CD4⁺ T lymphocytes play an essential role in regulating the immune system and promoting anti-tumor responses through interactions restricted by MHC II [22]. In contact with various antigens, naive CD4⁺ T cells can differentiate into distinct subsets, including Tregs, Th1, Th2, Th17, and follicular helper T cells. These subsets of CD4⁺ T cells produce a variety of cytokines that modulate inflammatory responses and activate gene programs in CD8⁺ T cells, thereby enhancing their anti-tumor response [23–25]. CD4⁺ CTLs represent a distinct subset of CD4 cells that exhibit cytotoxic activity. Both preclinical and clinical investigations have established the direct anti-tumor effects of CD4⁺ on CTLs by the release of cytotoxic granules containing granzyme B and perforin [26–28]. In the context of CCA, the overall quantity of tumor-infiltrating Tregs and CD4⁺ CTLs serves as independent prognostic markers for tumor grading and the Union for International Cancer Control (UICC) stage [18].

Distribution of CD4⁺ T cells in the peritumoral and intratumoral regions of CCA has been explored by numerous studies; however, their findings remain conflicting. Specifically, five studies have reported a significant increase of CD4⁺ T-cell infiltration in the peritumoral region, compared to the tumor center in ICCA [18, 29–32]. Conversely, one study has demonstrated that CD4⁺ T-cell infiltration at the tumor edge is significantly higher than in the tumor center in both ICCA and extrahepatic CCA (eCCA) [18]. The BTC tumor edge serves as the primary location for active infiltration by FOXP3⁺CD4⁺ Tregs that express elevated levels of lymphocyte activation gene 3 (LAG-3) and T cell immunoglobulin and mucin domain-containing protein 3 (TIM3) [33, 34]. Furthermore, a high density of FOXP3^–CD4⁺ Th cells at the tumor edge is independently associated with longer PFS and OS in patients undergoing palliative gemcitabine plus cisplatin treatment [33]. Conversely, another study has found no significant difference in the distribution of CD4⁺ T cells between the peritumoral and intratumoral regions [35]. The correlation between the distribution of inflammatory cells in CCA and local microvessel density (MVD) has also been established. Compared with metastatic CCA patients, immune cells, such as CD4⁺ T lymphocytes, of patients with locally advanced tumors are more infiltrated. Additionally, patients with metastatic CCA exhibit a significant increase in MVD, compared to those with localized CCA, indicating a negative correlation between MVD and inflammatory cell infiltration [36]. A further investigation demonstrated that a reduced MVD was associated with a poorer prognosis in patients undergoing treatment for ICCA. This was indicated by the significantly shorter recurrence-free survival observed in low MVD patients than in their high MVD counterparts. The correlation between MVD and CD8⁺ TILs infiltration was highly positive, while it was negative between MVD and FOXP3⁺ TILs, which could be attributed to the modulation of TILs recruitment [37].

CD8⁺ T lymphocytes infiltrate in the TME to eliminate cancer cells [42]. These lymphocytes proliferate and, activated by a specific antigen, differentiate into effector cells that are specifically referred to as CD8⁺ cytotoxic CTLs. Retaining their killing function, CD8⁺ CTLs target and eliminate cells displaying a specific antigen through two distinct mechanisms: the perforin/granzyme pathway and the FAS/FASL pathway. On the former pathway, CD8⁺ CTLs induce apoptosis of target cells by releasing cytotoxic molecules, including perforin and granzyme B as well as cytokines. On the latter pathway, they induce apoptosis by binding the FASL to the FAS receptor on the target cell. The quantity and efficacy of CD8⁺ T cells significantly influence the anti-neoplastic response and thus the prognosis of patients with CCA [37, 43].

PD1/PDL1 has emerged as a prevalent clinical biomarker for prognosticating the effectiveness of immunotherapy in solid tumors [60]. To note, extensive clinical investigations have been conducted to assess the expression of PD1/PDL1 and its correlation with the response to ICI therapy in CCA. Nevertheless, unlike other solid tumors, the predictive value of PDL1 in CCA remains inconclusive based on existing studies. In the KEYNOTE-158 study, it was observed that the objective response rate (ORR) among patients with positive PDL1 expression was merely 6.6%, whereas the ORR for patients lacking PDL1 expression was 2.9% [61]. This finding suggests a modest association between PDL1 expression and the response to ICI. The findings from KEYNOTE-158 and KEYNOTE-028 trials indicate that the effectiveness of pembrolizumab therapy appears to be unaffected by the expression of PDL1 in CCA cells [62]. Interestingly, the detection rate of PDL1 also varied depending on the clone type, heterogeneity of samples, and limitations in sample size [19]. Specifically, studies using clone streptavidin-perosidase (SP) 142 reported a PDL1 positivity rate of 8.6%, whereas studies employing 5H1 (an PDL1 antibody to quantify PDL1 in tumors and immune cells) reported a rate of 72% [63]. Additionally, a study using E1L3 (an PDL1 antibody to quantify PDL1 in tumors and immune cells) revealed PDL1 expression in 30.5% of cancer cases [64]. Conversely, a recent clinical trial (NCT02829918) demonstrated that the clinical effectiveness of nivolumab was contingent upon the expression of PDL1 [65].

Many signaling pathways, including Wnt/β-catenin, TGF signaling pathway, (an inhibitor of atypical protein kinase C-i) aPKC-i/phosphorylated transcription factor Sp1 (P-Sp1)/snail family transcriptional repressor (Snail) signaling pathway, hippo pathway, B7H1/PD1 pathway, and FAS/FASL pathway, are implicated in the malignant potential of CCA, as well as in immune evasion by inducing apoptosis of TILs [66–73]. Furthermore, apart from immune checkpoints, there is a growing research emphasis on exploring other molecular markers as potential targets for immunotherapy in this particular cancer type. Martin-Serrano et al. [40] conducted a transcriptomic analysis on a cohort of 900 ICCA cases by integrating components of the stroma, tumor, and immune microenvironment. As a result, they proposed a novel classification for ICCA termed the “stromal interaction molecule (STIM)” classification. Over 60% of ICCA cases exhibited a non-inflammatory phenotype, characterized by the presence of stemness-related pathways, BRCA1 associated protein 1 (BAP1), and isocitrate dehydrogenase 1 (IDH1) mutations, as well as focal loss of Salvador family WW domain containing protein 1 (SAV1). This phenotype is characterized by an abundance of immune inhibitory components [40].

It has been reported that cholangiocarcinogenesis is facilitated by Yes-associated protein (YAP) through transcriptional enhanced associate domain (TEAD)-dependent transcriptional activation of β-catenin, where the YAP interacts with the transcription factor TEAD to regulate gene expression [74]. In ICCA, β-catenin forms a binding association with YAP and is essential for the complete transcriptional activity of YAP, thereby highlighting the functional interplay between the YAP and β-catenin pathways in cholangiocarcinogenesis [70]. Song et al. [75] reported that focal adhesion kinase (FAK) activation induces the proto-oncogene YAP, contributing to ICCA initiation and progression. FAK inhibitors, alone or in combination with anti-cyclin-dependent kinase 4/6 (anti-CDK4/6) inhibitors, may be effective treatments for ICCA [75]. Since YAP1 stimulates the transformation of hepatocytes into ICCA, it is necessary to investigate the association between the YAP1 pathway and the immune inhibitory microenvironment in CCA [76].

In a recent study, Zhang et al. [77] identified the pivotal gene Mucin 1 (MUC1) in the progression of CCA and predicted its potential downstream carcinogenic pathways through high-throughput sequencing datasets from the Gene Expression Omnibus (GEO) database. As a result of the interaction between MUC1 and a cell surface receptor, epidermal growth factor receptor (EGFR), the EGFR/phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt) signaling pathway is activated, resulting in the accumulation of FOXP3⁺ Treg cells. In vivo, this accumulation of cells enhances the malignant phenotype of CCA cells, which ultimately facilitates metastasis and promotes CCA growth [77]. These studies offer prospective and valid options for the treatment of CCA.

IDH1 and IDH2 are frequently mutated metabolic genes in human cancers [78, 79]. IDH mutations are particularly prevalent in ICCA among epithelial malignancies [80]. The presence of mutant IDH (mIDH) in ICCA promotes its development through the production of 2-hydroxyglutarate (2HG) and the inhibition of hepatocyte nuclear factor 4 (HNF-4) [81]. The pharmacological inhibition of mIDH1 has shown promise in delaying the progression of mIDH1 ICCA. However, the activation of immune cells and stimulation of Treg cells ultimately lead to tumor progression [82, 83]. Using genetically engineered mouse models driven by mIDH1, researchers inactivated the ten-eleven translocation 2 gene (TET2) of DNA demethylase with 2HG produced by mutant IDH enzymes. Significantly, researchers discovered that the combination therapy of mIDH1 inhibitor AG120 (an orally active inhibitor of mIDH1 enzyme) and anti-CTLA-4 antibody induced substantial cell death and sustained tumor-specific T-cell responses in tumor-bearing mice [84]. These findings support further development of immunotherapy for patients with mIDH1-mutant ICCA.

Circular RNAs (circRNAs) are also a biomarker of iCCA. Circular RNA HMGCS1-016 (circHMGCS1-016) promotes invasion of ICCA through the miR-1236-3p/CD73 and galectin 8 (GAL-8) pathway while downregulating CD8⁺ T-cell infiltration. Furthermore, circHMGCS1-016 holds promise as a novel potential biomarker for predicting resistance to PD1 inhibition in ICCA [85]. In addition, CD73 plays a role in promoting the proliferation, migration, invasion, and epithelial-mesenchymal transition of ICCA. Analysis of clinical samples from 259 resected ICCA cases reveal a significant association between high CD73 expression and an increased proportion of FOXP3⁺/CD8⁺ TILs and CD163⁺/CD68⁺ tumor-associated macrophages (TAMs) [86]. Moreover, an increased expression of LAIR2 was correlated with poorer patient survival, indicating the potential of LAIR2 as a marker for exhausted T-cell populations [87].

CAR T therapy is an immunotherapeutic approach that entails the genetic modification of T lymphocytes to express an artificial receptor capable of identifying and binding to particular target antigens (Figure 4). Extensive evidence has demonstrated the remarkable efficacy and long-lasting clinical responses of CAR T therapy in diverse hematological malignancies and solid tumors [88–92]. CAR T therapy, a breakthrough in cancer treatment, is currently being used to combat refractory CCA.

Human EGFR 2 (HER2)-CAR T have demonstrated high patient tolerability without dose-limiting toxicities [93]. Antigens such as EGFR and B7H3 have been found to be highly expressed on CCA cells. Studies have shown that B7H3-CAR T and EGFR-CAR T possess specific antitumor activity. In addition, researchers have successfully engineered protein targeting to glycogen (PTG)-T16R-single chain fragment variable (scFv)-CAR T with immune checkpoint molecules such as PD1, TIM3, TIGIT, TGF-β receptor (TGF-βR), IL-10 receptor (IL-10R), and IL-6R knocked out. These hexavalent inhibitory molecule-knockout CAR T have demonstrated potent anti-CCA immunity and long-term efficacy both in vitro and in vivo [94]. Supimon et al. [95] have devised anti-MUC1-CAR (αM.CAR)/switch receptor (SR), which comprises a MUC1 scFv and a PD1-CD28 SR. These T cells have quickly eliminated CCA cells overexpressing MUC1, with partial mitigation of T-cell exhaustion induced by PD1 upregulation [96]. The fourth-generation CAR T were composed of an anti-MUC1 scFv and three co-stimulatory domains (CD28, CD137, and CD27) associated with CD3ζ. The study successfully demonstrated the specific cytotoxic activity of these CAR T against CCA cells [95]. Another study on antitumor efficacy in ICCA revealed the toxicity of Tn-MUC1 CAR T targeting glycosylated MUC1, both in vitro and in vivo [97]. In a recent study, fourth-generation A20-4G CAR T exhibited remarkable cytotoxicity against CCA cells expressing integrin integrin alphavbeta6/8 (αvβ6) [98]. Furthermore, T cells producing αCD133-αCD3 bispecific T-cell engagers (BiTEs) have been successfully engineered and thus have become highly specific against CD133-positive CCA [99].

CAR T therapy is used not only for primary CCA but also for secondary CCA. To address the unfavorable prognosis of metastatic CCA, researchers have adopted autologous infiltrating lymphocytes. The method has proved to be highly effective. In a particular study, the adoptive transfer of ERBB2IP-reactive CD4⁺ Th1 lymphocytes successfully mitigated the advancement of the disease [100]. White et al. [101] used whole exome and transcriptome sequencing to identify TCR-mediated, MHC II-restricted autologous CD4⁺ Th1 cells within TILs from a patient with metastatic CCA. These cells were able to recognize specific peptides that make up the FGFR2-tudor domain containing 1 (TDRD1) protein breakpoint in cancer cells. The studies presented empirical evidence of the ability of TILs from patients with invasive metastatic CCA to selectively identify peptides originating from FGFR2 fusions. Consequently, the findings of the aforementioned studies corroborate future investigations into adoptive transfer and fusion-reactive T-cell therapy, particularly for the treatment of fusion-driven solid tumors.