Colony-stimulating factor 3 and its receptor promote leukocyte immunoglobulin-like receptor B2 expression and ligands in gastric cancer

Abstract

BACKGROUND

Colony-stimulating factor 3 (CSF3) and its receptor (CSF3R) are known to promote gastric cancer (GC) growth and metastasis. However, their effects on the immune microenvironment remain unclear. Our analysis indicated a potential link between CSF3R expression and the immunosuppressive receptor leukocyte immunoglobulin-like receptor B2 (LILRB2) in GC. We hypothesized that CSF3/CSF3R may regulate LILRB2 and its ligands, angiopoietin-like protein 2 (ANGPTL2) and human leukocyte antigen-G (HLA-G), contributing to immunosuppression.

AIM

To investigate the relationship between CSF3/CSF3R and LILRB2, as well as its ligands ANGPTL2 and HLA-G, in GC.

METHODS

Transcriptome sequencing data from The Cancer Genome Atlas were analyzed, stratifying patients by CSF3R expression. Differentially expressed genes and immune checkpoints were evaluated. Immunohistochemistry (IHC) was performed on GC tissues. Correlation analyses of CSF3R, LILRB2, ANGPTL2, and HLA-G were conducted using The Cancer Genome Atlas data and IHC results. GC cells were treated with CSF3, and expression levels of LILRB2, ANGPTL2, and HLA-G were measured by quantitative reverse transcriptase-polymerase chain reaction and western blotting.

RESULTS

Among 122 upregulated genes in high CSF3R expression groups, LILRB2 showed the most significant increase. IHC results indicated high expression of LILRB2 (63.0%), ANGPTL2 (56.5%), and HLA-G (73.9%) in GC tissues. Strong positive correlations existed between CSF3R and LILRB2, ANGPTL2, and HLA-G mRNA levels (P < 0.001). IHC confirmed positive correlations between CSF3R and LILRB2 (P < 0.001), and HLA-G (P = 0.010), but not ANGPTL2 (P > 0.05). CSF3 increased LILRB2, ANGPTL2, and HLA-G expression in GC cells. Heterogeneous nuclear ribonucleoprotein H1 modulation significantly altered their expression, impacting CSF3’s regulatory effects.

CONCLUSION

The CSF3/CSF3R pathway may contribute to immunosuppression in GC by upregulating LILRB2 and its ligands, with heterogeneous nuclear ribonucleoprotein H1 playing a regulatory role.

Key Words: Gastric cancer; Immunosuppressive receptor; Colony-stimulating factor 3; Colony-stimulating factor 3 receptor; Leukocyte immunoglobulin-like receptor B2; Angiopoietin-like protein 2; Human leukocyte antigen-G; Heterogeneous nuclear ribonucleoprotein H1

Core Tip: This study investigates the relationship between colony-stimulating factor 3 (CSF3)/colony-stimulating factor 3 receptor (CSF3R) and leukocyte immunoglobulin-like receptor B2, along with its ligands angiopoietin-like protein 2 and human leukocyte antigen-G, in gastric cancer (GC). High CSF3R expression correlates with increased levels of leukocyte immunoglobulin-like receptor B2, along with its ligands angiopoietin-like protein 2, and human leukocyte antigen-G in both transcriptomic data and immunohistochemical analysis. CSF3 upregulates these immune checkpoints in GC cells, with heterogeneous nuclear ribonucleoprotein H1 modulating this effect. These findings suggest that the CSF3/CSF3R pathway may play a role in promoting immunosuppression in GC.

INTRODUCTION

Inhibitory receptors (IRs) are common immune checkpoints (ICs) found in various immune cell subpopulations[1]. They transmit inhibitory signals to immune cells, dampening immune activity to maintain immune system balance and prevent autoimmunity[1]. Tumor cells exploit IRs to evade immune surveillance, suppressing the immune system[1]. IC inhibitors (ICIs) treat tumors by blocking IR activation, promoting immune cell activation, and enhancing their ability to eliminate tumor cells[1]. Currently, the most widely used ICIs in clinical practice target the programmed cell death protein ligand 1 (PD-L1)/programmed cell death protein 1 (PD-1) pathway. Gastric cancer (GC) is a highly prevalent malignancy worldwide, with a high incidence and mortality rate. China has the highest incidence of GC, accounting for about half of new cases globally in 2020[2]. Patients with advanced GC (AGC) often have a short-term survival. Combining chemotherapy with PD-1 inhibitors has significantly extended the overall survival (OS) of AGC patients. Studies such as ORIENT-16 and CheckMate-649 have demonstrated that adding ICIs to chemotherapy markedly improves the OS of AGC patients, particularly those with high PD-L1 expression[3,4]. Furthermore, human epidermal growth factor receptor 2 (HER2)-positive AGC patients can benefit from ICI treatment. Pembrolizumab, an ICI, when combined with trastuzumab and chemotherapy, further enhances the objective response rate in HER2-positive AGC[5]. ICIs are now considered a crucial treatment following anti-HER2 therapy for AGC. Despite the significant OS prolongation with ICIs, the survival of AGC patients remains less than two years. Many patients exhibit primary or secondary resistance to PD-L1/PD-1 inhibitors[6]. In addition to PD-1, other important IRs include cytotoxic T lymphocyte-associated antigen-4, T cell immunoreceptor with Ig and inhibitory motifs (ITIMs) domains, and others. Inhibitors targeting these receptors have been extensively studied in clinical trials for cancer treatment[7,8]. While IRs are predominantly expressed in immune cells such as activated T cells, B cells, dendritic cells (DCs), natural killer cells, and monocytes, recent studies have also found IR expression on tumor cells[9]. IRs expressed on tumor cells, termed endogenous IRs, such as endogenous PD-1 in cancer cells, can either promote or inhibit tumor growth upon activation[10,11]. Hence, it is crucial to explore new mechanisms of tumor-induced immune tolerance and identify alternative checkpoint molecules to overcome immune resistance and extend patient survival in GC.

Colony-stimulating factor 3 (CSF3) is a type of hematopoietic growth factor, with its receptor CSF3 receptor (CSF3R) being expressed in hematopoietic stem cells, bone marrow progenitor cells, platelets, and mature bone cells[12]. When CSF3 binds to CSF3R, it promotes the proliferation and differentiation of bone marrow cells, enhances the function of mature neutrophils, and mobilizes bone marrow granulocytes to the peripheral blood[13]. Our previous research has shown that high expression of CSF3 and CSF3R in GC is positively correlated with tumor-node-metastasis (TNM) stage and lymph node metastasis, respectively[14]. Patients with higher CSF3 expression have a shorter OS[14]. Leukocyte immunoglobulin-like receptor B2 (LILRB2), also known as immunoglobulin-like transcript 4, belongs to the immunoglobulin superfamily and is mainly expressed in myeloid cells such as monocytes, macrophages, DCs, and granulocytes[15]. It is also expressed in various other cell types like hematopoietic stem cells, osteoclast precursor cells, platelets, and neurons, and regulates their biological functions[15]. Ligands for LILRB2 include angiopoietin-like protein 2 (ANGPTL2), human leukocyte antigen-G (HLA-G), complement component C4d, and semaphorin-4Ad, and these ligands are expressed in different cell types[15]. Upon activation of LILRB2, an immunosuppressive signal is transmitted, leading to inhibition of immune cell activation[16]. CSF3 can stimulate monocytes to change to DC cells by up-regulating the expression of LILRB2 and HLA-G, and then immunological activity is inhibited. These DCs exhibit poor capacity to induce T-cell proliferation and interferon-gamma secretion[17]. In our research, we found that LILRB2 was the most relevant immunosuppressive receptor to CSF3R in GC tissues. Therefore, this study aimed to investigate the correlation between CSF3/CSF3R and LILRB2 and its ligands in GC, hoping to gain a better understanding of the immune environment of GC.

MATERIALS AND METHODS

Bioinformatics analysis

To investigate CSF3R gene expression in GC patients, we utilized The Cancer Genome Atlas (TCGA) cohort. The TCGA-Stomach adenocarcinoma data set was used to download the data (https://portal.gdc.cancer.gov/). The CSF3R gene was selected as the target gene, and it was categorized into groups based on high and low levels of expression for subsequent differential analysis. To conduct the differential analysis, the 375 samples were divided into two groups: 188 samples with low expression and 187 samples with high expression. The criteria for conducting the differential screening were set as follows: A P-value of 0.05 and an |log2 fold change| > 1.

Immunohistochemical staining

Samples for this study were obtained from 46 GC patients who underwent surgery at the Fourth Hospital of Hebei Medical University in Shijiazhuang, China, between May 2020 and October 2020. Of these patients, 33 were men and 13 were women. The TNM stage of GC patients was classified according to the 8^th edition of the American Joint Committee on Cancer staging system. Written informed consent was obtained from all patients. The study protocol was approved by the Review Board and Ethics Committee of the Fourth Hospital of Hebei Medical University and conducted in accordance with the principles outlined in the World Medical Association Declaration of Helsinki. GC tissue specimens were fixed with 10% paraformaldehyde and embedded in paraffin for further sectioning. Immunohistochemical staining for LILRB2, HLA-G, ANGPTL2, and CSF3R was performed using primary antibodies: Anti-LILRB2 (1:100, Sangon Biotech, Shanghai, China), anti-HLA-G (1:100, Sangon Biotech, Shanghai, China), anti-ANGPTL2 (1:100, Sangon Biotech, Shanghai, China), and anti-CSF3R (1:100, Sangon Biotech, Shanghai, China). Evaluation of all specimens was conducted by two independent pathologists who were blinded to the clinical records of the patients. The scoring system was based on the sum of the percentage of stained cells (0: 0%-5%, 1: 6%-25%, 2: 26%-50%, 3: 51%-75%, 4: 76%-100%) multiplied by a number (0-3) reflecting the intensity of staining (0: Negative, 1: Weak, 2: Moderate, 3: Strong). A score of 4 or lower indicated low expression of LILRB2, HLA-G, ANGPTL2, or CSF3R, while a score higher than 4 indicated high expression.

Cell culture

AGS and BGC-823 cells were purchased from Procell Life Science and Technology Co., Ltd. and cultured in Ham’s F-12 medium (Cytiva Company, Marlborough, MA, United States) supplemented with 10% fetal bovine serum and antibiotics. The cells were maintained at 37 °C in a humidified incubator with 5% CO₂. In the experimental group, CSF3 (final concentration 100 ng/mL) was added to the regular culture medium to treat the cells. Following treatment, cells from both the experimental and control groups were collected at 6 hours, 12 hours, 24 hours, 48 hours, and 72 hours, respectively. CSF3 was added to cells with downregulated heterogeneous nuclear ribonucleoprotein H1 (hnRNPH1) and the control group. Cells from both groups were collected at 48 hours, respectively.

Quantitative reverse transcriptase-polymerase chain reaction analysis

Total RNA was extracted from cells using Redzol reagent (SBS Genetech, Beijing, China), and cDNA was synthesized using the First-Strand cDNA Synthesis Kit (Cat. No: QP056, GeneCopeia, United States). The mRNA expression levels of LILRB2, ANGPTL2, and HLA-G in each sample were determined by quantitative reverse transcriptase-polymerase chain reaction using specific primers. The mRNA levels in each sample were normalized to the relative quantity of GAPDH. All experiments were conducted in triplicate. The specific primer sequences used were as follows: LILRB2: Forward 5’-GCATTTGGCGGCTTCATT-3’, reverse 5’-GTGCGACCACCTGCGATT-3’; ANGPTL2: Forward 5’-TCCTGCACGAGATCATCCG-3’, reverse 5’- GGTGCTGGTACTTGTGCTCC-3’; HLA-G: Forward 5’-ATTACCTCGCCCTGAACG-3’, reverse 5’-CCAGAAGGCACCACCACA-3’; GAPDH: Forward 5’-GTGGACCTGACCTGCCGTCT-3’, reverse 5’-GGAGGAGTGGGTGTCGCTGT-3’.

Western blotting

Cells in each group were completely lysed using RIPA lysis buffer (Solarbio, China) to obtain total cellular protein. The protein content was promptly measured using spectrophotometry. Subsequently, 25% volume of 5 × loading buffer was added to denature the proteins by boiling the samples at 100 °C for 5 minutes. An appropriate amount of protein sample was loaded onto 8% or 10% sodium-dodecyl sulfate gel electrophoresis gel (BOSTER, China) for electrophoresis. The wet transfer method was employed to transfer proteins to a polyvinylidene fluoride membrane (Millipore, United States). The membrane was then blocked with 5% skim milk powder at 37 °C for 2 hours. For immunoblotting, the following primary antibodies were used: Rabbit anti-LILRB2 polyclonal antibody (1:1000, Thermo, United States), rabbit anti-ANGPTL2 polyclonal antibody (1:1000, Proteintech, China), rabbit anti-HLA-G polyclonal antibody (1:1000, Proteintech, China), rabbit anti-β-Tubulin monoclonal antibody (1:1000, Abcam, United Kingdom) and rabbit anti-GAPDH monoclonal antibody (1:1000, Abcam, United Kingdom). The primary antibodies were incubated on the polyvinylidene fluoride membrane overnight at 4 °C. Subsequently, the membrane was washed three times with tuberculous antigen-based skin test for 10 minutes each time. Following primary antibody incubation, diluted goat anti-rabbit immunoglobulin G secondary antibody (horseradish peroxidase labeled, 1:5000, Abcam, United Kingdom) was added and incubated at 37 °C for 1 hour. After the secondary antibody incubation, the membrane was washed again. Chemiluminescence detection was performed using Western Lightning™ Chemiluminescence Reagent (Servicebio, China). The membrane was fully covered with the reagent and then scanned and imaged using an Epson Perfection V39 scanner. The brightness value of each protein band was analyzed. To normalize protein expression, the brightness value of each sample’s protein band was divided by the brightness value of the corresponding GAPDH band. The brightness value of the control group was set to 1 for comparative analysis. Finally, a histogram was generated to visualize the results.

Transient transfection

Small interfering RNA (siRNA) duplexes targeting hnRNPH1 were procured from GenePharma Company. The sequences of the siRNAs were as follows: 5’-GGAUUACCUUACAGAGCUATT-3’ and 5’-UAGCUCUGUAAGGUAAUCCTT-3’. To knockdown hnRNPH1 in cells, a siRNA targeting the hnRNPH1 coding sequence was designed and incorporated into a lentiviral vector obtained from GenePharma Company, China. A scramble siRNA was used as a negative control. The efficacy of knockdown was evaluated by both quantitative reverse transcriptase-polymerase chain reaction and western blot analysis.

Plasmid transfection

To achieve overexpression of hnRNPH1 in cells, we generated an expression construct by subcloning the polymerase chain reaction-amplified full-length human hnRNPH1 cDNA into the pcDNA3.1(+) vector (GenePharma Company, China). As a negative control, an empty vector was used. The efficiency of overexpression was assessed by western blot analysis.

Statistical analysis

The statistical analysis was conducted using SPSS version 27.0, with the χ² test utilized to analyze the correlation of count data. GraphPad Prism version 9 software was employed for further statistical analysis and visualization of the data. Group comparisons were performed using one-way analysis of variance. All experiments were conducted in triplicate, and the results are presented as mean ± SD. Statistical significance was set at P < 0.05.

RESULTS

LILRB2 is the immune IR most closely associated with CSF3R expression

The GC tissues were divided into the high CSF3R expression group (187 samples) and the low CSF3R expression group (188 samples) based on the median value of the CSF3R expression level. Differential expression analysis was conducted between the two groups. It was found that all 122 differentially expressed genes (DEGs) were significantly up-regulated in the high CSF3R expression group, with none of the DEGs being down-regulated (Figure 1A and B). Enrichment analysis of the DEGs showed that DEGs were mainly associated with cellular pathways such as Staphylococcus aureus infection, rheumatoid arthritis, and leishmaniasis, as identified by Kyoto Encyclopedia of Genes and Genomes analysis. Gene Ontology enrichment analysis revealed that the DEGs were closely associated with biological processes such as leukocyte migration, positive regulation of cytokine production, and myeloid leukocyte migration. The cellular component analysis showed that the DEGs were predominantly located on the external side of the plasma membrane. In terms of molecular function, the DEGs were mainly associated with immune receptor activity, chemokine activity, chemokine receptor binding, and other pathways (Figure 1C and D). The top 10 significantly up-regulated genes among the DEGs were ranked as follows: MNDA, FPR1, LILRB2, HCK, LCP2, NCF2, ITGAX, CLEC4E, SRGN, and PILRA. As DEGs were primarily accumulated in immune receptor activity in molecular function, we examined the expression variations of 23 common ICs between tissues with high and low CSF3R expression. In 22 of these ICs, expression was upregulated in the high CSF3R expression group compared to the low expression group. The six most significantly differentially expressed ICs were LILRB2, hepatitis a virus cellular receptor, chemokine receptor 2, CD80, inducible T-cell costimulator, and CD28 (all P < 0.001) (Figure 2). The expression of other ICs in the two groups is shown in Supplementary Figure 1. Bioinformatics analysis revealed that LILRB2 was highly up-regulated in the tissues with high CSF3R mRNA expression.

Figure 1 Analysis of differentially expressed genes between patients with high and low colony-stimulating factor 3 receptor expression. A: Volcano plots of differentially expressed genes (DEGs); B: Heatmap of DEGs; C: Enrichment analysis of DEGs by Kyoto Encyclopedia of Genes and Genomes; D: Enrichment analysis of DEGs by Gene Ontology.

Figure 2 Immune checkpoints showing the most significant differences between high and low colony-stimulating factor 3 expression groups. A: Leukocyte immunoglobulin-like receptor B2; B: Hepatitis A virus cellular receptor 2; C: Chemokine receptor 2; D: CD80; E: Inducible T-cell costimulator; F: CD28. ^dP < 0.0001.

LILRB2 and its ligand ANGPTL2 and HLA-G are highly expressed in GC tissues

We performed immunohistochemistry (IHC) staining to assess the protein expression levels of LILRB2 and its ligands, ANGPTL2 and HLA-G, in GC tissues. The findings revealed that among the 46 GC tissues, the expression levels of LILRB2, ANGPTL2, and HLA-G were found to be high in 63.0% (29), 56.5% (26), and 73.9% (34) samples, respectively (Figure 3A-C). ANGPTL2 expression was significantly correlated with TNM stage (χ² = 13.400, P < 0.001) and lymph node metastasis (χ² = 21.854, P < 0.001). However, no correlation was observed with other pathological factors such as age, gender, depth of tumor infiltration, and Lauren classification (P > 0.05). On the other hand, the protein expression of LILRB2 and HLA-G in GC tissues did not show any correlation with age, gender, depth of tumor infiltration, lymph node metastasis, tumor TNM stage, and Lauren classification (P > 0.05) (Table 1).

Figure 3 Immunohistochemical staining of gastric cancer tissues (200 ×). A: Leukocyte immunoglobulin-like receptor B2 (high expression); B: Angiopoietin-like protein 2 (high expression); C: Human leukocyte antigen-G (high expression); D: Colony-stimulating factor 3 receptor (high expression); E: Leukocyte immunoglobulin-like receptor B2 (low expression); F: Angiopoietin-like protein 2 (low expression); G: Human leukocyte antigen-G (low expression); H: Colony-stimulating factor 3 receptor (low expression).

Table 1 Relationship between the expression of leukocyte immunoglobulin-like receptor B2, angiopoietin-like protein 2, and human leukocyte antigen-G in gastric cancer tissues and clinicopathological characteristics, n (%).

Characteristics	LILRB2		ANGPTL2		HLA-G
	High	Low	High	Low	High	Low
Age (years)
≤ 60	18 (62.1)	6 (35.3)	14 (53.8)	10 (50.0)	16 (47.1)	8 (66.7)
> 60	11 (37.9)	11 (64.7)	12 (46.2)	10 (50.0)	18 (52.9)	4 (33.3)
Gender
Male	20 (69.0)	13 (76.5)	18 (69.2)	15 (75.0)	22 (64.7)	11 (91.7)
Female	9 (31.0)	4 (23.5)	8 (30.8)	5 (25.0)	12 (35.3)	1 (8.3)
TNM stage
I/II	16 (55.2)	9 (52.9)	8 (30.8)	17 (85.0)a	17 (50.0)	8 (66.7)
III	13 (44.8)	8 (47.1)	18 (69.2)	3 (15.0)	17 (50.0)	4 (33.3)
Tumor depth
T1/2	11 (37.9)	7 (41.2)	7 (26.9)	11 (55.0)	13 (38.2)	5 (41.7)
T3	18 (62.1)	10 (58.8)	19 (73.1)	9 (45.0)	21 (61.8)	7 (58.3)
Lymph nodes metastases
No	10 (34.5)	9 (52.9)	3 (11.5)	16 (80.0)a	12 (35.3)	7 (58.3)
Yes	19 (65.5)	8 (47.1)	23 (88.5)	4 (20.0)	22 (64.7)	5 (41.7)
Lauren typing
Intestinal-type	12 (41.4)	6 (35.3)	12 (46.2)	6 (30.0)	13 (38.2)	5 (41.7)
Diffuse-type	9 (31.0)	7 (41.2)	7 (26.9)	9 (45.0)	12 (35.3)	4 (33.3)
Mixed-type	8 (27.6)	4 (23.5)	7 (26.9)	5 (25.0)	9 (26.5)	3 (25.0)

^aP < 0.001.

LILRB2: Leukocyte immunoglobulin-like receptor B2; ANGPTL2: Angiopoietin-like protein 2; HLA-G: Human leukocyte antigen-G; TNM: Tumor-node-metastasis.

Correlations between CSF3R and LILRB2, ANGPTL2 and HLA-G

To further investigate the relationship between CSF3/CSF3R and LILRB2, we examined the correlation between CSF3R and LILRB2, as well as its ligands, in GC tissues using IHC. The protein expression level of CSF3R was significantly correlated with LILRB2 (χ² = 24.156, P < 0.001) and HLA-G (χ² = 6.639, P = 0.010), but not with ANGPTL2 (χ² = 1.458, P = 0.227) (Table 2). The IHC expression of CSF3R is depicted in Figure 3D. Figure 3E-H shows low expression of LILRB2, ANGPTL2, HLA-G, and CSF3R as detected by IHC. No significant correlation was found between LILRB2 and ANGPTL2 (χ² = 2.584, P = 0.108) or between LILRB2 and HLA-G (χ² = 2.064, P = 0.151) (Table 3). However, a positive correlation was found between the protein expression of ANGPTL2 and HLA-G (χ² = 10.494, P = 0.001). Upon analyzing the TCGA-Stomach adenocarcinoma dataset, significant correlations were identified between CSF3R and LILRB2 (r = 0.785, P < 0.001), ANGPTL2 (r = 0.335, P < 0.001), and HLA-G (r = 0.240, P < 0.001). A weak correlation was observed between LILRB2 and ANGPTL2 (r = 0.405, P < 0.001), as well as between LILRB2 and HLA-G (r = 0.232, P < 0.001). However, no correlation was found between ANGPTL2 and HLA-G (r = -0.028, P = 0.591) (Figure 4).

Figure 4 Correlation of mRNA expression. A-F: Correlation of mRNA expression levels between colony-stimulating factor 3 receptor, leukocyte immunoglobulin-like receptor B2, angiopoietin-like protein 2, and human leukocyte antigen-G. CSF3R: Colony-stimulating factor 3 receptor; LILRB2: Leukocyte immunoglobulin-like receptor B2; ANGPTL2: Angiopoietin-like protein 2; HLA-G: Human leukocyte antigen-G.

Table 2 Correlation between the protein expression levels of leukocyte immunoglobulin-like receptor B2, angiopoietin-like protein 2, and human leukocyte antigen-G, and colony-stimulating factor 3 receptor in gastric cancer tissues by immunohistochemistry, n (%).

		CSF3R
		High	Low
LILRB2	High	29 (85.3)	0 (0.0)
	Low	5 (14.7)	12 (100.0)a
ANGPTL2	High	21 (61.8)	5 (41.7)
	Low	13 (38.2)	7 (58.3)
HLA-G	High	29 (85.3)	5 (41.7)
	Lowa	5 (14.7)	7 (58.3)b

^aP < 0.001.

^bP = 0.01.

LILRB2: Leukocyte immunoglobulin-like receptor B2; ANGPTL2: Angiopoietin-like protein 2; HLA-G: Human leukocyte antigen-G; CSF3R: Colony-stimulating factor 3 receptor.

Table 3 Correlation between the protein expression levels of leukocyte immunoglobulin-like receptor B2, angiopoietin-like protein 2, and human leukocyte antigen-G in gastric cancer tissues by immunohistochemistry, n (%).

		LILRB2
		High	Low
ANGPTL2	High	19 (65.5)	7 (41.2)
	Low	10 (34.5)	10 (58.8)
HLA-G	High	24 (82.8)	10 (58.8)
	Low	5 (17.2)	7 (41.2)

LILRB2: Leukocyte immunoglobulin-like receptor B2; ANGPTL2: Angiopoietin-like protein 2; HLA-G: Human leukocyte antigen-G.

CSF3 promotes LILRB2, ANGPTL2 and HLA-G expression in GC cells

LILRB2, ANGPTL2 and HLA-G are expressed in GC cells, and the mRNA expression levels were decreased as the culture time increased (Figure 5). When GC cells were exposed to CSF3 for various durations (6 hours, 12 hours, 24 hours, 48 hours, and 72 hours), the expression levels of mRNA (Figure 5) and protein (Figure 6) for LILRB2, ANGPTL2, and HLA-G were notably elevated compared to the control group lacking CSF3 treatment. This suggests that CSF3 promotes the mRNA and protein expression of LILRB2, ANGPTL2, and HLA-G in GC cells.

Figure 5 The mRNA expression of leukocyte immunoglobulin-like receptor B2, angiopoietin-like protein 2, and human leukocyte antigen-G. A-C: The mRNA expression levels at different time points; D-F: The mRNA expression levels at different time points after administration of colony-stimulating factor 3 compared with control groups. ^aP < 0.001; ^bP < 0.01; ^cP < 0.05. LILRB2: Leukocyte immunoglobulin-like receptor B2; ANGPTL2: Angiopoietin-like protein 2; HLA-G: Human leukocyte antigen-G; CSF3: Colony-stimulating factor 3; NC: Negative control.

Figure 6 Protein expression of leukocyte immunoglobulin-like receptor B2, angiopoietin-like protein 2, and human leukocyte antigen-G at different time points after administration of colony-stimulating factor 3 compared with control groups. A-C: Protein expression levels at different time points. D: Protein expression at different time points. ^aP < 0.001; ^cP < 0.05. LILRB2: Leukocyte immunoglobulin-like receptor B2; ANGPTL2: Angiopoietin-like protein 2; HLA-G: Human leukocyte antigen-G; CSF3: Colony-stimulating factor 3.

HnRNPH1 plays a regulatory role in CSF3-mediated expression

We discovered a correlation between CSF3/CSF3R and hnRNPH1 expression (the data are not presented in this article). Consequently, we conducted further investigations to determine if CSF3/CSF3R regulates the expression of LILRB2, ANGPTL2, and HLA-G through hnRNPH1. Upon reducing hnRNPH1 expression, both mRNA and protein levels of LILRB2, ANGPTL2, and HLA-G were significantly decreased (Figure 7A-D). Conversely, increasing hnRNPH1 expression led to a significant increase in mRNA and protein levels of LILRB2, ANGPTL2, and HLA-G (Figure 7A-D). These findings indicate that hnRNPH1 influences the expression of LILRB2, ANGPTL2, and HLA-G in GC cells. We then downregulated hnRNPH1 expression in cells using siRNA and stimulated the cells with CSF3. The cells were divided into four groups: Control, control + CSF3, si-hnRNPH1 and si-hnRNPH1 + CSF3. We observed a significant decrease in the upregulation effect of CSF3 on these genes when hnRNPH1 expression was downregulated, compared to the control groups (Figure 7E-H). These findings suggest that hnRNPH1 plays an important role in the regulation of LILRB2, ANGPTL2, and HLA-G by CSF3.

Figure 7 Colony-stimulating factor/colony-stimulating factor 3 receptor regulates the expression of leukocyte immunoglobulin-like receptor B2, angiopoietin-like protein 2, and human leukocyte antigen-G through heterogeneous nuclear ribonucleoprotein H1. A-C: The mRNA expression levels when heterogeneous nuclear ribonucleoprotein H1 (hnRNPH1) expression is downregulated or upregulated; D: Protein expression levels when hnRNPH1 expression is downregulated or upregulated; E-G: The mRNA expression levels when hnRNPH1 expression is downregulated with or without colony-stimulating factor 3 stimulation; H: Protein expression levels when hnRNPH1 expression is downregulated with or without colony-stimulating factor 3 stimulation. ^aP < 0.0001; ^bP < 0.001; ^cP < 0.01; ^dP < 0.05. LILRB2: Leukocyte immunoglobulin-like receptor B2; ANGPTL2: Angiopoietin-like protein 2; HLA-G: Human leukocyte antigen-G; hnRNPH1: heterogeneous nuclear ribonucleoprotein H1; CSF3R: Colony-stimulating factor 3 receptor.

DISCUSSION

LILRB2 is a transmembrane protein with four extracellular tandem Ig-like domains, a 23-amino acid transmembrane region, and a cytoplasmic tail containing three immunoreceptor tyrosine ITIMs[18,19]. Mainly expressed in monocytes, macrophages, and DCs, LILRB2 is a potent immune suppression receptor that binds classical and nonclassical major histocompatibility complex class I molecules[20]. When LILRB2 is activated, tyrosine kinases phosphorylate the tyrosine residues in ITIM, leading to the delivery of inhibitory signals and suppression of cell activation[21]. LILRB2 can inhibit the release of cytotoxic molecules from monocytes or macrophages against tumor cells, inhibit the inflammatory reaction mediated by monocytes and macrophages, and inhibit the maturation of DC cells, resulting in the impotence of CD4⁺ T cells, and promote the production of myeloid-derived suppressor cells, thus inhibiting the ability of immune cells to respond, and promoting tumor immune escape[22-24]. When the expression of LILRB2 is increased in infiltrating DCs of hepatocellular carcinoma, this may lead to an immunosuppressive microenvironment through the secretion of the inflammatory interferon-gamma[25]. The expression of LILRB2 in tumor tissue is associated with the prognosis of patients. A study showed that LILRB2 is scarcely expressed in normal endometrium, but is expressed in endometrial cancer tissue, resulting in shorter OS[26]. In GC tissue, patients with high LILRB2 expression have shorter OS compared to those with low LILRB2 expression (42.9 months vs 84.5 months)[27]. The OS of male patients with low LILRB2 expression was 67.3 months and was 47.1 months in those with high LILRB2 expression. The OS of female patients with low LILRB2 expression was 89.3 months and was 17.9 months in those with high LILRB2 expression[27]. Therefore, high LILRB2 expression in GC tissue suggests a poor prognosis, particularly in female patients[27].

Two ligands, ANGPTL2 and HLA-G, have been extensively studied for their interactions with LILRB2. ANGPTL2 belongs to the ANGPTL family and is a secreted glycoprotein[28]. ANGPTL2 prevents apoptosis, promotes angiogenesis and the activities and proliferation of hematopoietic stem cells[28]. The expression of ANGPTL2 is significantly higher in GC tissues than in paracancerous tissues, and its high expression is positively associated with pathological stage and lymph node metastasis[29]. ANGPTL2 and LILRB2 overexpression are found in non-small cell lung cancer and colorectal cancer tissues, but not in normal tissues[30,31]. Colorectal cancer patients with high expression of LILRB2 and ANGPTL2 have poorly differentiated tumor cells and more advanced disease[31]. HLA-G is a type of non-classical HLA class I molecule that leads to immune evasion of cancer cells and the progression of tumors[32,33]. High levels of HLA-G expression in malignant tumors are strongly associated with metastasis, advanced disease stages, chemoresistance, and poor survival[32-34]. In GC patients, those who tested positive for HLA-G had significantly shorter OS compared to those who tested negative[35]. In another study, only female GC patients with high HLA-G expression showed a poor prognosis[27]. In colorectal cancer patients, simultaneous overexpression of LILRB2 and HLA-G correlated with larger tumor volumes, later TNM stages, and shorter OS[36]. Thus, the overexpression of LILRB2, ANGPTL2, and HLA-G may affect the prognosis of patients by promoting tumor growth. In this study, LILRB2, ANGPTL2, and HLA-G were highly expressed in 63.0%, 6.5%, and 73.9% GC tissues, respectively. Limited by the small number of tissue specimens collected, only ANGPTL2 was found to be correlated with TNM stage and lymph node metastasis. Thus, more samples are needed to determine the correlation between LILRB2, ANGPTL2, and HLA-G and clinicopathology.

CSF3 is a member of the hematopoietic factor family, which can enhance the proliferation and differentiation of hematopoietic stem cells, promote maturation of neutrophils, mobilize neutrophils from bone marrow into the peripheral blood, and activate the effector functions of mature neutrophils[37]. Our previous study has indicated that the expression level of CSF3 and CSF3R was up-regulated in GC tissue[14]. High expression of these factors in GC patients is linked to a higher risk of metastasis and shorter survival[14]. CSF3 induces the proliferation and migration of GC cells, and stimulates the tubule formation of vascular endothelial cells, which may further promote the growth and metastasis of GC[14]. In this study, a significant positive correlation was found between the expression of CSF3R, LILRB2, ANGPTL2 and HLA-G in GC.

Apart from suppressing immune cell activity, the activation of endogenous IRs expressed on tumor cells can trigger multiple signal transduction pathways, such as mechanistic target of rapamycin, Hippo, and mitogen-activated protein kinase, promoting tumor growth[11,38-41]. Elevated expression of LILRB2 in cancer cells (iLILRB2) has been identified in various tumor types, including lung cancer, breast cancer, prostate cancer, and melanoma[42]. LILRB2 mRNA is highly expressed in lung cancer, breast cancer, prostate cancer, and melanoma, when compared to the normal epithelium[42]. The overexpression of iLILRB2 can enhance the proliferation and migration of tumor cells, activate the extracellular signal-regulated kinases 1 and 2 (ERK1/2) signal pathway, and promote fatty acid synthesis[42]. Overexpression of iLILRB2 can activate the intracellular ERK1/2 signaling pathway, promoting fatty acid synthesis, thus leading to increased tumor cell proliferation and migration[42]. Blocking the intracellular ERK1/2 signaling pathway or fatty acid synthesis can inhibit the proliferation and migration of tumor cells induced by LILRB2 overexpression, and prevent senescence in CD4⁺ and CD8⁺ T cells that is mediated by LILRB2[42]. In our in vitro experiments, it was demonstrated that CSF3 promoted the expression of iLILRB2, ANGPTL2, and HLA-G in GC cells. Reducing hnRNPH1 expression significantly decreased mRNA and protein levels of LILRB2, ANGPTL2, and HLA-G. Conversely, increasing hnRNPH1 expression significantly increased their levels. Downregulation of hnRNPH1 attenuates the upregulatory effect of CSF3 on these genes, indicating its pivotal role in CSF3-mediated gene expression.

CONCLUSION

LILRB2 belongs to the family of immunosuppressive receptors. The observed positive association between CSF3/CSF3R and LILRB2, along with its ligands in GC, highlights the pivotal role of the CSF3/CSF3R pathway in promoting an immunosuppressive microenvironment. HnRNPH1 emerges as a crucial regulator, as its downregulation significantly reduces the upregulatory effect of CSF3 on LILRB2, ANGPTL2, and HLA-G genes, emphasizing its significance in CSF3-mediated gene expression. Further investigation is essential to delineate the precise involvement of CSF3/CSF3R in the GC immune microenvironment and its potential implications in ICIs resistance.

ACKNOWLEDGEMENTS

We would like to sincerely thank Yue-Ping Liu and Chun-Xia Ding from the Department of Pathology at the Fourth Hospital of Hebei Medical University for their invaluable assistance in specimen collection. We also extend our heartfelt gratitude to Wen-Xuan Liu from the Department of Epidemiology and Health Statistics at Hebei Medical University for her statistical review. Additionally, we appreciate the support of Jing Zuo and Yu-Dong Wang from the Department of Oncology for their contributions to our experimental research.