelines of the Helsinki Declaration and was approved by the Human Ethics Committee of The Second Xiangya Hospital, Central South University. P.R. China. Written informed consent was obtained from individual or guardian participants.

Competing interests

The authors declare no competing interests.