CD161, a promising prognostic biomarker in hepatocellular carcinoma, correlates with immune infiltration

Patients and specimens

Clinical data and surgical specimens were retrospectively collected from 109 patients diagnosed with HCC at the Fourth Hospital of Hebei Medical University between November 2017 and January 2023. All patients had undergone curative surgical resection, and the presence of hepatocellular carcinoma was confirmed by postoperative pathology. A range of clinical and pathological indicators were gathered, including age, gender, history of hepatitis, tumor size, tumor number, albumin-bilirubin (ALBI) score, alpha-fetoprotein (AFP) levels, neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), tumor-node-metastasis (TNM) staging, China National Liver Cancer (CNLC) stage (Xie et al., 2023), Barcelona Clinic Liver Cancer (BCLC) staging (Reig et al., 2022), portal vein tumor thrombus (PVTT) occurrence, and microvascular invasion (MVI) status. The study received approval from the ethics committee of the Fourth Hospital of Hebei Medical University (certificate no. 2024KS003), and written informed consent was obtained from all participants.

Immunohistochemistry (IHC)

The expression levels of CD161 in HCC and adjacent normal tissuestissues were assessed by immunohistochemistry analysis. Following standard IHC procedures, formalin-fixed paraffin-embedded (FFPE) samples were blocked with 10% goat serum (cat. no. ZLI-9056; ZSGB-BIO Technologies) for 10 min at 37 °C prior to incubation with primary antibodies against CD161 (1:500, HA601063, HUABIO) at 4 °C overnight. The negative controls were prepared using phosphate-buffered saline instead of the primary antibody. Positive controls were generated according to the manufacturer’s instructions provided with the antibody. A goat anti-rabbit secondary antibody was then applied at room temperature for 20 min. Lastly, hematoxylin was used to counterstain nuclei. The expression levels of CD161 were observed under a light microscope at x200 magnification with five randomly selected fields of view for a separate semi-quantitative analysis, and the average was calculated as the final immunohistochemical score. The immunoreactivity score was blindly evaluated by two senior pathologists. The immunoreactivity score (IRS) of CD161 was calculated using the proportion of positive cells (PP) and the staining intensity (SI), IRS = SI*PP (Zhou et al., 2022). For SI scoring, colorless scored 0 points, pale yellow scored 1 point, brown scored 2 points, and tan scored 3 points. For PP scoring: 0 for negative cells, 1 for positive cells with a 10% chance of being positive, 2 for 11% with a 50% chance of being positive, 3 for 51% with a 100% chance of being positive. The median immunoreactivity score served as the cut-off value to define high and low expression in the samples.

Data acquisition

In this study, the mRNA expression profiles and clinical features of HCC patients (n = 365) with complete survival data were collected from The Cancer Genome Atlas (TCGA) database (https://cancergenome.nih.gov/). The processed microarray data of 221 HCC samples from GSE14520 were retrieved from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) to confirm the relationship between CD161 expression and clinicopathological features and HCC survival. To confirm the biological function of CD161 expression in HCC, another dataset (ICGC-LIRI-JP) including the mRNA expression profiles of 230 patients was obtained from ICGC (https://dcc.icgc.org/; The ICGC Data Portal officially closed in June 2024. The related data is now available at https://www.icgc-argo.org). The TCGA, ICGC, and GSE14520 contain both core tumor tissue and adjacent tissue. Both cancer tissues and adjacent normal tissues were used to examine the differential expression of CD161 and its diagnostic significant for HCC. All other analyses used exclusively core cancer tissue. The mRNA sequencing data of the TCGA-LIHC and ICGC-LIRI-JP cohorts were converted into transcripts per kilobase million (TPM) values and were then log2(TPM+1) transformed, which has been recommended as the most accurate quantification method with minimal statistical biases (Jin, Wan & Liu, 2017). An immunotherapy cohort (GSE140901) containing 24 HCC patients was acquired from the GEO database. The IMvigor210 cohort (Mariathasan et al., 2018), a study of atezolizumab in patients with locally advanced or metastatic urothelial carcinoma, was downloaded from http://research-pub.gene.com/IMvigor210CoreBiologies. The protein levels of CD161 in human normal liver tissue and HCC were examined using the Human Protein Atlas (HPA: https://www.proteinatlas.org/) database.

Prognostic analysis

Kaplan–Meier analyses were performed to evaluate the overall survival and recurrence-free survival in both TCGA and GSE14520 datasets. Furthermore, the prognostic significance of various parameters was assessed using univariate and multivariate Cox proportional hazards models.

Functional enrichment analysis

The genes most relevant to CD161 were uploaded to the Database for Annotation, Visualization, and Integrated Discovery (DAVID). The enrichment results of Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were retrieved. The figures displayed the top six results in ascending order of P-value (P < 0.05).

Gene set variation analysis (GSVA)

The gene list of immune processes was downloaded from the Gene Set Enrichment Analyses database (GSEA, https://www.gsea-msigdb.org/gsea/index.jsp). An enrichment score was calculated for each HCC sample using the provided package in the R environment. A heatmap was generated based on the enrichment results and Pearson’s correlation coefficients were computed.

Immune cell infiltration analysis

The immune cell infiltration score in both TCGA and ICGC databases underwent xCell deconvolution analysis, CIBERSORTx (https://cibersortx.stanford.edu/; Newman et al., 2015), and MCP-counter analysis. The levels of immune cell infiltration were compared between different groups, which were stratified based on the median CD161 expression level in the TCGA and ICGC databases. To further elucidate the relationship between CD161 and immunophenotyping, data was used from a published work (Thorsson et al., 2018), which identified six immune subtypes (Wound Healing, IFN-γ Dominant, Inflammatory, Lymphocyte Depleted, Immunologically Quiet, and TGF-β Dominant) based on the TCGA database. The leukocyte fraction, lymphocyte infiltration signature score, B-cell receptor (BCR) Shannon, and T-cell receptor (TCR) Shannon were compared across different CD161 groups.

Statistical analyses

All statistical analyses and visualization were performed using SPSS 25.0 (IBM Corp), R software (version 4.2.1), Python (version 3.9) and Prism 10.0 (GraphPad, San Diego, CA, USA). Unpaired Student t-tests or Mann–Whitney U tests (also known as the Wilcoxon rank-sum test) were used to compare the differences between the two groups. One-way ANOVA or Kruskal–Wallis tests were used for comparisons of more than two groups. The associations between CD161 expression and the patients’ clinicopathological characteristics and treatment efficacy were evaluated by the chi-square test or Fisher’s exact test. The receiver operating characteristic (ROC) curve was generated using the pROC package to evaluate the diagnostic capability of CD161. The significance of the correlation between the two groups was assessed by Pearson correlation analysis. The prognostic value was evaluated using Kaplan Meier curves, and the significance of the prognostic value was tested by a log-rank test. A P-value of < 0.05 was deemed statistically significant.

Patient characteristics

A total of 109 patients who underwent liver resection for HCC were included in this study, with their details summarized in Table 1. The median age at diagnosis was 59 years (range: 32–86 years). Most patients were male (81.7%), had a history of Hepatitis B/C (90.8%), had a solitary tumor (81.7%), and possessed a tumor size of ≤ 5 cm (60.6%). The majority of individuals in this patient cohort were classified as TNM stages I–II (80.7%) and had an ALBI grade of 1 (76.1%). As of the cut-off date of May 1, 2024, two patients were lost to follow-up, and three patients died from other causes. Ultimately, a total of 104 patients completed the follow-up with a median follow-up duration of 35.4 months.

Expression of CD161 is lower in HCC tissues

The mRNA levels of CD161 were assessed in paired tumor and normal liver tissues across TCGA, ICGC, and GSE14520 datasets. Low CD161 expression was observed in the tumor compared to normal liver tissue in the three databases (Figs. 1A–1C). To validate these findings at the protein level, a total of 109 liver cancer tissues from 109 HCC patients, along with matched adjacent normal tissues from 53 of these patients, were collected for the assessment of CD161 protein expression through immunohistochemistry. IHC analysis of HCC tissues showed that CD161 was mainly expressed in the cytoplasm and cell membrane, consistent with a previous study (Xu et al., 2023). Among the 109 liver cancer tissues in this cohort, 103 (94.5%) exhibited positive expression of CD161. Additionally, positive CD161 expression was observed in all 53 adjacent normal tissues. Representative IHC staining images displayed different levels of CD161 expression in both tumor and adjacent normal tissues (Figs. 2A–2J). The Human Protein Atlas database (http://www.proteinatlas.org/pathology) was also used to investigate protein levels of CD161 in HCC (Fig. 2K). The immunohistochemical results showed that CD161 was significantly upregulated in 53 adjacent non-cancerous tissues compared to 109 liver cancer tissues (Fig. 1E). Additionally, in the 53 matched pairs of HCC and adjacent normal tissues, CD161 was downregulated in cancerous tissues compared to adjacent normal tissues (Fig. 1D). Furthermore, the diagnostic capability of CD161 was evaluated by analyzing the ROC curve. The area under the curve (AUC) values were 0.658, 0.689, 0.792, and 0.808 for TCGA, ICGC, GSE14520, and the study cohort, respectively, suggesting a robust diagnostic potential for CD161 in HCC (Figs. 1F–1I).

Expression of CD161 is associated with tumor size and PVTT

The associations between CD161 expression and clinicopathological characteristics were also investigated. Using a median immunohistochemical score of 5.6 as the threshold, 62 patients (56.9%) were categorized as exhibiting high CD161 expression, while the remaining 47 cases (43.1%) were classified as having low CD161 expression. As shown in Table 1, CD161 expression was positively correlated with smaller tumor sizes (≤5 cm) and the absence of portal vein tumor thrombus. The relationship between CD161 expression and clinical pathological features was further confirmed in the TCGA and GSE14520 datasets (detailed information can be found in Table S1). Higher levels of CD161 expression were typically observed in patients with smaller tumors and was linked to absence of PVTT in HCC, as indicated by the results from analyzing multiple public and in-house cohorts. Furthermore, higher levels of CD161 expression were also observed in surviving patients in comparison to deceased patients (Fig. 3).

CD161 is closely linked to a favorable prognosis in HCC patients

CD161’s prognostic significance was investigated in our cohort as well as in the TCGA and GSE14520 databases. In the TCGA and GSE14520 databases, the median value of CD161 expression was selected as the cutoff value to categorize HCC patients into CD161 high-expression and low-expression groups within each database. Based on a Kaplan–Meier analysis, the CD161 high-expression groups had longer OS and RFS compared to the low-expression groups (Fig. 4), although RFS analysis in the GSE14520 dataset did not reach statistical significance. Overall, these findings suggested that CD161 conferred a protective effect on HCC patients. To further elucidate the role of CD161 in the prognosis of HCC patients, univariable and multivariable Cox analyses were performed. Univariate analysis of the Cox proportional hazards model showed that the factors significantly associated with OS and RFS were tumor size, tumor number, PVTT, TNM stage, CNLC stage, AFP level, and CD161 expression. The clinicopathological parameters that were screened out in the univariate analysis (p < 0.05) were then included in the multivariate survival analysis. The results indicated that TNM stage, AFP level, and CD161 expression were independent prognostic factors for predicting RFS and OS (Table 2).

CD161 is involved in the immune response and immune regulation in HCC

A Pearson correlation analysis (|R| > 0.5, P < 0.05) was used to identify the genes most related to CD161 in both the TCGA and ICGC databases. GO and KEGG analyses were conducted based on these gene sets. In the TCGA database, the biological processes most closely linked to CD161 included immune response, adaptive immune response, T cell activation, and the T cell receptor signaling pathway (Fig. 5A). Additionally, the cellular components most associated with CD161 were the plasma membrane and external side of the plasma membrane (Fig. 5B). The molecular function most associated with CD161 was transmembrane signaling receptor activity (Fig. 5C). The signaling pathways most closely associated with CD161 included cytokine-cytokine receptor interaction and the chemokine signaling pathway (Fig. 5D). The biological functions, cellular components, molecular processes, and signaling pathways most associated with CD161 in the ICGC database mirrored those in the TCGA database (Figs. 5E–5H).

Based on the results of the GO and KEGG analyses, CD161 was considered to be crucial for immune response, T cell activation, cytokine-cytokine receptor interaction, and chemokine signaling. Therefore, an in-depth investigation into the effects of CD161 on these cellular functions and signaling pathways was conducted. Gene set variation analysis (GSVA) was used to determine the enrichment score of the immune process in both the TCGA and ICGC databases. The correlation analysis revealed a positive association between CD161 expression and the majority of immune responses, as well as between CD161 expression and functions and signaling pathways related to cytokines and chemokines (Fig. 6A). Subsequent validation in the ICGC database confirmed these results (Fig. 6B). These findings suggested that CD161 likely plays a pivotal role in immune response and immune regulation in HCC.

CD161 is related to an inflamed tumor microenvironment in HCC

Considering that CD161 was associated with multiple immune-related processes, the specific immunological role of CD161 in HCC was investigated. The median value of CD161 expression from the TCGA and ICGC databases was selected as the threshold to categorize CD161 expression into high and low groups. A majority of chemokines and receptors (Fig. 7A), cytokines and receptors (Fig. 7B), major histocompatibility complex (MHC) molecules (Fig. 8A), and co-stimulators (Fig. 8B) were notably correlated with CD161 in HCC. To explore the association between CD161 and immunophenotyping, CD161 expression was compared across different immune subtypes based on previously published research (Thorsson et al., 2018). The findings indicated that CD161 expression peaked in the IFN-γ dominant subtype, followed by the inflammatory subtype, and was lowest in the lymphocyte depleted subtype (Fig. 9A). Subsequently, comparisons were made among different CD161 groups regarding the leukocyte fraction, lymphocyte infiltration signature score, BCR Shannon score, and TCR Shannon score. Higher values were observed for all of these metrics in the group with high CD161 expression (Figs. 9B–9D). Taken together, CD161 is positively associated with an inflamed tumor microenvironment in HCC.

CD161 is associated with high immune cell infiltration

To explore the relationship between CD161 expression and immune cell infiltration, a correlation analysis was conducted using xCell, CIBERSORT, and the MCP-counter method in both the TCGA and ICGC databases. A positive association between CD161 expression and the infiltration levels of most immune cells, especially DC cells, CD8+ T cells, and CD4+ T cells, was revealed by xCell in the TCGA database (Fig. 10A). The xCell analysis also found that the CD161 high-expression group exhibited significantly higher immune scores, stroma score, and microenvironment score (Figs. 10B–10C) in the TCGA database. The X cell analysis in the ICGC database produced consistent results with those from the TCGA database (Fig. 11). Similarly, CIBERSORT (Figs. 12A–12B) and MCP-counter (Figs. 12C–12D) analyses also revealed a positive correlation between CD161 expression and the infiltration levels of DC cells, CD8+ T cells, and CD4+ T cells. These findings suggest that CD161 is related to high immune cell infiltration in HCC. This further indicates the positive correlation between CD161 and the inflammatory immune microenvironment, suggesting its potential as a biomarker for immunotherapy.

CD161 predicts immunotherapeutic responses

The above findings demonstrated that CD161 expression was positively associated with an inflamed TIME and immune cell infiltration, which raises the question of its possible use as a biomarker to predict the efficacy of immunotherapy. The relationship between CD161 expression and previously published immune-related signatures (Spranger, Bao & Gajewski, 2015; Danilova et al., 2016; Ayers et al., 2017; Hsu et al., 2021), potentially aiding in predicting responses to immunotherapy across different cancers, was analyzed using GSVA. The findings revealed a positive correlation between the expression of CD161 and these predictive genetic signatures (Figs. 13A–13B). Furthermore, a robust correlation was observed between CD161 and prominent inhibitory immune checkpoints including PD-1, PD-L1, TIM3, LAG3, CTLA-4, TIGIT, and CD200R1 across the TCGA and ICGC databases (Figs. 14A–14B). Subsequent exploration was conducted to examine the relationship between CD161 expression and treatment response in an immunotherapy cohort (GSE140901) comprised of 24 HCC patients. The CD161 high-expression group had a higher proportion of patients evaluated as partial response (PR) and stable disease (SD), while the low-expression group had a higher proportion of patients assessed as progressive disease (PD) (Fig. 14C). However, there was no significant statistical difference in the objective response rates (ORR) and disease control rates (DCR) in the GSE140901 set (Table S2). In addition, the association between CD161 expression and the efficacy of immunotherapy was investigated in the IMvigor 210 cohort. CD161 expression peaked in the inflamed subtype and was lowest in the desert subtype within the IMvigor 210 cohort (Fig. 14D). Furthermore, CD161 expression exhibited a positive correlation with the level of IHC-assessed PD-L1 staining on immune cells (IC; Fig. 14E). In the IMvigor 210 cohort, the CD161 high-expression group had a larger percentage of patients with complete response (CR), partial response, and stable disease, and the CD161 low-expression group had a higher percentage of patients experiencing progressive disease (Fig. 14F). Additionally, the CD161 high-expression group exhibited superior objective response rates and disease control rates compared to the CD161 low-expression group, despite the lack of statistical significance in ORR (Table 3).