2.1 Study population

This study included 55 NSCLC patients who had undergone lobectomy/sub-lobectomy and lymph node dissection at Shandong Cancer Hospital and Jinan Central Hospital between 2015 and 2016. The demographic characteristics of the participants are summarized in Table S1.

2.2 Multiplex immunofluorescence (mIF) detection

mIF staining was performed on NSCLC samples using a four-color fluorescence immunohistochemistry kit (abs50012, Absin, China). The formalin-fixed paraffin-embedded lung cancer tissues were sliced into 4-μm-thick sections. The slides were deparaffinized, rehydrated, and subjected to epitope retrieval by boiling in Tris-EDTA buffer (pH = 9) or citrate buffer (pH = 6) in a microwave oven at high heat and maintained for 15 min at low heat. Endogenous peroxidase activity was blocked using Antibody Diluent/Block for 10 min. Primary antibodies to CD3 (Abcam, ab699, 1:200), CD4 (Abcam, ab183685, 1:2000), CD8 (CST, 85336, 1:400), FOXP3 (Abcam, ab215206, 1:500), pan-cytokeratin (Abcam, ab7753, 1:500), CD68 (CST, 76437, 1:400), CD163 (CST, 93498, 1:500), and ILT5 (Immunaway, YN1914, 1:100) were incubated for 1 h at room temperature (20–25°C). Subsequently, anti-rabbit/mouse polymeric horseradish peroxidase was applied as a secondary label, with a 10-min incubation at room temperature. The slides were then incubated with fluorochromes (520, 570, and 650 nm wavelengths) diluted in amplification buffer for 10 min at room temperature. To prevent cross-reactivity, tissue slides were subjected to microwave antigen retrieval in Tris-EDTA buffer (pH = 9) or citrate buffer (pH = 6.0). Tissue slides were counterstained with 4′,6-diamidino-2-phenylindole and mounted using prolonged antifade mountant. All fluorescence-labeled slides were scanned using an OLYMPUS VS200 microscope (Olympus, Tokyo, Japan) at 40× magnification with appropriate exposure times, and five representative regions were selected for quantification at 20× magnification (0.1 mm²). The percentage of different cell subsets was determined as the ratio of positively stained cells to all nucleated cells per square millimeter.²⁵ All images were independently reviewed by two senior pathologists before data export, with at least 20% of each visual field reviewed.

Receiver operating characteristic analysis was performed to determine the cutoff value for ILT5 levels. Based on the optimal cut-off index, ILT5 expression in TAMs was stratified into ILT5-low (≤0.0215) and ILT5-high (>0.0215). Similarly, high and low frequencies of M2-like macrophages (CD163⁺ and CD68⁺) and T cells subgroups, including CD3⁺, CD4⁺, CD8⁺, and FOXP3⁺ T cells, were classified according to their respective optimal cutoff scores.

2.3 Bioinformatics analysis in public databases

The GSE50081 dataset was downloaded from the gene expression omnibus (GEO) database to analyze the survival of NSCLC patients based on ILT5 expression. The correlation between ILT5 and the infiltration level of M2-like TAMs was analyzed using the TIMER2.0 online tool. The TISCH online tool was used to analyze ILT5 expression in different immunocytes within the TME.

2.4 Cells and lentivirus infection

The human acute monocytic leukemia cell lines (THP-1 RRID: CVCL_0006) and lung cancer cell line A-549 (RRID: CVCL_0023) were purchased from The Cell Resource Center of the Chinese Academy of Sciences (Beijing, China). All cell lines were authenticated using short tandem repeat profiling within the last 3 years. All experiments were performed using mycoplasma-free cells. Human peripheral blood mononuclear cells (PBMCs) were isolated from fresh whole blood donated by healthy volunteers using Ficoll-Paque density centrifugation. Human naïve CD3⁺ T cells were purified from PBMCs by negatively selecting EasySep T cell enrichment kits (StemCell Technologies, 50103). The purity of naïve T cells was >97%, as confirmed by flow cytometry. Naïve T cells were activated using anti-human CD3 (BioLegend, 317325)-precoated 6-well plates and cultured with RPMI-1640 solution containing 10% fetal bovine serum (FBS), 1 μg/mL anti-human CD28 (Biolegend, 302933) and 100 IU/mL recombinant human IL-2 (Pepro Tech, 200-02). For lentiviral infection, THP-1 cells were seeded in 6-well plates, and cultured with 100 nM phorbol myristate acetate (PMA) for 24 h. Subsequently, ILT5 overexpression (control and LV-ILT5) and knockdown (control shRNA and shILT5-1 or shILT5-2) lentiviruses were infected using HitransGP Reagents (Genechem, REVG005, Shanghai, China) at a multiplicity of infection of 10–20. After 72 h of infection, THP-1 cells were stimulated with A-549 conditioned medium (CM) to induce their differentiation into TAMs.

2.5 RNA extraction and real-time quantitative polymerase chain reaction (PCR)

THP-1 cells were harvested and RNA was extracted using a total extraction kit (AC0205-B; Sparkjade). Briefly, 1 μg total RNA was used to synthesize the cDNA using the Evo M-MLV RT Premix for quantitative PCR (Accurate Biology, AG11706). ILT5 mRNA expression in THP-1 cells was determined using the SYBR Green Premix Pro Taq HS qPCR kit (Accurate Biology, AG11701) and analyzed by the 2^ΔΔCt method (normalized to beta-actin levels). Each experiment was performed in triplicate. Gene-specific primers used in this study are listed in Table S2.

2.6 Assessing the proliferation ability of T cell subsets using carboxyfluorescein succinimidyl ester (CFSE)

T cell proliferation was assessed using the CellTrace CFSE Cell Proliferation Kit (ThermoFisher, C34554). Briefly, pre-activated T cells were labeled with CFSE at a dilution of 1:1000. The labeled T cells were then cocultured with THP-1 cells (ILT5 up-/down-regulating) for 4 days. CFSE-labeled T cells were then stained with anti-human CD3-PercCP/Cyanine5.5 (BioLegend, 981008), anti-human CD4-PE-Cyanine7 (BioLegend, 317413), and anti-human CD8-APC (BioLegend, 300912). The percentage of dividing cells was analyzed using the FlowJo version 10.0 software (Tree Star).

2.7 Flow cytometry analysis

The proportions of TAM and T cell subsets were determined by flow cytometry after surface or intracellular staining with anti-mouse-specific or anti-humor-specific antibodies conjugated with various fluorescence. After treatment, interferon (IFN)-γ was stained with Cell Stimulation Cocktail plus protein transport inhibitors (eBioscience, 00-4975-93). The following anti-mouse and anti-human antibodies were used: APC/Cyanine7-anti-mouse-CD45 (BioLegend, 109824), FITC-anti-mouse-CD11b (BioLegend, 101206), PerCP/Cyanine5.5-anti-mouse-F4/80 (BioLegend, 123128), PE-anti-mouse-CD163 (BioLegend, 155307), FITC-anti-mouse-CD3 (BioLegend, 100204), PerCP/Cyanine5.5-anti-mouse-CD4 (BioLegend, 100434), PE-anti-mouse-FOXP3 (BioLegend, 320008), PE-anti-mouse-IFN-γ (BioLegend, 505808), PE-antihuman-CD80 (BioLegend, 305208), PE-anti-human-CD86 (BioLegend, 374206), APC-eFluor 780-anti-human-CD3 (eBioscience, 47-0038-42), eFluor 450- anti-human-CD4 (eBioscience, 48-0049-42), Alexa Fluor 700- anti-human-CD8 (BioLegend, 301028), PE-antihuman-FOXP3 (eBioscience, 12-4776-42), PE-antihuman-IFN-γ (BioLegend, 502509). The stained cells were analyzed on a FACS Calibur flow cytometer (BD Bioscience) and data were analyzed using FlowJo10.0 software (Tree Star).

2.8 In vivo studies

PIR-B gene knockout and wild-type C57BL/6 mice (6–8 weeks old) were purchased from Shanghai Model Organisms (Shanghai, China). All animal experiments were approved by the Institutional Animal Care Committee of Shandong First Medical University. The mice were subcutaneously injected with 2 × 10⁵ Lewis lung carcinoma cells (LLC RRID: CVCL_4358) into the right flanks. The LL/2 (LLC1) (RRID: CVCL_4358) is considered identical to the 3LL cell line (RRID: CVCL_5653). Tumor volumes were measured every 3 days using calipers and calculated as 0.5 × length×width². The mice were sacrificed when tumors reached 2 cm in diameter. The tumors were then isolated and weighed. Immune cells were isolated from the peripheral blood, spleen, and tumors for further staining and analysis.

2.9 Statistical analysis

All statistical analyses were performed using SPSS version 26.0 (SPSS, Chicago, USA) and GraphPad Prism 8.0 software (GraphPad Software Inc., La Jolla, CA, USA). The correlation between ILT5 expression and immune cell density/clinicopathological variables was determined using Chi-squared/Fisher's exact test, as appropriate. Survival curves were plotted using the Kaplan–Meier method and analyzed with the log-rank test. Statistical significance was set at p <.05.

3.1 ILT5 is enriched in TAMs of NSCLC tissues, predicting poor patient outcomes

Given that ILT5 is a classical immunosuppressive molecule, we hypothesized that ILT5 is enriched in immune cells. To explore the localization of ILT5 in the TME, the TISCH online tool was used to determine the levels of ILT5 in different immune cell types. These data suggested that ILT5 was most abundant in monocytes and macrophages of the TME (Figure 1A). Subsequently, we determined ILT5 expression in TAMs in our patient cohort of 55 NSCLC patients who had undergone lobectomy/sub-lobectomy and lymph node dissection for lung cancer. ILT5 expression in TAMs was analyzed using mIF. Our findings revealed that ILT5 was enriched in CD68⁺TAMs and was mainly localized in the membrane and cytoplasm (Figure 1B). Further statistical analysis showed that ILT5 expression was considerably higher in TAMs than in adjacent normal tissues (Figure 1C). To determine the significance of ILT5 expression in TAMs, we evaluated the correlation between ILT5 levels and TNM stages. The results indicated that higher ILT5 expression in TAMs was correlated with increased lymph node involvement (Figure 1D) and advanced tumor stages (Figure 1E). Furthermore, we analyzed the correlation between ILT5 levels and patient survival, which indicated that patients with high ILT5 expression exhibited significantly poorer OS than those with low ILT5 expression (Figure 1F). These findings were further confirmed using the GEO database, which yielded consistent results (Figure 1G). Collectively, these results suggest that ILT5 is enriched in the TAMs of NSCLC and indicates advanced disease and poor prognosis.

3.2 ILT5 polarizes TAMs toward the M2-like phenotypes

Given the high ILT5 expression in TAMs and their negative prognostic value, we aimed to determine the role of ILT5 in the recruitment and phenotypic plasticity of TAMs, which are two possible explanations for their immunosuppressive function.²⁶ We determined the frequencies of both the CD68⁺ TAMs and CD163⁺CD68⁺ M2-like subsets. As shown in Figure 2A, B, no difference in the frequency of CD68⁺ macrophages was observed between the ILT5-high and ILT5-low groups. However, the proportion of M2-like TAMs was significantly higher in the ILT5-high group than in the ILT5-low group (Figure 2C). To validate this, we analyzed the relationship between M2-like macrophages and ILT5 levels using the TIMER2.0 database, which demonstrated a positive linear correlation (Figure 2D). These results suggest that ILT5 positively correlates with the M2-like polarization of TAMs. Using in vitro gain-of-function and loss-of-function assays, we validated the direct effects of ILT5 on polarizing TAMs. ILT5 overexpression in THP-1 cells significantly increased the mRNA levels of CD163 and CD206, the two main M2 markers for TAMs (Figure 2E). Moreover, CD163 and CD206 protein levels were also increased following ILT5 overexpression (Figure 2F). On the contrary, ILT5 knockdown in THP-1 cells markedly decreased the mRNA and protein levels of CD163 and CD206 (Figure 2G, H). Taken together, these results indicate that ILT5 in TAMs induces their M2-like phenotypes.

To determine the clinical significance of M2-like TAMs in NSCLC, we analyzed patient OS based on the density of CD163⁺ TAMs. As anticipated, patients with low M2-like TAMs had superior OS than those with high M2 (Figure 3A). Similarly, enriched M2-like TAMs predicted increased lymph node involvement (Figure 3B) and advanced tumor stages (Figure 3C). We combined ILT5 expression with the proportion of M2-like TAMs to evaluate their prognostic value. Patients with high ILT5 expression and enriched M2-like TAMs displayed remarkably poorer OS and a higher probability of regional lymph node metastasis and advanced TNM stages than those of their ILT5-low and M2-low counterparts (Figure 3D–F). Notably, the hazard ratio (HR) for OS was substantially higher using the combined biomarkers (HR = 6.693) than that using ILT5 expression (HR = 5.127) or M2-like TAM proportion (HR = 3.689) alone. These results suggest that TAM-derived ILT5 induces M2-like polarization and is an indicator of poor patient outcomes.

3.3 ILT5 in TAMs reprograms immunosuppressive T cell context

TAMs facilitate tumor progression through various mechanisms, including promotion of neoplastic cell survival, induction of angiogenesis, and mediating of immune escape.²⁷ Because ILT5 induces the M2-like phenotype of TAMs, we clarified whether ILT5 in TAMs affects T cell immunity. We analyzed the number of T cell subsets in the ILT5-low and ILT5-high populations. Our findings revealed that T cell density (CD3⁺) was substantially higher in patients with low ILT5 levels in TAMs than those with high ILT5 expression (Figure 4A, B). Further subset analysis revealed that patients in the ILT5-low group had a higher proportion of CD4⁺ and CD8⁺ T cells but not Tregs (FOXP3⁺), than those in the ILT5-high group (Figure 4A, C–E). Next, we determined the distribution of different T cell subsets based on ILT5 levels and found that the proportion of CD4⁺ and CD8⁺ T cells in total T cells was comparable between the ILT5-low and ILT5-high groups (Figure 4F, G). However, the ratio of Treg cells to total T cells was markedly higher in ILT5-high patients than in ILT5-low patients (Figure 4H). These results suggested that ILT5 within TAMs decreased the number of T cells and increased the proportion of Treg cells in the TME.

Using in vivo TAM-T cell coculture assays, we further explored the underlying mechanisms of T cell subset alteration. THP-1 cells with ILT5 overexpression or knockdown were first induced into TAMs, and then co-cultured with naive T cells. ILT5 overexpression in TAMs inhibited the proliferation of CD3⁺, CD4⁺ and CD8⁺ T cells (Figure 4I) and decreased the secretion of IFN-γ in T cells (Figure 4J). In contrast, ILT5 knockdown in TAMs induced the proliferation of the aforementioned T cell subsets and increased IFN-γ secretion in T cells (Figure 4K, L). Furthermore, we examined the differentiation of T cell subsets upon coculturing with ILT5-modified TAMs. As shown in Figure 4M–O, ILT5 overexpression in TAMs did not alter the proportion of CD4⁺ and CD8⁺ T cells but upregulated that of FOXP3⁺ Treg cells. ILT5 knockdown in TAMs reduced the proportion of Treg cells rather than that of CD4⁺ and CD8⁺ T cells (Figure 4P–R). In summary, TAM-derived ILT5 restricts the proliferation and IFN-γ release of T cells, and induces Treg differentiation.

Additionally, we analyzed the predictive role of ILT5 levels and T cell subset density on patient outcomes. We found that the combination of high ILT5 expression in TAMs with low CD3⁺ (HR = 9.034), CD4⁺ (HR = 7.503), and CD8⁺ (HR = 17.66) T cell expression or high FOXP3⁺ T cell density (HR = 6.872) was powerfully associated with poor OS (Figure S1A–D) and advanced TNM stages (Figure S1E–H). The HR for OS was substantially higher when using combined biomarkers than when using ILT5 expression alone (HR = 2.334). Collectively, these results suggested that ILT5 reprogrammed the proliferation and differentiation of different T cell subsets, leading to an immunosuppressive T cell context and impaired survival.

3.4 PIR-B (ILT5 in mice) orchestrates immunosuppressive TME and promotes tumor growth in vivo

To confirm ILT5-regulated immunosuppression in vivo, we established tumor transplantation models using LLC cell lines derived from PIR-B knockout mice. PIR-B-knockout or wild-type C57BL/6 mice were subcutaneously injected with 2 × 10⁵ LLC cells into the right upper flanks. Tumor size was evaluated every 3 days after tumor inoculation. Mice were sacrificed when the tumor reached a maximum length of 20 mm, and then tumors were isolated and weighed. The results showed that tumor growth was considerably slower in PIR-B knockout mice than in wild-type mice (Figure 5A). Additionally, the final tumor size and weight were significantly decreased in PIR-B-knockout mice than in wild-type mice (Figure 5B, C). Furthermore, we determined the phenotypes and subsets of TAMs and T-cells within the TME. As anticipated, PIR-B knockout did not alter the number of CD11b⁺ monocytes in the spleen, peripheral blood, or F4/80⁺ TAMs in tumors (Figure 5D, E). However, the frequency of M2-like TAMs in the tumors, rather than in the spleen and blood of tumor-bearing mice, was significantly decreased following PIR-B knockout (Figure 5F, G). Next, we explored ILT5-regulated T cell expression and subset distribution within the TME. The results indicated that the proportion of CD3⁺T cells was significantly higher in tumors but not in the spleen or peripheral blood following PIR-B knockout (Figure 5H). Subset analysis revealed that PIR-B knockout increased CD4⁺T cells in the spleen and CD8⁺T cells in tumor tissues (Figure 5I, J), whereas reduced FOXP3⁺Tregs in the tumors (Figure 5K). In addition, the killing ability of tumor-infiltrating T cells, evaluated by the levels of IFN-γ, was enhanced by PIR-B knockout (Figure 5L). Taken together, these results indicated that ILT5 knockout reversed the immunosuppressive TAMs and T cell context, restricting tumor growth in vivo.