Decoding the biological properties and transcriptomic landscapes of human natural killer cells derived from bone marrow and umbilical cord blood

Abstract

Longitudinal studies have highlighted allogeneic natural killer (NK) cell-based cytotherapy for cancer immunosurveillance and immunotherapy, yet the deficiency of systematic and detailed comparison of NK cells from candidate sources including umbilical cord blood (UC) and bone marrow (BM) largely hinders the large-scale application. Herein, we isolated resident NK cells (rUC-NK, rBM-NK) from mononuclear cells (MNC), and analyzed the corresponding expanded NK cell counterparts (eUC-NK, eBM-NK). Then, the eUC-NK and eBM-NK were turned to multifaceted bioinformatics from the aspects of gene expression profiling and genetic variations. The percentages of total or activated NK cells in rBM-NK group were approximate 2-fold higher over those in the rUC-NK group, respectively. Instead, the proportion of total NK cells in eUC-NK was higher than that in the eBM-NK group, and in particular, the CD25⁺ memory-like NK cell subset. Furthermore, eUC-NK and eBM-NK manifested multidimensional similarities and diversities in gene expression pattern and genetic spectrum, whereas both eUC-NK and eBM-NK exhibited effective tumor killing capacity. Collectively, we dissected the cellular and transcriptomic signatures of NK cells generated from UC-MNC and BM-MNC, which supplied new literature for further exploring the characteristics of the indicated NK cells and would benefit the clinical application for cancer immunotherapy in future.

Introduction

As one of the life-threatening diseases with high morbidity and mortality, cancers have emerged as a heavy burden for individual patients and economic society worldwide [1,2]. For example, cancers were recognized as the second most common causes of death among kids in America, and over 10,000 children and 5,000 adolescents were diagnosed with cancers in 2021 [1]. For decades, we and other investigators in the field have been devoted to figure out the pathogenesis of cancers from the view of genetic variations and environmental stimulus such as heterogeneity, uncontrolled cell division, cancer stem cells (CSCs), proto-oncogenes and anti-oncogenes, lonizing radiation, dangerous carcinogens, and even the coronavirus disease 2019 (COVID-19) pandemic [3-6].

To date, a variety of treatment regiments have arisen for oncotherapy, including surgery (e.g., robotic surgery, laparoscopic rectal surgery), chemotherapy, oncolytic virotherapy, radiotherapy, hormone therapy, photothermal therapy, peptide-based neoantigen vaccine, RNA vaccine, nanotheranostics and immunotherapy [7-11]. Despite the considerable advancements in cancer administration, more effective implementation remedies are still urgently needed for the improvement of the incidence and survival rates and the concomitant optimization of cancer diagnosis and treatment [12,13]. For instance, the progress in medical practice has strengthened for lung cancer but stagnated for prostate and breast cancers instead [5]. Of note, state-of-the-art renewal has highlighted the feasibility of cancer immunotherapy as a revolutionizing and promising route for targeted cancer control interventions and the resultant reductions in cancer mortality [6,8,14].

Natural killer (NK) cells are heterogeneous populations and play a vital role in both innate and adoptive immune responses dispense with preliminary antigen presentation attributes to the non-MHC-restricted recognition [7,12,15]. For decades, NK cells have been generated from perinatal blood (e.g., umbilical cord blood, placental blood) and adult blood (e.g., bone marrow, peripheral blood), NK cell lines (e.g., NK-92, YT), and even stem cells (e.g., hematopoietic stem cells, pluripotent stem cells) [7,16-21]. Bone marrow-derived NK cells are lymphocytes resided in the hematopoietic microenvironment, which have been recognized as crucial constituents for host defense against diverse infections and tumors [22]. Currently, due to the deficiency of peripheral blood, umbilical cord blood has been regarded as alternative source for high-efficient NK cell generation and cancer immunotherapy [7,23,24]. However, the systematic and detailed information of the similarities and variations between the tissue resident or ex vivo expanded BM-NK cells and UC-NK cells is still largely obscure, which will largely hinder the further development of allogeneic NK cell-based novel regimens.

For the purpose, in this study we utilized our well-established “3IL”-based strategy for ex vivo NK cell induction from umbilical cord blood- and bone marrow-derived mononuclear cells (MNCs). With the aid of multifaceted cellular evaluation and high-throughput sequencing, we verified the similarities and distinctions of the aforementioned NK cells in biological properties (e.g., immunophenotype, cellular vitality, activation and cytotoxicity) and transcriptomic landscapes (e.g., gene expression profiling, gene ontology, signaling pathway, genetic variations). Taken together, our data provide new references for decoding the biological and transcriptomic signatures of UC-NK cells with BM-NK cells, which will benefit the further investigations upon NK cell-based cancer immunotherapy.

Materials and methods

Enrichment of mononuclear cells (MNCs)

Human bone marrow tissue and umbilical cord blood were obtained from healthy donors with the approval of the Ethics Committee of Shangrao ETD-H&B Hospital (KLL-2020-04) and Gansu Provincial Hospital (2023-120) according to the guidelines of Helsinki. In general, BM-MNCs and UC-MNCs were isolated by utilizing the standard Ficoll (Sigma-Aldrich, USA)-based density gradient centrifugation method as we recently described [19,25].

In vitro NK cell induction from MNCs

In vitro amplification and activation of NK cells were conducted as we recently reported by utilizing our well-established “3IL”-based strategy. In details, 2 × 10⁶/ml BM-MNCs or UC-MNCs were seeded in NK MACS Medium (Miltenyi Biotech, Germany) supplemented with cytokine cocktail addition (rhIL-2, 1000 U/mL; rhIL-15, 10 ng/ml; rhIL-18, 50 ng/ml) [19]. The medium was half changed every two days until day 14. Trypan Blue staining and flow cytometry (FCM) assay were used for cell counting and quantification, respectively. The list of the cytokines was available in Supplementary Table 1.

Flow cytometry (FCM) assay

FCM assay was employed for the proportion analyses of total NK cells (CD3^-CD56⁺), specific NK cell subsets (CD16⁺, CD25⁺, NKp44⁺, NKp46⁺, NKG2D⁺) and relative populations (CD4⁺, CD8⁺, Annexin V⁺, 7-AAD⁺, CD107a⁺) as we described before [20,21]. In details, the MNCs (UC-MNCs, BM-MNCs), resident NK cells (rUC-MNCs, rBM-MNCs), and expanded NK cells (eUC-MNCs, eBM-MNCs) were incubated with the indicated fluorescence conjugated antibodies in dark for 30 min, and washed with 1 × PBS for further detection. Then, the aforementioned cells were turned to FACS Canto II (BD Biosci, USA) and FlowJo 10.0 software (Tree Star, USA) for analysis, respectively. The list of the antibodies was available in Supplementary Table 2.

Cellular vitality assay

The cellular vitality of the indicated NK cells was accomplished with cell cycle assay and apoptosis assay as we recently reported [19,21]. For cell cycle assay, NK cells were pre-treated with 70% (v/v) cooled ethanol and fixed at 4°C. After washing with 1 × PBS, the cells were turned to Propidium iodide (PI) staining and BD LSR II (BD Biosci, USA) for analysis. For cell apoptosis assay, 1 × 10⁶ cells were washes with precooled 1 × PBS and incubated with the Annexin V Apoptosis Detection Kit including Annexin V-FITC and 7-AAD solution (Sigma-Aldrich, USA). Finally, the cells were turned to FACS Canto II (BD Biosci, USA) for apoptotic cell analysis.

RNA-sequencing (RNA-SEQ) and bioinformatics analysis

For total mRNA preparation, eUC-NK and eBM-NK at day 14 were lysed with TRIZol reagent (ThermoFisher, USA) according to the manufacturer’s instructions [21,26]. The mRNAs were turned to BGI Genomics (Shenzhen, China) for sequencing. The bioinformatics analysis such as Principal Component Analysis (PCA), HeatMap, Volcano Plot, hierarchical cluster analysis, Gene Set Enrichment Analysis (GSEA, https://www.gsea-msigdb.org/gsea/index.jsp), Gene Ontology (GO, http://geneontology.org/), Kyoto Encyclopedia of Genes and Genomes (KEGG, https://www.kegg.jp/) were conducted as we recently described [27,28]. The list of DEGs between eBM-NKs and eUC-NKs was available in Supplementary Table 3.

Cytotoxicity assessment of NK cells

The cytotoxicity of the indicated eUC-NK cells and eBM-NK cells was performed as we previously reported [19,20]. Generally, tumor cell lines (K562, Nalm6, U937) were harvested and labelled with CellTrace Violet (Invitrogen, USA) as we described with several modifications [19,29]. Then, the eUC-NK cells or eBM-NK cells were cocultured with the aforementioned tumor cells at an effector-to-target ratios (E:T=1:1) for 8 hrs. After that, the cells were labeled with the indicated antibodies (e.g., anti-CD3, ant-CD56, ant-CD16, anti-CD107a, 7-AAD) and Precision Count Beads (BioLegend, USA). Finally, the incubated cells were washed with 1 × PBS and tested by FACS Canto II (BD Biosci, USA). NK cell Cytotoxicity = (1 - N₂/N₁) × 100%. N₁ represents the total number of living tumor cells in control group, N₂ represents the total number of living tumor cells in the experimental group [19,21].

Statistical analyses

All statistical analyses were accomplished with the Graph Pad Prism 6.0 (San Diego, USA) software as we previously described [30,31]. The student’s unpaired T test was conducted to analyze the data between two unpaired groups, whereas the one-way ANOVA test was utilized to analyze the data among multiple unpaired groups. Data were shown as mean ± SEM (N=3 independent experiments), when P value <0.05 was recognized as statistically significant. NS, not significant; *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001.

Results

Resident NK cells in bone marrow revealed higher proportion and activity over those in umbilical cord blood

To assess the potent similarities and distinctions of NK cells generated from bone marrow and umbilical cord blood, we initially enriched MNCs and explored the corresponding resident NK cells (rNK). According to the FCM diagrams, the proportions of CD3^-CD56⁺ total NK cells and CD3^-CD56⁺CD16⁺ activated NK cells in the rBM group were approximate 2-fold higher than those in the rUC, respectively (Figure 1A-D). Furthermore, except for the CD3^-CD56⁺NKp44⁺ subset, minimal differences were observed in the percentages of other activated NK cell subsets such as the NKG2D⁺, NKp46⁺ and CD25⁺ counterparts (Figure 1C, 1D). Taken together, BM-MNCs rather than UC-MNCs revealed higher percentages of total and activated rNK cells.

Figure 1.

The comparison of resident NK cells in MNCs enriched from bone marrow and umbilical cord blood. (A, B) Representative flow cytometry (FCM) diagrams (A) and statistical analysis (B) of resident bone marrow-derived mononuclear cells (rBM) and umbilical cord blood-derived mononuclear cells (rUC). (C, D) Representative FCM diagrams (C) and statistical analysis of the NK cell (D) subpopulations of rBM-NK and rUC-NK. All data were shown as mean ± SEM (N=3). NS, not significant; ****, P<0.0001.

Expanded NK cells from UCB manifested superiority in amplification over those generated from BM but with comparable cellular vitality

Having dissected the rNK in BM-MNCs and UC-MNCs, we next turn to explore the biological signatures of the corresponding expanded NK cells (eBM-NK, eUC-NK) by utilizing our well-established “3IL”-based strategy [19,21]. As shown by the phase contrast images, the seeded cells (day 0) in both BM-NK and UC-NK groups revealed unicellular morphology, and more typical NK cell clusters were arisen during the ex vivo induction (Figure 2A). Compared to the BM-NK group, approximate 2-fold change of eNK cells was observed in the UC-NK group at day 14 (Figure 2B).

Figure 2.

The comparison of the expanded NK cells in content and cellular vitality. (A) Representative phase contract images of MNC-derived NK cells with our well-established “3IL”-based strategey. Scale bar =100 μm. (B) Fold change of NK cells from the resident NK cells (rNK) at day 0 to the corresponding expanded NK cells (eNK) at day 14. The fold changes were normalized to rNK at day 0. (C, D) Representative FCM diagrams (C) and statistical analysis (D) of apoptotic NK cells in eBM-NK and eUC-NK at day 14. (E, F) The representative distribution (E) and statistical analysis (F) of eBM-NK and eUC-NK in the indicated substages of cell cycle. All data were shown as mean ± SEM (N=3). NS, not significant; **, P<0.01; ****, P<0.0001.

Simultaneously, we took advantage of FCM analyses to evaluate the cellular vitality of eNK at day 14. As shown by the FCM diagrams and statistical analysis, no significant differences were observed in the percentage of apoptotic NK cells (Figure 2C, 2D). Furthermore, with the aid of cell cycle assay, we noticed that the proportion of G0/G1 subset rather than the G2/M subset in the eBM-NK group was moderately higher over that in the eUC-NK group (Figure 2E, 2F). Collectively, our data indicated the preferable ex vivo amplification property of eUC-NK over eBM-NK, together with the minimal differences in cellular vitality.

Expanded NK cells and resident NK cells exhibited multifaceted diversity in subpopulations

To further explore the cellular phenotypes of eUC-NK and eBM-NK, we conducted FCM assay and found that the percentages of both CD3^-CD56⁺ total NK cells and CD3^-CD56⁺CD25⁺ memory-like NK cells were much higher in the eUC group after the 14-day’s ex vivo expansion and activation, whereas the proportion of CD3^-CD56⁺NKp44⁺ activated NK cell subset was lower compared to the eBM group instead (Figure 3A-D). Distinguish from those in the rNK, we noticed there were minimal differences in the percentage of CD3^-CD56⁺CD16⁺ total activated NK cell between the eBM-NK and eUC-NK group (Figure 3D). Similarly, the proportions of CD3^-CD56⁺ total NK cells and activated NK cell subsets in eNK (eBM-NK, eUC-NK) at day 14 were consistently elevated and higher than the corresponding rNK (rBM-NK, rUC-NK) at day 0 except for the CD3^-CD56⁺CD16⁺ total activated NK cells in the rBM-NK and eBM-NK groups, respectively (Figure 3E, 3F). Overall, these findings indicated the further maturation of rNK towards eNK as well as high-efficient ex vivo NK cell generation after the “3IL”-based stimulation, and in particular, the eUC-NK with more robust amplification and activation properties.

Figure 3.

The comparison of resident NK cells and expanded NK cells. (A, B) Representative FCM diagrams (A) and statistical analysis (B) of eBM and eUC at day 14. (C, D) Representative FCM diagrams (C) and statistical analysis (D) of the NK cell subpopulations in rNK (rBM-NK, rUC-NK) at day 0 and the corresponding eNK (eBM-NK, eUC-NK) at day 14. (E, F) Statistical analysis of total NK cells (E) and NK cell subpopulations (F) among the indicated rNK (rBM-NK, rUC-NK) and eNK (eBM-NK, eUC-NK). All data were shown as mean ± SEM (N=3). NS, not significant; ***, P<0.001; ****, P<0.0001.

The gene expression profiling of expanded BM-NK cells and UC-NK cells manifested multidimensional similarity and diversity

Having illuminated the biological properties of the eNK, we were next curious about the multifaceted characteristics at molecular levels. For the purpose, we took advantage of RNA-SEQ for transcriptomic analysis of eBM-NK (eBM-NK-1, eBM-NK-2, eBM-NK-3) and eUC-NK (eUC-NK-1, eUC-NK-2, eUC-NK-3). Initially, as shown by the stacked bar chart and Violin Plot diagrams, we noticed the similarities between eBM-NK and eUC-NK in overall gene expression pattern (Figure 4A, 4B). In details, the distribution of the significantly upregulated and downregulated differently expressed genes (DEGs) genes as well as the non-DEGs between the indicated eNK were intuitively shown by the Volcano Plot (Figure 4C).

Figure 4.

Comparison of gene expression profiling of eBM-NK and eUC-NK. (A) The stacked bar chart of gene numbers with the indicated expression levels based on TPM values in eBM-NK (eBM-NK-1, eBM-NK-2, eBM-NK-3) and eUC-NK (eUC-NK-1, eUC-NK-2, eUC-NK-3). (B, C) The violin plot (B) and Volcano plot (C) of gene expression pattern in the aforementioned eBM-NK and eUC-NK. (D) The PCA analysis of the indicated eBM-NK and eUC-NK based on FPKM values. (E) The correlation of affinity among the indicated eBM-NK and eUC-NK. (F) HeatMap diagram of the differentially expressed genes (DEGs) in eBM-NK and eUC-NK. (G, H) The gene ontology biological process (GOBP) assay based on the differentially expressed genes (G) and transcripts (H) between eBM-NK and eUC-NK.

Simultaneously, with the aid of principal component analysis (PCA) and correlation analysis, the genetic relationship of individual eBM-NK and eUC-NK were visually presented (Figure 4D, 4E). Furthermore, by conducting the hierarchical cluster analysis based on TPM (transcripts per kilobase of exon model per million mapped reads) values of DEGs, we found eBM-NK (eBM-NK-1, eBM-NK-2, eBM-NK-3) and eUC-NK (eUC-NK-1, eUC-NK-2, eUC-NK-3) revealed preferable relatedness within groups rather than between groups, respectively (Figure 4F). Subsequently, with the aid of gene ontology biological process (GOBP) assay of DEGs, we noticed cancer immunotherapy-associated biological processes such as immune response, negative regulation of cell proliferation, ERK1/2 cascade, and chemotaxis were specifically enriched (Figure 4G). Similarly, according to the GOBP assay of the differentially expressed transcripts, we observed the enrichment of transcription-associated processes including positive or negative regulation of transcription, chromatin organization, and regulation of immune response (Figure 4H). Collectively, eBM-NK and eUC-NK manifested multifaceted similarities and diversities in gene expression profiling, which further indicated the consistency and variations with the biological properties.

The landscapes of signaling pathway and variation spectrum of expanded BM-NK cells and UC-NK cells

To further estimate the genetic characteristics of the indicated eNK, we turned to kyoto encyclopedia of genes and genomes (KEGG) analysis, and found that the DEGs were mainly involved in cytokine-cytokine receptor interaction-, antigen processing and presentation-, NK cell mediated cytotoxicity-, TGF-β signaling pathway-, NF-Kappa B signaling pathway-, and chemokine signaling pathway-related processes (Figure 5A). The KEGG network assay of the aforementioned DEGs further intuitively revealed the relevance of the central signaling pathways (Figure 5B). Therewith, by conducting gene set enrichment analysis (GSEA), we observed the specific enrichment of the significantly different gene sets between eBM-NK and eUC-NK such as PPAR signaling pathway, TLR1/2 ligand vs NOD2 monocyte, fatty acid metabolism and glycolysis, whereas minimal differences were shown in KARS signaling pathway and complement and coagulation cascades (Figure 5C, 5D). Furthermore, the protein-protein interaction (PPI) network and KDA diagram consistently reflected the profile of the spatial relationship of the representative core DEGs, including CCL2, JUN, CD40LG, IFNG, TNF, and FOXP3 (Figure 5E, 5F).

Figure 5.

Signaling pathway analysis and variation spectrums between eBM-NK and eUC-NK. (A, B) KEGG analysis (A) and pathway correlation assay (B) of hAMSCs in eBM-NK and eUC-NK based on DEGs. (C, D) GSEA diagrams of significantly different gene sets (C) and non-different gene sets (D) between eBM-NK and eUC-NK. (E) PPI analysis of the DEGs (log₂FC >1.0) between eBM-NK and eUC-NK. (F) KDA analysis of the DEGs (log₂FC >1.0) between eBM-NK and eUC-NK. (G) Variable shear events in eBM-NK and eUC-NK. (H) The distribution of the differentially VSE (DVSE) between eBM-NK and eUC-NK. (I) Circos diagrams revealed the distribution of fusion genes in eBM-NK and eUC-NK.

For the purpose of further illuminating the potential similarities and differences in genetic modifications accompanied with the DEGs between the indicated eNK, we meticulously compared the overview of variable shear event (VSE), and verified the similarity in as_a3ss, as_a5ss, as_mxe, as_ri and as_se distribution (Figure 5G). Furthermore, among the differentially VSE (DVSE), the as_se subset were the principal variations in genetic modification (Figure 5H). Additionally, as shown by the Circos diagrams, the loci regional distribution of the indicated somatic variations such as gene fusion events, SNPs, INDELs and FPKM values could be intuitively observed in the chromosomes of eBM-NK and eUC-NKs (Figure 5I). Collectively, these data indicated the multidimensional conservations and alterations in the spectrum of genetic modifications between eBM-NK and eUC-NKs.

Expanded BM-NK cells and UC-NK cells showed effective lethality towards multiple tumor cells in vitro

Due to the boundedness of rNK in quantity and immaturity, we turned to the eNK for ex vivo cytotoxic assessment. Therefore, we took advantage of the NK cell and tumor cell line coculture model at the effector-target ratio (E:T=1:1), and verified that eBM-NKs revealed a higher percentage of CD107⁺ subset than that in eUC-NK against K562 and U937 cells, whereas no significant differences were observed against Nalm6 cells (Figure 6A, 6B). Meanwhile, the cytotoxicity of eBM-NKs against the target Nalm6 cells rather than the K562 and U937 tumor cells was equal to that of eUC-NKs, which was confirmed by BV421-based fluorescence according to the statistical analyses (Figure 6C, 6D). Taken together, our data suggested that the eBM-NK revealed moderate advantage over eUC-NK in ex vivo cytotoxicity.

Figure 6.

Comparison of the tumor killing cytothoxicity of eBM-NK and eUC-NK. (A) Representative FCM diagram (A) and statistical analysis (B) of CD107a⁺ eBM-NK and eUC-NK in coculturing with multiple tumor cell lines (K562, Nalm6, U937) at the effector-to-target ratios (E:T=1:3); (C, D) The comparison of the cytotoxicity of eBM-NK and eUC-NK against the aforementioned tumor cell lines based on the calculations of tumor cell number (C) and proportion of tumor cells (D). All data were shown as mean ± SEM (N=3). NS, not significant; *, P<0.05; ***, P<0.001.

Discussion

Cancers, including the metastatic solid tumors and hematologic malignancies, have caused thousands on thousands of deaths worldwide due to the high morbidity and mortality. In spite of the encouraging breakthroughs in cancer amelioration and treatment, yet the favorable and persistent prognosis of cancer patients are still challenging and urgently to be improved. Of note, state-of-the-art literatures have highlighted the rosy prospect of adoptive NK cell-based cytotherapy for cancer immunotherapy, yet the large-scale application in clinical practices are largely restricted to the deficiency of systematic and detailed dissection of the signatures at cellular and molecular levels. Herein, we took advantage of the “3IL”-based procedure for ex vivo amplification and activation of rNK into eNK, and verified the multidimensional biological properties and transcriptomic characteristics of NK cells derived from bone marrow and umbilical cord blood, together with the cytotoxicity assessment against tumor cell lines, which would collectively benefit the further exploration of NK cell-based remedies for cancer administration.

For decades, surgeries (e.g., robotic surgery) in combination with chemoradiotherapy (e.g., radiotherapy, chemotherapy) and drug therapy have been considered as classical therapeutic modalities for localized cancers without diffusion and metastasis [10,32]. In recent years, novel strategies in the field of cancer immunotherapy including autogenous and allogeneic immune cells (e.g., TILs, CILs, NK cells, CAR-T cells, CAR-NK cells), lymphocyte-promoting cytokines (e.g., GM-CSF, IFN-γ), checkpoint inhibitors (e.g., PD-L1, CTLA4), cancer vaccines (e.g., mRNAs), and monoclonal antibodies (mAb) has been continuously developed to fulfil cancer management [33-36]. Of note, chimeric antigen receptor-transduced T lymphocytes (CAR-Ts) have been proved by preclinical and clinical investigations with remarkable efficacy for various hematological malignancies such as B-cell non-Hodgkin’s lymphoma (NHL) and relapsed B acute lymphoblastic leukemia (r/rB-ALL) [37-41]. Similarly, T cell receptor-engineered T cells (TCR-Ts) are recognized as splendid alternatives to eliminate a relatively broad range of cancers by redirecting the recognition capability of cancer-associated surface antigens [42]. Distinguish from the aforementioned remedies, adoptive NK cells or CAR-NK cells hold advantaged superiority in cancer immunotherapy via simultaneously modulating the innate and adaptive immunity dispense with antigen presentation and sensitization. Moreover, the NK cell-based immunotherapy is not inclined to cause severe adverse reactions including graft-versus-host disease (GvHD), immune cell-associated neurotoxicity syndrome (ICANS), cytokine release syndrome (CRS) [12,43,44]. Therefore, NK cells and CAR-NKs are currently considered as “off-the-shelf” products and a promising virgin ground of cancer immunotherapy and immunosurveillance innovation.

For decades, we and other investigators in the field have been devoted to fulfilling the goal of relapse and recurrent cancer administration by utilizing the NK cell-based immunotherapy for boosting the immune system to strengthen anti-cancer responses and eventually obliterate the metastatic malignancies [7,45,46]. For instance, we have conducted continuous optimization of ex vivo procedures for large-scale preparation of NK cells with preferable cellular vitality and elevated cytotoxicity against a series of tumor cells, including initiating cell source selection (e.g., NK-92, KHYG-1, PBMCs, UC-MNCs, BM-MNCs, placental tissue, CD3^-CD56⁺ rNK), coculture with feeder cells (e.g., K562, K562-mbIL2-I L15-IL18-41BBL), cytokine cocktail-initiated programming (e.g., IL-2, IL-7, IL-18, IL-21), cell sorting or magnetic activated cell sorting (MACS), and physicochemical irritation (e.g., bioreactor, culture vessels) [21,47]. Of note, we verified that the “discarded” perinatal blood including placental blood and umbilical cord blood are preferable alternatives for fulfilling the high-efficient and cost-effective preparation of GMP-grade NK cell products for cancer immunotherapy in future.

NK cells function mainly via antibody-dependent cell-mediated cytotoxicity (ADCC), receptor-ligand related cytotoxicity, secretion of granzyme and perforin, and cytokine-based paracrine effects (e.g., GM-CSF, IFN-γ) [7,48]. Despite the robust prospective of NK cell-based cancer immunotherapy, yet the large-scale application in clinical practice is still restricted largely attributes to the limitation of alternative sources and cost-effective strategies for preparing NK cells with functional maturation after ex vivo “NK cell education” [49]. Meanwhile, the systematic and detailed illumination of NK cells with different origins is simultaneously obscure, which further hinders the development of NK cell-based cytotherapy for cancer immunotherapy. Therefore, depicting the biological signatures and transcriptomic characteristics has been acknowledged to supply basic knowhow for developing and improving novel remedies for cancer management [50,51]. For the purpose, in this study we verified the multifaceted characterizations of the biological and transcriptomic signatures of eBM-NK and eUC-NK, together with the immunophenotypes of the corresponding rBM-NK and rUC-NK as well. Collectively, our findings suggested umbilical cord blood as preferable alternatives over bone marrow for NK cell-based cancer immunotherapy.

Disclosure of conflict of interest

None.

Abbreviations

UCB

Umbilical cord blood

NK

natural killer

DEGs

differentially expressed genes

HSCs

hematopoietic stem cells

PCA

principal component analysis

FPKM

fragments per kilobase million

VSE

variable shear event

DVSE

differentially variable shear event

PPI

protein-protein interaction

GSEA

gene set enrichment analysis

KEGG

Kyoto encyclopedia of genes and genomes

GOBP

gene ontology biological process

TPM

transcripts per kilobase of exon model per million mapped reads

MNCs

mononuclear cells

rNK

resident natural killer cells

eNK

expanded natural killer cells

rBM-NK

resident NK cells in bone marrow-derived MNCs