Skip to main content
Download PDF

High ratio of resident to exhausted CD4 + T cells predicts favorable prognosis and potentially better immunotherapeutic efficacy in hepatocellular carcinoma

BMC Cancer volume 24, Article number: 1152 (2024) Cite this article

Abstract

Background

Tumor-infiltrating lymphocytes (TILs) are significantly implicated in regulating the tumor immune microenvironment (TIME) and immunotherapeutic response. However, little is known about the impact of the resident and exhausted status of TILs in hepatocellular carcinoma (HCC).

Methods

Single-cell RNA sequencing data was applied to discover resident and exhausted signatures of TILs. Survival outcomes, biological function, immune infiltration, genomic variation, immunotherapeutic efficacy, and sorafenib response were further explored the clinical significance and molecular association of TILs in HCC. Moreover, a candidate gene with predictive capability for the dismal subtype was identified through univariate Cox regression analysis, survival analysis, and the BEST website.

Results

Single-cell analysis revealed that CD8 + T, CD4 + T, and NK cells were strongly associated with resident and exhausted patterns. Specific resident and exhausted signatures for each subpopulation were extracted in HCC. Further multivariate Cox analysis revealed that the ratio of resident to exhausted CD4 + T cells in TIME was an independent prognostic factor. After incorporating tumor purity with the ratio of resident to exhausted CD4 + T cells, we stratified HCC patients into three subtypes and found that (i) CD4 residencyhighexhaustionlow subtype was endowed with favorable prognosis, immune activation, and sensitivity to immunotherapy; (ii) CD4 exhaustionhighresidencylow subtype was characterized by genome instability and sensitivity to sorafenib; (iii) Immune-desert subtype was associated with malignant-related pathways and poor prognosis. Furthermore, spindle assembly abnormal protein 6 homolog (SASS6) was identified as a key gene, which accurately predicted the immune-desert subtype. Prognostic analysis as well as in vitro and in vivo experiments further demonstrated that SASS6 was closely associated with tumor prognosis, proliferation, and migration.

Conclusions

The ratio of resident to exhausted CD4 + T cells shows promise as a potential biomarker for HCC prognosis and immunotherapy response and SASS6 may serve as a biomarker and therapeutic target for prognostic assessment of HCC.

Peer Review reports

Introduction

Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related deaths worldwide [1]. Clinical routine therapies, such as surgical resection, radiotherapy, chemotherapy, and multi-kinase inhibitors, have failed to significantly improve clinical outcomes of HCC patients [2, 3]. Recently, immunotherapy represented by immune checkpoint inhibitors (ICIs) has revolutionized the treatment of solid tumors, but only 25% of patients have achieved durable benefits [4]. More recently, durvalumab plus tremelimumab yielded superior overall survival (OS) versus sorafenib and atezolizumab plus cabozantinib yielded superior progression-free survival [5]. Unfortunately, no robust biomarker is available to help stratify patients and guide clinical decision-making. Therefore, novel biomarkers devoted to predicting prognosis and immunotherapeutic efficacy are urgent in HCC.

The tumor immune microenvironment (TIME) profoundly impacts the development and progression of tumors by balancing between suppressive and cytotoxic responses [6]. Tumor-infiltrating lymphocytes (TILs) are the primary components of TIME and have diverse functions, such as killing tumor cells and secreting cytokines [7]. Indeed, the characteristics and infiltration fraction of TILs are the key factors to determine immunotherapeutic response [8, 9]. Recently, the composition, differentiation status, and exhausted pattern of TILs have been abundantly explored [10,11,12,13]. Previous studies have revealed that triple-negative breast cancer patients carrying higher levels of tissue-resident memory T (TRM) cells have significantly improved prognosis and increased response rates to anti-PD-1 antibodies [10]. In patients with non-small cell lung cancer, TRM cells promote cytotoxic responses within tumors and support the rationale used to reverse tumor-induced T cell exhaustion in patients using anti-PD-1 antibodies [11]. Conversely, TILs can predominantly undergo distinct differentiation patterns to acquire tissue exhausted characteristics, potentially impacting immunotherapeutic responses [12,13,14]. Foroutan M et al. found that the ratio of exhausted to resident TILs affected the prognosis of colorectal cancer [12], and our previous study also confirmed that the ratio of resident to exhausted CD4 + T cells is associated with the prognosis of gastric cancer [14]. These findings have prompted efforts to explore the impact of resident and exhausted TILs on tumor prognosis as well as their role and potential mechanisms in immunotherapy to effectively improve efficacy. However, the role of the resident and exhausted patterns of TILs in HCC prognosis and response to immunotherapy remains to be further explored.

In this work, we identify specific gene signatures for resident and exhausted patterns in CD8 + T, CD4 + T, and NK cells by single-cell RNA sequencing (scRNA-seq). Moreover, only the ratio of resident to exhausted CD4 + T cells was an independent prognostic factor. After incorporating tumor purity with the ratio of resident to exhausted CD4 + T cells, HCC patients were stratified into three subtypes endowed with distinct prognosis, biological function, immune infiltration, immunotherapeutic efficacy, and sorafenib sensitivity. Moreover, receiver operating characteristic (ROC) curves, univariate Cox regression analysis and survival analysis were employed to identify candidate biomarkers which could accurately predict the worst prognosis subtype. Spindle assembly abnormal protein 6 homolog (SASS6) possessed robust and accurate ability in HCC prognosis and the high expression of SASS6 predicted worse OS. Overall, our study deepened the understanding of the characteristics of TILs in TIME and identified a promising biomarker for evaluating prognosis and potential response to immunotherapy in HCC.

Materials and methods

Data acquisition

Single-cell RNA sequencing data

The 10X and Smart-seq2 single-cell transcriptome data from GSE140228 (including 73,261 cells) were downloaded from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) [15]. In addition, we additionally recruited two scRNA-seq datasets, GSE151530 and GSE166635, from previous literature to verify the robustness and reproducibility of our results [16, 17].

Cancer cell lines

Transcriptomic data for cancer cell lines were generated from the cancer cell line encyclopedia (CCLE, https://sites.broadinstitute.org/ccle) [18].

Multi-omics data for TCGA- LIHC bulk cohort

The RNA-sequencing data, clinical information, and mutational data of TCGA-LIHC were retrieved from the UCSC Xena (http://xena.ucsc.edu/).

Treatment cohorts

Three hundred nineteen patients from seven immunotherapeutic cohorts, including GSE100797 (melanoma), GSE111636 (urothelial tumors), GSE135222 (non-small cell lung carcinoma), GSE136961 (non-small cell lung carcinoma), GSE35640 (melanoma), GSE91061(melanoma), and GSE93157 (lung carcinoma, head and neck squamous cell carcinoma and melanoma), were collected to assess the immunotherapeutic response. The information of 67 HCC patients accepting sorafenib was downloaded from GSE109211 to predict the sorafenib efficacy.

Quality control and cell type identification

The single cell quality control process was performed with the Seurat R package. After filtering out doublets and eligible cells, cells with the following criteria were retained: (i) 200 to 5000 genes were detected; (ii) the total unique molecular identifiers (UMIs) number ≤ 40,000; (iii) UMIs arising from mitochondrial percentage ≤ 15%. The top 2000 variable genes were selected for principal component analysis. The top 30 principal components were utilized for Uniform Manifold Approximation and Projection (UMAP) analysis. The UMAP scatter plots were drawn to validate the accuracy of cell annotations by canonical genes: CD8A, MS4A1, LYZ, GNLY, CD3E, and FCER1A. Resident and exhausted markers, including CD69, ITGAE (CD103), ITGA1, PDCD1 (PD-1), CTLA4, and LAG3, were selected to identify TIL subpopulations influenced by resident and exhausted patterns [12].

Obtaining initial and final resident and exhausted signatures

  1. 1

    Only cells from tumor tissue of 10X scRNA-seq data were retained to reflect the TIME. Additionally, an extensive list of resident and exhausted genes was extracted based on previous reports [12, 19,20,21] (Additional file 1: Table S1).

  2. 2

    Removing canonical resident and exhausted markers that showed a significant negative correlation with each other [12]. Subsequently, we identified canonical resident markers, including CD69, ITGAE, RGS1, and CXCR6 for NK cells, the same genes added ITGA1 to CD8 + T cells, and resident marker ZNF683 (Hobit) was included in CD4 + T cells compared with CD8 + T cells. Meanwhile, canonical exhausted markers for CD8 + T, CD4 + T, and NK cells consisted of HAVCR2 (Tim-3), PDCD1, CTLA4, LAYN, CXCL13, and LAG3.

  3. 3

    Calculated correlation between resident/exhausted genes in comprehensive gene lists described in step 1 and canonical resident/exhausted markers. Genes with the correlation coefficient > 0.25 were retained for CD8 + and CD4 + T cells. But for NK exhausted genes, the threshold was set to 0.2 due to the differences in gene expression distribution.

  4. 4

    Deleted the duplicated or highly correlated genes between resident and exhausted gene lists. Subsequently, the sparse canonical correlation analysis was performed to further reduce the correlation between resident and exhausted gene sets by removing genes with a correlation coefficient of the first component ≤ -0.1. Finally, we obtained the initial resident and exhausted signatures for CD8 + T, CD4 + T, and NK cells (Additional file 1: Table S2).

  5. 5

    Based on the initial resident and exhausted signatures, the singscore package was used to calculate gene-set enrichment scores at the single cell level [22]. For CD8 + T, CD4 + T, and NK cells, exhausted cells owned the characteristics with exhausted score above the 90th percentile and resident score below the 60th percentile, and a similar threshold to resident cells. However, due to the peculiarities of the distribution of gene expression, the top threshold for NK cells was established at the 85th percentile.

  6. 6

    Performed differential expression analysis between resident and exhausted cells by the Wilcoxon rank-sum test and the hurdle model. Genes with adjusted p-value < 0.05 and log2FC > 0.3 were collected.

  7. 7

    Refined the signatures by comparing transcript abundance percentile for resident vs. exhausted cells. Trajectory inference with Slingshot was performed based on the final gene markers [23].

Validation of single-cell data

The efficacy of the resident and exhausted signatures was validated using Smart-seq2 data from GSE140228. Similar to the previous quality control procedure, the final resident and exhausted signatures were selected to score for CD8 + T, CD4 + T, and NK cells, respectively. Subsequently, resident and exhausted cells were identified according to predetermined thresholds for each cell type to validate the stability of the resident and exhausted signatures.

Refining genes to pass “bulk threshold”

For HCC cell lines data, only genes with counts per million > 1 in more than five cell lines were filtered. Then, we calculated scaling factors to convert the raw library into an effective library and further performed reads per kilobase per million mapped reads transformation. Next, resident and exhausted signatures less than the 75th percentile of the value expressed across all genes in HCC cell lines were filtered. Subsequently, we correlated the scoring information of three cell types with the expression information of overlapping genes in TCGA and reassigned the overlapping genes according to the correlation. Ultimately, the selected genes were considered to pass the “bulk threshold” and were suitable to be analyzed in bulk data.

Subtyping and enrichment analysis

The ESTIMATE package was used to predict the tumor purity of each patient in TCGA [24], and the resident and exhausted scores were calculated by Single-Sample Gene Set Enrichment Analysis (ssGSEA). Subsequently, we stratified patients with tumor purity > 90% as the immune-desert subtype in TCGA-LIHC and the remaining patients as the high-scoring and low-scoring subtypes according to the mean value of resident, exhausted, and ratio scores among CD4 + T, CD8 + T, and NK cells. Gene Set Enrichment Analysis (GSEA) was used to perform Hallmark, Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG), and further identified specific functional pathways for three subtypes, respectively.

Immune cell infiltration assessment

Using ssGSEA, the infiltration abundance of 28 immune cells was measured in HCC bulk tissues. CIBERSORT was conducted to further validate our findings. Co-stimulatory and co-inhibitory molecules were extracted from a previous report, and the relative differences were compared among different subtypes [25]. Additionally, the immune subtypes, leukocyte fraction, lymphocyte infiltration signature scores, T cell receptor (TCR), and cancer immunity cycle (CIC) were compared among different subtypes [15, 26].

Genomic variation

The maftools package was used to calculate the top 20 hypervariable genes among three subtypes. The recurrent amplified and deleted genome regions information for TCGA-LIHC was downloaded from the Firebrowse database (http://firebrowse.org/). We focused on the regions of amplified copy number alternation (CNA) > 15% or deleted CNA > 20%. After correction for differences in the number of patients among three subtypes, the proportion of hypervariable genes and CNA in each subtype was calculated. We finally calculated the fraction genome alteration (FGA), fraction genome gain (FGG), fraction genome loss (FGL), arm gain, arm loss, focal gain, and focal loss of each sample.

Immunotherapeutic assessment

Patients with immunotherapeutic information were stratified into three subtypes based on resident and exhausted scores of CD4 + T cells. We assessed percentile and the area under the ROC curve (AUC) differences in immunotherapeutic responses among different subtypes. Subclass mapping (Submap) was used to compare the expression similarity between three subtypes and different immunotherapeutic responders/nonresponders [27]. However, since GSE136961 and GSE93157 contained only a few genes for ssGSEA scoring, these two datasets were only used to perform Submap analysis, but not for the calculation of the differences in immunotherapy responses.

Identification of prognostic hub gene

To elucidate the fundamental prognostic mechanisms of the unfavorable subtype, screening strategies for essential biomarkers were examined from a genetic standpoint.

Firstly, AUC was calculated by the pROC package to identify genes with predictive capability for the dismal subtype. Then, genes with AUC > 0.7 were selected for univariate Cox regression analysis, employing a threshold of hazard ratio > 1 and P < 0.05. Thirdly, genes whose expression associated with tumor purity greater than 0.4 were preserved. Subsequently, Kaplan–Meier survival analysis and BEST website (https://rookieutopia.com/app_direct/BEST/) were further utilized to validate these candidate genes obtained by the above methods. Ultimately, SASS6 was selected to be the diagnostic gene for the worst prognosis, despite its extensive study in other cancers but lack of exploration in HCC.

Cell line and cell transfection

HCC cell line (MHCC-97H) was used in our research, which was cultured in Dulbecco’s Modified Eagle’s medium (DMEM)/High-Glucose (Solarbio, Beijing, China) containing 10% fetal bovine serum (Cyagen, Guangzhou, China) at 37 °C with 5% CO2. Negative control (NC) and SASS6 small interfering RNA (siRNA) (RiboBio, Guangzhou, China) were transfected into MHCC-97H cells according to the manufacturer’s instruction. Quantitative real-time polymerase chain reaction (qRT-PCR) was employed to confirm the transfection efficiency. The expression value of SASS6 was normalized to GAPDH and then log2 transformation for subsequent analysis. The 2^-ΔΔCT approach was applied to compute the relative RNA expression.

Cell counting kit-8 (CCK-8) assay

Transfected MHCC-97H cells were seeded in 96-well plates and cultured under suitable conditions for 24 h. Once the cells were adherent, the CCK-8 assay solution (US Everbright, Suzhou, China) was added to each well. After 2 h of additional incubation, the optical density (OD) value at 450 nm was measured with a microplate reader (Infinite F50, Tecan, Switzerland).

5-Ethynyl-2'-deoxyuridine (EdU) incorporation assay

The assessment of cell proliferation capacity was conducted using the EdU incorporation assay. Prior to the assay, MHCC-97H cells were pretreated and subsequently exposed to 50 μm EdU for 2 h at 37 °C. Next, the cells were fixed using a 4% formaldehyde solution for 30 min, and then permeabilized using 0.5% Triton X-100 for 10 min. Subsequently, each well was treated with 1 × Apollo solution for 30 min, followed by staining of the nuclei with 1 × Hoechst 33,342 for an additional 30 min. Finally, the EdU-positive cells (red) and Hoechst-positive cells (blue) were observed and quantified using a fluorescence microscope.

Scratch assay

The NC and SASS6 siRNA cells were cultured in 6-well plates and incubated at 37 °C overnight in a 5% CO2 incubator. The medium was then removed, and the surface of the cells was gently scratched and marked using a 200 uL pipette tip. Subsequently, the cells were washed. Photographs of the scratches were taken at 0 h and 48 h, respectively. This experiment was repeated three times. The distance of cell migration towards the injured area was measured during the specified time period.

Transwell migration assay

Transwell chambers were employed for the assessment of cellular migration capability. Approximately, 2.5 × 104 cells transiently transfected cells were cultured in the upper chamber using serum-free medium, while a complete medium was added to the lower chamber. The cells were incubated at 37 °C for 48 h, subsequently washed with physiological saline, and fixed with methanol. After that, the cells were stained and subsequently photographed under a microscope, with the stained cells being quantified.

Animal experiment

For subcutaneous xenograft model, six-week-old male BALB/c nude mice were obtained from Vital River Laboratory (Beijing, China) and kept under pathogen-free conditions. Cells in the logarithmic growth phase were harvested and resuspended in PBS. A total of 1 × 10^7 cells (either shCtrl-MHCC-97H or shSASS6-MHCC-97H cells) were injected subcutaneously into the right flank of each mouse. Tumor dimensions were measured using calipers, and tumor volume was calculated using the formula: V = L × W^2 × 0.5236 (where L is the long axis and W is the short axis). After 20 days, the mice were anesthetized by intraperitoneal injection of sodium pentobarbital (50 mg/kg) and subsequently euthanized by cervical dislocation, and the tumors were harvested and weighed. All animal experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee of Zhengzhou University.

Statistical analysis

R software (v-4.2.0) was used for all data processing and statistical analysis. Pearson’s correlation analysis was applied to calculate the correlations between two continuous variables. Categorical variables comparison was performed by the chi-squared or fisher exact test, and continuous variables by the T-test or Wilcoxon rank-sum test. Multiple subtypes were compared based on the ANOVA or Kruskal-Walli’s test. Differential expression analysis between resident and exhausted cells by the Wilcoxon rank-sum test and the hurdle model. Based on the survival package, Kaplan–Meier analysis was performed. GSEA analysis was performed by the clusterProfiler package. The ROC curve was implemented through the pROC package. The Slingshot package conducted the trajectory analysis. The ESTIMATE package estimated the tumor purity. All statistical tests were two-sided P < 0.05 was regarded as statistically significant.

Results

CD8 + T, CD4 + T, and NK cells strongly correlated with resident and exhausted patterns

The workflow of our study is shown in Additional file 2: Figure S1. Cell clustering and annotation were performed according to the standard Seurat pipeline. We found that TILs from different patients and tissues were mixed (Additional file 2: Figure S2A) and mainly divided into five major subpopulations: T cells, NK cells, Myeloid, B cells, and B plasma (Fig. 1A and B). To validate the accuracy of cell annotation, canonical cell markers including CD8A, MS4A1, LYZ, GNLY, CD3E, FCER1A, and so on were utilized to delineate expression profiles, which were highly consistent with specific cell regions (Fig. 1C and Additional file 2: Figure S2B). Furthermore, we gathered the signature gene sets of CD4 + T cells, CD8 + T cells, and NK cells, and verified our clustering accuracy by evaluating scores for each cell subset using the AddModuleScore function (Additional file 2: Figure S3 and Additional file 1: Table S3). Subsequently, canonical resident and exhausted markers were selected to identify which subpopulations were significantly featured by resident and exhausted patterns. Interestingly, these markers were dramatically concentrated in CD8 + T, CD4 + T, and NK cell regions (Fig. 1D and E).

Fig. 1
figure 1

CD8 + T, CD4 + T, and NK cells expressed high levels of resident and exhausted markers. A-B Cell clustering and annotation based on the meta-data (A) and canonical cell markers (B). C Expression level of six canonical genes. The color of the dots represented the expression level of canonical genes. D-E Expression level of canonical resident (D) and exhausted (E) signatures based on UMAP

Full size image

Obtaining and verifying resident and exhausted signatures

To identify specific resident and exhausted signatures in HCC, we introduced a seven-step pipeline (see Methods). For each cell subpopulation, we first retained genes highly correlated with canonical resident and exhausted markers. Due to the crosstalk of resident and exhausted patterns, genes were further refined by removing overlapping and significantly related genes between resident and exhausted gene sets (Additional file 1: Table S2) [28]. Subsequently, we calculated gene-set enrichment scores to designate resident and exhausted cells (Additional file 2: Figure S4A). Spatial analysis further demonstrated that resident and exhausted cells were well separated (Additional file 2: Figure S4B).

Genes were extracted based on differential expression analysis between resident and exhausted cells to accurately identify the resident and exhausted signatures of each cell subpopulation. Meanwhile, the comparison between resident and exhausted percentile thresholds further refined signatures. Using final gene signatures, single cells were rescored and recategorized. We then observed that resident and exhausted cells were robustly separated (Fig. 2A). Further analysis elaborated the evolutionary trajectory from residency to exhaustion to distinguish between resident and exhausted patterns (Additional file 2: Figure S4C).

Fig. 2
figure 2

Obtaining and verifying resident and exhausted signatures. A Resident and exhausted cells based on final resident and exhausted signatures in the CD8 + T, CD4 + T, and NK cells. Blue represented resident cells, and red represented exhausted cells. B Selecting resident and exhausted signatures that passed “bulk threshold”. Resident and exhausted signatures were annotated based on whether they passed the “bulk threshold”. C-D Resident and exhausted signatures were represented after reassigning duplicated genes. Red represented present signatures, and blue represented absent signatures. E Cell clustering and annotation plot according to canonical cell markers in validation data

Full size image

As previously reported, resident and exhausted signatures are highly expressed in TILs but not tumor cells [12]. Thus, resident and exhausted signatures passing the “bulk threshold” were selected based on the percentile abundance comparison between tumor cells and TILs (Fig. 2B). To further reduce the crosstalk among three cell types, overlapping genes were reassigned based on the correlation between signature scores and gene expression for three cell types (Fig. 2C and D, Additional file 1: Table S4). Ultimately, Smart-seq2 data was utilized to confirm the stability of results from 10X scRNA-seq data (Fig. 2E and Additional file 2: Figure S4D). In addition, in the GSE151530 and GSE166635 datasets, after quality control, major cluster annotation, and T cell cluster annotation, we extracted CD4 + T cells and scored them with the final resident and exhausted signatures. As expected, the resident and exhausted CD4 + T cell subsets were significantly separated from each other, again suggesting the robustness and reproducibility of our results (Additional file 2: Figure S5A-I and S6A-C). Notably, given the vital role of tumor-infiltrating B cells (TIBs) in anti-tumor immunity and immunotherapy, as well as their impact on patient prognosis [29, 30]. We included B cells in our seven-step pipeline, the results exhibited that conventional resident and exhausted signatures do not adequately define the B cell status in HCC (Additional file 2: Figure S6D-F).

Prognostic significance of the ratio of resident to exhausted CD4 + T cells

We observed a significant association between resident and exhausted patterns in TCGA-LIHC (Fig. 3A). In bulk tissues, tumor purity is a crucial factor for quantifying the components of the tumor microenvironment [31, 32]. In this study, correlation analysis showed a significant decrease in tumor purity with increasing resident and exhausted scores in HCC patients (Fig. 3A), suggesting tumor purity might be a confounding factor for investigating TIME. Thus, tumor purity, resident scores, exhausted scores, and the ratio of resident to exhausted scores in three cell types were combined to form nine subtyping methods. For each method, patients were stratified into three subtypes, including residencyhighexhaustionlow, exhaustionhighresidencylow, and immune-desert subtypes. Survival analyses revealed that the immune-desert subtype had the worst prognosis, whereas subtypes with high levels of resident CD8 + T cells and a high ratio of resident to exhausted CD4 and NK cells displayed a favorable prognosis (Fig. 3B). Furthermore, multivariate Cox analysis demonstrated that only the ratio of resident to exhausted CD4 + T cells had an independent prognostic impact on HCC patients after removing confounding factors (Fig. 3C). Herein, subsequent analysis will focus on the subtyping method of the ratio of resident to exhausted CD4 + T cells.

Fig. 3
figure 3

Survival analysis, and multivariate Cox analysis. A Scatter plot among all resident scores, all exhausted scores, and tumor purity. The x-axis represented resident gene score, the y-axis represented exhausted gene score, and color changing represented tumor purity. The correlation R value was 0.78, and P < 2.2e-16. B Survival analysis in CD8 + T, CD4 + T, and NK cells. The resident, exhausted, and ratio scores were evaluated based on final resident and exhausted signatures. The residencyhighexhaustionlow subtype was represented by red; exhaustionhighresidencylow subtype was represented by blue and immune-desert subtype was represented by green. P < 0.05 represented a statistical difference among three subtypes. C Multivariate Cox analysis based on the age, sex, tumor grade, and all score information. HR was represented by blue squares, and 95% CI was represented by lines. 95%CI = 95% confidence interval, HR = hazard ratio

Full size image

Biological characteristics and immune landscape of three subtypes

Due to the survival difference among the three subtypes, we hypothesized that the subtypes may be endowed with different biological attributes. Enrichment analysis revealed that the CD4 residencyhighexhaustionlow subtype focused on immune activation pathways, including cytokine receptor interaction, activation of immune response, and acute inflammatory response (Fig. 4A), the CD4 exhaustionhighresidencylow subtype was featured by negative regulation of immune response, negative regulation of lymphocyte activation, and antigen processing and presentation pathways (Fig. 4B), while the immune-desert subtype was associated with proliferation-related pathways such as DNA repair, DNA replication, and E2F targets (Fig. 4C). The immune landscape among three subtypes was profiled to further delineate the difference in TIME among the three subtypes. Rich immune cells were particularly evident in the CD4 residencyhighexhaustionlow subtype and rare in the immune-desert subtype (Fig. 4D and E). Likewise, CIBERSORT results coincided with these findings (Additional file 2: Figure S7A and S7B).

Fig. 4
figure 4

Biological characteristics and immune landscape of three subtypes. A-C GSEA showed specific pathways in the CD4 residencyhighexhaustionlow (A), CD4 exhaustionhighresidencylow (B), and immune-desert subtype (C). D The heatmap of 28 immune cells among three subtypes. E The infiltration difference of 28 cells among three subtypes. The residencyhighexhaustionlow subtype was represented by red; the exhaustionhighresidencylow subtype was represented by blue, and the immune-desert subtype was represented by green. The asterisks represented the statistical p value (**P < 0.01; ***P < 0.001; ***P < 0.0001). F-G The immune landscape difference of leukocyte faction (F) and lymphocyte infiltration fraction (G) among three subtypes. GSEA = Gene Set Enrichment Analysis

Full size image

To reflect overall immune cell differences across three subtypes, leukocyte fraction and lymphocyte infiltration fraction were extracted from a previous report [33], which deteriorated from CD4 residencyhighexhaustionlow subtype to immune-desert subtype (Fig. 4F and G). In parallel, multiple co-stimulatory and co-inhibitory molecules were highly expressed in the CD4 residencyhighexhaustionlow subtype (Fig. 5A and Additional file 2: Figure S7C). Additionally, based on the classical immune subtypes described by V. Thorsson et al. [33], the association of CD4 ratio subtyping with reported immune subtypes was further explored. We found that C3 (inflammatory) was significantly involved in the CD4 residencyhighexhaustionlow subtype, while C4 (lymphocyte depleted) was mainly concentrated in the immune-desert subtype, consistent with the previous findings (Fig. 5B). Furthermore, to broadly evaluate the potential response to immunotherapy among different subtypes, TCR richness and CIC were assessed and showed that the CD4 residencyhighexhaustionlow subtype was endowed with a higher score, implying that this subtype more likely benefited from immunotherapy (Fig. 5C, Additional file 2: Figure S7D).

Fig. 5
figure 5

Genomic variation landscape among different subtypes. A The relative expression level of co-stimulatory molecules among three subtypes. B Difference of cancer immune cycle among three subtypes. C The proportion differences of immune subtypes among three subtypes. The height represented the percentile. The asterisks represented the statistical p value (*P < 0.05; *P < 0.01; ***P < 0.001; ****P < 0.0001). D Landscape of molecular variation among three subtypes. Each column represented an individual patient. The left number showed the frequency of mutational genes and regions. The right bar plot indicated the different mutation fractions among three subtypes. E-F Specific mutational differences of FGA, FGG, FGL (E), arm gain, arm loss, focal gain, and focal loss (F) among three subtypes. The residencyhighexhaustionlow subtype was represented by red; the exhaustionhighresidencylow subtype was represented by blue, and the immune-desert subtype was represented by green. FGA = fraction genome altered, FGG = fraction genome gain, FGL = fraction genome lost

Full size image

Genomic variation landscape among different subtypes

To systematically reveal the genomic traits of each subtype, we further characterized the landscape of genomic variation among three subtypes. As illustrated in Fig. 5D, the immune-desert subtype significantly harbored more mutations of TP53 (40%, P = 0.016), CTNNB1 (50%, P < 0.001), and OBSCN (60%, P = 0.03); the CD4 exhaustionhighresidencylow subtype concentrated in more mutation of ABCA13 (56%, P = 0.024); whereas the CD4 residencyhighexhaustionlow subtype displayed overall rare mutations across all hypervariable genes (Additional file 1: Table S5). Additionally, we further investigated chromosomal instability, chromosomal alteration regions (17q25.3 and Xq28 amplifications) were statistically gathered in the immune-desert subtype, and chromosomal loss regions significantly enriched in CD4 exhaustionhighresidencylow subtype (Fig. 5D, Additional file 1: Table S5). Similarly, FGA, FGG, and FGL in immune-desert and CD4 exhaustionhighresidencylow subtypes had higher levels, suggesting robust ability of tumor proliferation and immune escape (Fig. 5E, Additional file 1: Table S5). The copy number load was consistent with the above-mentioned findings (Fig. 5F).

Different responses to immunotherapy and sorafenib among three subtypes

To explore the difference in immunotherapeutic responses among three subtypes, patients from seven immunotherapeutic cohorts were assigned to three subtypes based on the same pipeline. Specifically, the CD4 residencyhighexhaustionlow subtype favored immunotherapy benefits (Fig. 6A-D, Additional file 2: Figure S7E). The ROC analysis was further performed to validate the efficacy of the ratio of resident to exhausted CD4 + T cells in predicting immunotherapeutic response. The results showed that the ratio of resident to exhausted CD4 + T cells demonstrated a stable and robust performance with AUC = 0.692 in GSE100797 (Fig. 6E), 0.967 in GSE111636 (Fig. 6F), 0.664 in GSE135222 (Fig. 6G), 0.764 in GSE35640 (Fig. 6H), and 0.655 in GSE91061 (Additional file 2: Figure S7F). Submap analysis further indicated a significantly similar expression pattern between CD4 residencyhighexhaustionlow subtype and immunotherapeutic responders after Bonferroni correction (Fig. 6I-N, Additional file 2: Figure S7G). As previously reported, the efficacy of sorafenib is also associated with the characteristics and infiltration fractions of TILs [34]. Therefore, a sorafenib cohort (GSE109211) was employed to test the possibility in our subtypes. We found that CD4 exhaustionhighresidencylow subtype was more sensitive to sorafenib and the AUC reached 0.819 (Fig. 6O and P). Overall, the ratio of resident to exhausted CD4 + T cells might serve as a promising biomarker for assessing the efficacy of immunotherapy and sorafenib.

Fig. 6
figure 6

Different responses to immunotherapy and sorafenib among three subtypes. A-H Percentage histogram of immunotherapeutic response in GSE100797 (melanoma) (A), GSE111636 (urothelial tumors) (B), GSE135222 (non-small cell lung carcinoma) (C) and GSE35640 (melanoma) (D). Red represented response to immunotherapy, and cyan represented non-response to immunotherapy. The AUC of immunotherapy in GSE100797 (E), GSE111636 (F), GSE135222 (G), and GSE35640 (H) were represented. I-N Submap was used to predict the immunotherapeutic response in GSE100797 (I), GSE111636 (J), GSE135222 (K), GSE136961 (non-small cell lung carcinoma) (L), GSE35640 (M) and GSE93157 (lung carcinoma, head and neck squamous cell carcinoma and melanoma) (N). O-P Percentage histogram (O) and AUC (P) of sorafenib response. The area under the AUC was in the lower right corner. Blue represented the nominal p-value, and red represented Bonferroni corrected p-value. P values were represented in figure if P < 0.05. Submap = subclass mapping. HCC = hepatocellular carcinoma. AUC = Area under the ROC Curve

Full size image

SASS6, a hub gene tightly associated with prognosis

In light of the darkest survival outcome among patients in the immune-desert group, an optimal biomarker with practical clinical applications was necessary. Ultimately, we successfully identified a candidate gene, SASS6, capable of accurately distinguishing patients within the immune-desert group in the TCGA-LIHC cohort (AUC = 0.712, Additional file 2: Figure S8A). Based on Kaplan–Meier curve analyses, elevated SASS6 expression was significantly associated with poorer OS for the TCGA-LIHC cohort (Additional file 2: Figure S8B). Using the BEST website, we further validated the above results. The Cox regression analysis demonstrated that SASS6 consistently acts as a risk factor in more than half of the cohorts with OS (Additional file 2: Figure S8C). To verify the prognostic impact of SASS6, after stratifying HCC patients into high and low SASS6 expression groups according to the median value, Kaplan–Meier survival analysis and log-rank test showed that the high group had a significantly shorter OS (Additional file 2: Figure S8D and S8E). We also found SASS6 expression in the tumor was markedly higher than that in adjacent normal tissues (Additional file 2: Figure S8F). Furthermore, the level of SASS6 expression was significantly associated with cancer progression (Additional file 2: Figure S8G). The GSEA, KEGG, and GO enrichment analysis also revealed that SASS6 was involved in cancer proliferation, invasion, and metastasis signaling pathways (Additional file 2: Figure S8H-J, Figure S9A and S9B). Moreover, we conducted a thorough analysis of the association between SASS6 and the abundance of immune cell infiltration using the BEST website. Our results show that SASS6 is negatively associated with the abundance of infiltration of crucial killer cells such as CD4 + T, CD8 + T, and NK cells (Additional file 2: Figure S9C). Additionally, our ROC curve analysis demonstrated that SASS6 can serve as a reliable predictor of response to immunotherapy (Additional file 2: Figure S9D).

Effects of SASS6 knockdown in HCCin vitroandin vivo

To validate the biological function of SASS6 in HCC cells, a special siRNA was designed to suppress the expression of SASS6 in the MHCC-97H cell line. As exhibited in Fig. 7A, the siRNA-2 efficiently reduced the expression of SASS6 in MHCC-97H cells (Additional file 1: Table S6). The CCK8 assay demonstrated that the proliferation ability of MHCC-97H cells was significantly inhibited after SASS6 knockdown (P < 0.05) (Fig. 7B, Additional file 1: Table S7). Further validation by EdU assay revealed that the suppression of SASS6 in MHCC-97H cell line markedly restrained cell proliferation compared to the control group (Fig. 7C). Scratch assay showed that depletion of SASS6 significantly inhibited the healing of scratched wound (Fig. 7D and E). Moreover, we examined the effect of SASS6 knockdown on HCC cell migration through Transwell assay, and the significantly reduced number of cells in the siRNA group suggested that SASS6 knockdown significantly inhibited migration ability (Fig. 7G and H). To assess whether SASS6 knockdown affects tumor formation in vivo, MHCC-97H SASS6 knockdown cell lines were generated by infecting cells with a SASS6 shRNA lentiviral vector. The targeting efficiency of the shRNA was confirmed by RT-qPCR (Fig. 7F, Additional file 1: Table S8). We performed an in vivo tumorigenesis study by subcutaneously injecting shCtrl-MHCC-97H and shSASS6-MHCC-97H cells into nude mice. As anticipated, mice injected with shSASS6-MHCC-97H cells showed significantly lower average tumor volumes and weights compared to those injected with control cells (F ig. 7I-K, Additional file 1: Table S9 and S10). Taken together, the inhibition of proliferation and migration of MHCC-97H cells after SASS6 knockdown suggests that SASS6 may promote the proliferation and metastasis of HCC.

Fig. 7
figure 7

Effects of SASS6 on HCC cells proliferation and migration in vitro and in vivo. A Expression was significantly reduced after SASS6 knockdown in HCC cell line. B CCK-8 proliferation assay of MHCC-97H cells in NC group and siRNA group. C EdU incorporation assay of HCC cell line in NC group and siRNA group. D-E Scratch assay to detect the healing ability of HCC cells in NC group and siRNA group. F Expression level of SASS6 in shCtrl and shSASS6 HCC cell line. G-H Transwell assay to detect the migratory ability of HCC cells in NC group and siRNA group. I-K Subcutaneous xenografts of MHCC-97H cells infected with either shCtrl lentivirus or shSASS6 lentivirus (n = 5). Tumor images at necropsy are shown, tumor volumes measured at designated time points, and average weights of the xenografted tumors. The asterisks represented the statistical p value (*P < 0.05; *P < 0.01; ***P < 0.001; ****P < 0.0001). qRT-PCR = quantitative real-time polymerase chain reaction PCR = polymerase chain reaction. CCK-8 = cell counting kit-8. OD = optical density. NC = negative control. EdU = 5-Ethynyl-20-deoxyuridine

Full size image

Discussion

Currently, immunotherapy has made considerable progress in solid tumors, but unfortunately, only a subset of patients generates benefits. TILs are the main components in TIME, and their resident and exhausted characteristics have a significant effect on the immunotherapeutic response [7, 12, 13]. Previous studies demonstrate the high percentage of TRM cells is associated with improved outcomes following checkpoint treatment [13]. However, studies focusing on the effect of T cell exhaustion status on immunotherapeutic response have yielded conflicting results. Hus and colleagues discovered that high CD8 + T cell exhausted status might indicate better efficacy of ICIs therapy [35], whereas M. Barsch et al. presented evidence that a high ratio of exhausted T cells suggests poor prognosis and scarce response to immunotherapy [13]. Our study aims to investigate the impact of resident and exhausted patterns on prognosis and immunotherapeutic response and identify a stable and powerful biomarker that predicts the HCC prognosis.

Recent studies have highlighted the importance of CD8 + T cells resident and exhausted patterns, whereas the sight of these patterns also gradually increased in CD4 + T cells and NK cells [11,12,13, 36, 37]. Similarly, mounting evidence showed that the resident and exhausted patterns in CD8 + T, CD4 + T, and NK cells could significantly affect the survival and immunotherapeutic response [10, 13, 28, 37]. In this study, we found that the resident and exhausted patterns of TILs mainly exhibited effects on CD8 + T, CD4 + T, and NK cells. Thus, we included these cells in our scope to comprehensively elucidate the impact of resident and exhausted patterns on immunotherapy. Although resident and exhausted signatures have been described, significant signature differences and different outcomes pose a challenge to accurately distinguish resident and exhausted patterns based on previous signatures [9, 12, 13]. Thus, we extracted resident and exhausted signatures according to the seven-step pipeline, which had great efficacy in distinguishing resident and exhausted patterns. The ratio subtype of resident to exhausted TILs was reported with the ability to reflect the resident and exhausted patterns together [12, 13]. Thus, we further increased the ratio subtypes of resident to exhausted CD8 + T, CD4 + T, and NK cells to better reflect the two patterns.

Patients with predominantly exhausted T cells are associated with a poor prognosis, whereas the high ratio of resident to exhausted TILs favored patient outcomes [13]. Here, no significant survival difference was observed between the exhaustionhigh and exhaustionlow subtypes of CD8 + T cells. Together, although there were statistically significant survival differences in some grouping methods, which were consistent with previous studies [11, 13]. Moreover, crosstalk among different factors has proven to lead to the survival result difference [12]. We then conducted multivariate Cox analysis and found that only the ratio of resident to exhausted CD4 + T cells significantly influenced survival. Oja, A. E. et al. found that resident CD4 + T cells are associated with patient survival and immunotherapeutic response [36]. Likewise, the ratio of resident to exhausted T cells has been reported to be a robust biomarker in predicting survival and immunotherapeutic response [12, 13]. Thus, we hypothesized that the ratio of resident and exhausted CD4 + T cells could potentially enhance the precise management of HCC. Due to the essential roles of tumor purity on survival and immunotherapeutic response [32], we stratified HCC patients into three subtypes based on both the ratio of resident and exhausted CD4 + T cells and tumor purity, in which the survival deteriorated gradually from CD4 residencyhighexhaustionlow subtype to immune-desert subtype.

Functional enrichment analysis revealed that the immune-desert subtype was endowed with malignant-related functions, such as E2F targets and DNA repair. As previously reported, high expression of E2F targets indicates that tumor cells have strong proliferation ability and may evade immune system recognition and attack by down-regulating immunogenicity [38], suggesting the immune-desert subtype has higher proliferative capacity. Also, we found that immune activation was assigned to the CD4 residencyhighexhaustionlow subtype, while the CD4 exhaustionhighresidencylow subtype was primarily endowed with negative immune-related pathways. Specifically, immune-related characteristics were significantly enriched in the CD4 residencyhighexhaustionlow subtype: rich immune cells (such as B, CD8 + T, and CD4 + T cells) were observed to be concentrated in the CD4 residencyhighexhaustionlow subtype. The abundance of TILs has been reported to be significantly correlated with survival and immunotherapeutic response [39, 40]. Several immune relevant factors have been shown to affect the clinical outcome. For example, co-stimulatory and co-inhibitory molecules, such as CD27, ICOS, PDCD1, and HAVCR2, are associated with immunotherapeutic response [41]. Moreover, the increasing TCR richness and CIC also predict better immunotherapeutic outcomes [26, 42,43,44,45]. Our data demonstrated that the CD4 residencyhighexhaustionlow subtype highly expressed these co-stimulatory/inhibitory molecules and was associated with the high immunity cycle score and TCR richness, indicating potentially better immunotherapeutic outcome.

To further comprehensively reveal the differences in survival among the three subtypes, the different landscapes of molecular variation were analyzed. From a global perspective, the overall variation was predominantly assigned to immune-desert and CD4 exhaustionhighresidencylow subtypes, which belonged to mutation-driven and copy number loss-driven, separately. Specifically, the immune-desert subtype was endowed with more alterations of TP53, CTNNB1, and OBSCN. TP53 mutation has been well-known to promote cancer cell survival and evade tumor immune surveillance [46]. CTNNB1 mutation is related to an immune-excluded response in HCC, which has been used to predict innate insensitivity to immunotherapy [47]. Of note, the copy number loss robustly occurred in the CD4 exhaustionhighresidencylow subtype. The copy number loss is reported to be mainly associated with cell proliferation, worse prognosis, and evading immune surveillance, which could predict the response to immunotherapy [48, 49]. Overall, these genomic variations suggested that distinct mutations not only led to different immune status and survival outcomes but also provided the potential mechanisms for the response to immunotherapy among three subtypes, which might throw light on guiding the precise treatment for HCC.

PD-L1 expression and immune exhausted status are clinically popular indicators for predicting the efficacy of HCC immunotherapy, but many studies have found them unsatisfactory in prediction accuracy [12, 13, 50]. Thus, to explore a potential better biomarker to predict the efficacy of immunotherapy, the immunotherapeutic response fraction among different subtypes and the AUC of immunotherapeutic response were assessed. Consistent with the immune landscape results, TCR richness, and CIC analyses, the result revealed that the CD4 residencyhighexhaustionlow subtype showed a better immunotherapeutic response, which was supported by Submap analysis. Overall, our findings provide new insights that may inform future strategies for optimizing immunotherapy in HCC, though more research is required before these insights can be translated into treatment modifications. Additionally, a previous study in the murine liver cancer model has revealed that Sorafenib could modulate immunosuppressive cell populations to enhance anti-tumor immunity [51]. Therefore, to verify the relationship between sorafenib efficacy and exhausted procedures, the response to sorafenib was assessed and the results showed that the CD4 exhaustionhighresidencylow subtype had a significantly higher sorafenib response, with the prediction accuracy reaching an AUC value of 0.819.

SASS6 is an essential protein required for centrosome replication, involved in initiating centriole formation and stabilizing centriole intermediates. Previous studies have revealed that SASS6 expression was significantly upregulated in colon cancer, breast cancer, esophageal squamous carcinoma, and other tumor tissues and associated with unsatisfactory prognosis of lung adenocarcinoma, colon cancer, kidney renal clear cell carcinoma, and esophageal squamous carcinoma [52,53,54]. In this study, we demonstrated that SASS6, which could accurately predict the worst prognosis subtype, was highly expressed in HCC and associated with poorer prognosis in a multicenter cohort through the BEST web tool. Further, through multiple functional assays, we confirmed that SASS6 indeed inhibited the proliferation and migration of HCC cells in vitro and in vivo. These results are consistent with previous research in other tumors, suggesting that SASS6 may serve as a key prognostic biomarker and therapeutic target in HCC, though additional studies are required to confirm its clinical relevance and therapeutic potential.

Our study demonstrated that the CD4 residencyhighexhaustionlow subtype predicted better immunotherapeutic responses, which can potentially facilitate accurate clinical guidance of immunotherapy in HCC. In clinical practice, our signature validated by more multicenter cohorts in the future could better guide precision treatment. Patients with higher ratio of resident to exhausted CD4 + T cells may experience better prognosis and potentially enhanced response to immunotherapy. On the other hand, a lower ratio of these cells demonstrates an intermediate prognosis and possibly greater sensitivity to sorafenib, while patients exhibiting higher SASS6 expression, indicative of an immune-desert subtype, are associated with significantly poorer prognosis and proliferation-related biological behavior and required combination therapy strategies.

This study suggested a role for resident and exhausted CD4 + T cells in predicting immunotherapy response, contributing novel insights into their potential prognostic value in HCC. Nevertheless, our study also had several limitations. Firstly, lacking public data on immunotherapy for HCC, we predicted the immunotherapeutic results for HCC based on immunotherapeutic cohorts of other cancer types. Further validation in immunotherapeutic cohorts of HCC is needed. Secondly, our study focused mainly on the impact of resident and exhausted patterns on HCC, and more investigations should be conducted on pan-cancer. We need to include more tumor types and datasets to improve the robustness and credibility of the conclusions. Thirdly, in this study, we only included major CD4 + T cells, CD8 + T cells, and NK cells to investigate the impact of the resident and exhausted status of TILs on HCC prognosis and response to immunotherapy. A comprehensive understanding of more TILs subsets such as regulatory T cells, gamma delta T cells, B cells and plasma cells is necessary in the future. Finally, given the potential impact of tumor heterogeneity on our results, additional clinical and experimental validation is imperative in the future. This may encompass the examination of larger sample sizes, more comprehensive investigations into underlying mechanisms, and cross-validation across diverse populations. Through further research endeavors, we will gain enhanced insights into disease prevention and treatment modalities for improved patient care.

Conclusion

We propose a new stratification strategy based on the ratio of resident to exhausted CD4 + T cells, which shows great performance in predicting prognosis and immunotherapeutic response. Subsequently, we highlighted the potential of SASS6 as a therapeutic target for the treatment of HCC. In the future, our findings, validated by preclinical and prospective multicenter clinical studies, will provide new insights for early diagnosis, stratified prognostic management, and individualized immunotherapy of HCC patients in the clinical practice.

Availability of data and materials

The datasets used in this study can be found on the respective websites listed in the article. Other necessary data and supporting codes will be made available by contacting the corresponding author.

Abbreviations

TILs:

Tumor-infiltrating lymphocytes

TIME:

Tumor immune microenvironment

HCC:

Hepatocellular carcinoma

ICIs:

Immune checkpoint inhibitors

OS:

Overall survival

ROC:

Receiver operating characteristic

GEO:

Gene Expression Omnibus

CCLE:

Cancer cell line encyclopedia

UMIs:

Unique molecular identifiers

UMAP:

Uniform Manifold Approximation and Projection

ssGSEA:

Single-Sample Gene Set Enrichment Analysis

GSEA:

Gene Set Enrichment Analysis

GO:

Gene Ontology

KEGG:

Kyoto Encyclopedia of Genes and Genomes

TCR:

T cell receptor

CIC:

Cancer immunity cycle

CNA:

Copy number alternation

FGA:

Fraction genome alteration

FGG:

Fraction genome gain

FGL:

Fraction genome loss

AUC:

The area under the receiver operating characteristic curve

Submap:

Subclass mapping

SASS6 :

Spindle assembly abnormal protein 6 homolog

NC:

Negative control

siRNA:

Small interfering RNA

qRT-PCR:

Quantitative real-time polymerase chain reaction

CCK-8:

Cell counting kit-8

OD:

Optical density

EdU:

5-Ethynyl-2'-deoxyuridine

References

  1. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209–49.

    Article  PubMed  Google Scholar 

  2. Iizuka N, Hamamoto Y, Oka M. Predicting individual outcomes in hepatocellular carcinoma. Lancet. 2004;364(9448):1837–9.

    Article  PubMed  Google Scholar 

  3. Bruix J, Qin S, Merle P, Granito A, Huang YH, Bodoky G, et al. Regorafenib for patients with hepatocellular carcinoma who progressed on sorafenib treatment (RESORCE): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet. 2017;389(10064):56–66.

    Article  CAS  PubMed  Google Scholar 

  4. Pons-Tostivint E, Latouche A, Vaflard P, Ricci F, Loirat D, Hescot S, et al. Comparative Analysis of Durable Responses on Immune Checkpoint Inhibitors Versus Other Systemic Therapies: A Pooled Analysis of Phase III Trials. JCO Precis Oncol. 2019;3:1–10.

    Article  PubMed  Google Scholar 

  5. Llovet JM, Castet F, Heikenwalder M, Maini MK, Mazzaferro V, Pinato DJ, et al. Immunotherapies for hepatocellular carcinoma. Nat Rev Clin Oncol. 2022;19(3):151–72.

    Article  CAS  PubMed  Google Scholar 

  6. Chew V, Toh HC, Abastado JP. Immune microenvironment in tumor progression: characteristics and challenges for therapy. J Oncol. 2012;2012: 608406.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Paijens ST, Vledder A, de Bruyn M, Nijman HW. Tumor-infiltrating lymphocytes in the immunotherapy era. Cell Mol Immunol. 2021;18(4):842–59.

    Article  CAS  PubMed  Google Scholar 

  8. Cursons J, Souza-Fonseca-Guimaraes F, Foroutan M, Anderson A, Hollande F, Hediyeh-Zadeh S, et al. A Gene Signature Predicting Natural Killer Cell Infiltration and Improved Survival in Melanoma Patients. Cancer Immunol Res. 2019;7(7):1162–74.

    Article  CAS  PubMed  Google Scholar 

  9. Yu X, Zhang L, Chaudhry A, Rapaport AS, Ouyang W. Unravelling the heterogeneity and dynamic relationships of tumor-infiltrating T cells by single-cell RNA sequencing analysis. J Leukoc Biol. 2020;107(6):917–32.

    Article  CAS  PubMed  Google Scholar 

  10. Byrne A, Savas P, Sant S, Li R, Virassamy B, Luen SJ, et al. Tissue-resident memory T cells in breast cancer control and immunotherapy responses. Nat Rev Clin Oncol. 2020;17(6):341–8.

    Article  PubMed  Google Scholar 

  11. Djenidi F, Adam J, Goubar A, Durgeau A, Meurice G, de Montpréville V, et al. CD8+CD103+ tumor-infiltrating lymphocytes are tumor-specific tissue-resident memory T cells and a prognostic factor for survival in lung cancer patients. J Immunol. 2015;194(7):3475–86.

    Article  CAS  PubMed  Google Scholar 

  12. Foroutan M, Molania R, Pfefferle A, Behrenbruch C, Scheer S, Kallies A, et al. The Ratio of Exhausted to Resident Infiltrating Lymphocytes Is Prognostic for Colorectal Cancer Patient Outcome. Cancer Immunol Res. 2021;9(10):1125–40.

    Article  CAS  PubMed  Google Scholar 

  13. Barsch M, Salié H, Schlaak AE, Zhang Z, Hess M, Mayer LS, et al. T-cell exhaustion and residency dynamics inform clinical outcomes in hepatocellular carcinoma. J Hepatol. 2022;77(2):397–409.

    Article  CAS  PubMed  Google Scholar 

  14. Lv J, Wang L, Jia W, Xu H, Weng S, Zhang Y, et al. Resident to exhausted CD4+ T cell ratio is associated with the prognosis of gastric cancer. Genes Dis. 2024;11(4): 101076.

    Article  CAS  PubMed  Google Scholar 

  15. Zhang Q, He Y, Luo N, Patel SJ, Han Y, Gao R, et al. Landscape and Dynamics of Single Immune Cells in Hepatocellular Carcinoma. Cell. 2019;179(4):829-45.e20.

    Article  CAS  PubMed  Google Scholar 

  16. Ma L, Hernandez MO, Zhao Y, Mehta M, Tran B, Kelly M, et al. Tumor Cell Biodiversity Drives Microenvironmental Reprogramming in Liver Cancer. Cancer Cell. 2019;36(4):418-430.e6.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Meng Y, Zhao Q, An L, Jiao S, Li R, Sang Y, et al. A TNFR2-hnRNPK Axis Promotes Primary Liver Cancer Development via Activation of YAP Signaling in Hepatic Progenitor Cells. Cancer Res. 2021;81(11):3036–50.

    Article  CAS  PubMed  Google Scholar 

  18. Ghandi M, Huang FW, Jané-Valbuena J, Kryukov GV, Lo CC, McDonald ER, et al. Next-generation characterization of the Cancer Cell Line Encyclopedia. Nature. 2019;569(7757):503–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Dogra P, Rancan C, Ma W, Toth M, Senda T, Carpenter DJ, et al. Tissue Determinants of Human NK Cell Development, Function, and Residence. Cell. 2020;180(4):749-63.e13.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Zhang X, Lan Y, Xu J, Quan F, Zhao E, Deng C, et al. Cell Marker: a manually curated resource of cell markers in human and mouse. Nucleic Acids Res. 2019;47(D1):D721–8.

    Article  CAS  PubMed  Google Scholar 

  21. Kumar BV, Ma W, Miron M, Granot T, Guyer RS, Carpenter DJ, et al. Human Tissue-Resident Memory T Cells Are Defined by Core Transcriptional and Functional Signatures in Lymphoid and Mucosal Sites. Cell Rep. 2017;20(12):2921–34.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Foroutan M, Bhuva DD, Lyu R, Horan K, Cursons J, Davis MJ. Single sample scoring of molecular phenotypes. BMC Bioinformatics. 2018;19(1):404.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Street K, Risso D, Fletcher RB, Das D, Ngai J, Yosef N, et al. Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC Genomics. 2018;19(1):477.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545–50.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Liu Z, Zhang Y, Shi C, Zhou X, Xu K, Jiao D, et al. A novel immune classification reveals distinct immune escape mechanism and genomic alterations: implications for immunotherapy in hepatocellular carcinoma. J Transl Med. 2021;19(1):5.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Kansy BA, Shayan G, Jie HB, Gibson SP, Lei YL, Brandau S, et al. T cell receptor richness in peripheral blood increases after cetuximab therapy and correlates with therapeutic response. Oncoimmunology. 2018;7(11): e1494112.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Hoshida Y, Brunet JP, Tamayo P, Golub TR, Mesirov JP. Subclass mapping: identifying common subtypes in independent disease data sets. PLoS ONE. 2007;2(11): e1195.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Philip M, Schietinger A. CD8(+) T cell differentiation and dysfunction in cancer. Nat Rev Immunol. 2022;22(4):209–23.

    Article  CAS  PubMed  Google Scholar 

  29. Lu Z, Liu R, Wang Y, Jiao M, Li Z, Wang Z, et al. Ten-eleven translocation-2 inactivation restrains IL-10-producing regulatory B cells to enable antitumor immunity in hepatocellular carcinoma. Hepatology. 2023;77(3):745–59.

    PubMed  Google Scholar 

  30. Yang Y, Chen X, Pan J, Ning H, Zhang Y, Bo Y, et al. Pan-cancer single-cell dissection reveals phenotypically distinct B cell subtypes. Cell. 2024;187(17):4790-4811.e22.

    Article  CAS  PubMed  Google Scholar 

  31. Mao Y, Feng Q, Zheng P, Yang L, Liu T, Xu Y, et al. Low tumor purity is associated with poor prognosis, heavy mutation burden, and intense immune phenotype in colon cancer. Cancer Manag Res. 2018;10:3569–77.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Gong Z, Zhang J, Guo W. Tumor purity as a prognosis and immunotherapy relevant feature in gastric cancer. Cancer Med. 2020;9(23):9052–63.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Thorsson V, Gibbs DL, Brown SD, Wolf D, Bortone DS, Ou Yang TH, et al. The Immune Landscape of Cancer. Immunity. 2018;48(4):812-30.e14.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Tang W, Chen Z, Zhang W, Cheng Y, Zhang B, Wu F, et al. The mechanisms of sorafenib resistance in hepatocellular carcinoma: theoretical basis and therapeutic aspects. Signal Transduct Target Ther. 2020;5(1):87.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Hsu CL, Ou DL, Bai LY, Chen CW, Lin L, Huang SF, et al. Exploring Markers of Exhausted CD8 T Cells to Predict Response to Immune Checkpoint Inhibitor Therapy for Hepatocellular Carcinoma. Liver Cancer. 2021;10(4):346–59.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Oja AE, Piet B, van der Zwan D, Blaauwgeers H, Mensink M, de Kivit S, et al. Functional Heterogeneity of CD4(+) Tumor-Infiltrating Lymphocytes With a Resident Memory Phenotype in NSCLC. Front Immunol. 2018;9:2654.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Gao Y, Souza-Fonseca-Guimaraes F, Bald T, Ng SS, Young A, Ngiow SF, et al. Tumor immunoevasion by the conversion of effector NK cells into type 1 innate lymphoid cells. Nat Immunol. 2017;18(9):1004–15.

    Article  CAS  PubMed  Google Scholar 

  38. Chen Z, Zhou L, Liu L, Hou Y, Xiong M, Yang Y, et al. Single-cell RNA sequencing highlights the role of inflammatory cancer-associated fibroblasts in bladder urothelial carcinoma. Nat Commun. 2020;11(1):5077.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Huang S, Song Z, Zhang T, He X, Huang K, Zhang Q, et al. Identification of Immune Cell Infiltration and Immune-Related Genes in the Tumor Microenvironment of Glioblastomas. Front Immunol. 2020;11: 585034.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Havel JJ, Chowell D, Chan TA. The evolving landscape of biomarkers for checkpoint inhibitor immunotherapy. Nat Rev Cancer. 2019;19(3):133–50.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. O’Neill RE, Cao X. Co-stimulatory and co-inhibitory pathways in cancer immunotherapy. Adv Cancer Res. 2019;143:145–94.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Wang X, Zhou T, Chen X, Wang Y, Ding Y, Tu H, et al. System analysis based on the cancer–immunity cycle identifies ZNF207 as a novel immunotherapy target for hepatocellular carcinoma. J Immunother Cancer. 2022;10(3): e004414.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Hu Y, Sun H, Zhang H, Wang X. An Immunogram for an Individualized Assessment of the Antitumor Immune Response in Patients With Hepatocellular Carcinoma. Front Oncol. 2020;10:1189.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Chen DS, Mellman I. Oncology meets immunology: the cancer-immunity cycle. Immunity. 2013;39(1):1–10.

    Article  PubMed  Google Scholar 

  45. Pio R, Ajona D, Ortiz-Espinosa S, Mantovani A, Lambris JD. Complementing the Cancer-Immunity Cycle. Front Immunol. 2019;10:774.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Ghosh M, Saha S, Bettke J, Nagar R, Parrales A, Iwakuma T, et al. Mutant p53 suppresses innate immune signaling to promote tumorigenesis. Cancer Cell. 2021;39(4):494-508.e5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Pinyol R, Sia D, Llovet JM. Immune Exclusion-Wnt/CTNNB1 Class Predicts Resistance to Immunotherapies in HCC. Clin Cancer Res. 2019;25(7):2021–3.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Marcon J, DiNatale RG, Sanchez A, Kotecha RR, Gupta S, Kuo F, et al. Comprehensive Genomic Analysis of Translocation Renal Cell Carcinoma Reveals Copy-Number Variations as Drivers of Disease Progression. Clin Cancer Res. 2020;26(14):3629–40.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. Davoli T, Uno H, Wooten EC, Elledge SJ. Tumor aneuploidy correlates with markers of immune evasion and with reduced response to immunotherapy. Science. 2017;355(6322):eaaf8399.

    Article  PubMed  PubMed Central  Google Scholar 

  50. Zhang Z, Wang ZX, Chen YX, Wu HX, Yin L, Zhao Q, et al. Integrated analysis of single-cell and bulk RNA sequencing data reveals a pan-cancer stemness signature predicting immunotherapy response. Genome Med. 2022;14(1):45.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Cao M, Xu Y, Youn JI, Cabrera R, Zhang X, Gabrilovich D, et al. Kinase inhibitor Sorafenib modulates immunosuppressive cell populations in a murine liver cancer model. Lab Invest. 2011;91(4):598–608.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Shinmura K, Kato H, Kawanishi Y, Nagura K, Kamo T, Okubo Y, et al. SASS6 overexpression is associated with mitotic chromosomal abnormalities and a poor prognosis in patients with colorectal cancer. Oncol Rep. 2015;34(2):727–38.

    Article  CAS  PubMed  Google Scholar 

  53. Du LL, Jing JX, Wang Y, Xu XQ, Sun T, Shi YC, et al. Knockdown of SASS6 reduces growth of MDA-MB-231 triple-negative breast cancer cells through arrest of the cell cycle at the G2/M phase. Oncol Rep. 2021;45(6):101.

    Article  CAS  PubMed  Google Scholar 

  54. Xu YJ, Zhu KS, Chen JQ, Lin LY, Huang ZR, Zhang JL, et al. SASS6 promotes proliferation of esophageal squamous carcinoma cells by inhibiting the p53 signaling pathway. Carcinogenesis. 2021;42(2):254–62.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

Not applicable.

Funding

Not applicable.

Author information

Author notes

  1. Anning Zuo, Jinxiang Lv and Wenlong Jia contributed equally to this work and share the first authorship.

Authors and Affiliations

  1. Department of Interventional Radiology, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, 450052, China

    Anning Zuo, Yuhao Ba, Yuyuan Zhang, Siyuan Weng, Hui Xu, Xinwei Han & Zaoqu Liu

  2. Interventional Institute of Zhengzhou University, Zhengzhou, Henan, 450052, China

    Anning Zuo, Xinwei Han & Zaoqu Liu

  3. Interventional Treatment and Clinical Research Center of Henan Province, Zhengzhou, Henan, 450052, China

    Anning Zuo, Xinwei Han & Zaoqu Liu

  4. Department of Gastroenterology, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan, 450052, China

    Jinxiang Lv

  5. Hepatic Surgery Center, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, China

    Wenlong Jia

  6. School of Basic Medical Sciences, College of Medicine, Zhengzhou University, Zhengzhou, Henan, 450052, China

    Shutong Liu

  7. Department of Hepatobiliary and Pancreatic Surgery, The First Affiliated Hospital of Xi’an Jiaotong University, Xi’an, Shanxi, 710049, China

    Long Liu

  8. Department of Pancreatic Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjin’s Clinical Research Center for Cancer, Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Tianjin, 300060, China

    Libo Wang

  9. Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100730, China

    Zaoqu Liu

  10. State Key Laboratory of Proteomics, Beijing Proteome Research Center, National Center for Protein Sciences (Beijing), Beijing Institute of Lifeomics, Beijing, 102206, China

    Zaoqu Liu

Authors
  1. Anning Zuo
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jinxiang Lv
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Wenlong Jia
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Yuhao Ba
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Shutong Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Yuyuan Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Siyuan Weng
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Hui Xu
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Long Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Libo Wang
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Xinwei Han
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Zaoqu Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

ZQL, XWH, and LBW provided direction and guidance throughout the preparation of this manuscript. ANZ, JXL, WLJ, ZQL, and LBW wrote and edited the manuscript. ZQL reviewed and made significant revisions to the manuscript. YHB, STL, YYZ, SYW, HX, LL, WLJ, LBW, and ZQL collected and prepared the related papers. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Libo Wang, Xinwei Han or Zaoqu Liu.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zuo, A., Lv, J., Jia, W. et al. High ratio of resident to exhausted CD4 + T cells predicts favorable prognosis and potentially better immunotherapeutic efficacy in hepatocellular carcinoma. BMC Cancer 24, 1152 (2024). https://doi.org/10.1186/s12885-024-12916-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12885-024-12916-0

Keywords

BMC Cancer

ISSN: 1471-2407

Contact us