Immune-checkpoint blockade (ICB) combined with molecular targeted therapy is an emerging trend in anticancer treatment. The blockade of inhibitory immune checkpoints, including programmed cell death-1 (PD-1) and its ligand programmed cell death ligand-1 (PD-L1), lymphocyte activation gene-3 (LAG-3), and cytotoxic T-lymphocyte–associated antigen 4 (CTLA-4), can reactivate the body’s immune system, thereby preventing the escape of tumor cells (Wei et al., 2018; He and Xu, 2020). In contrast, molecular targeted therapy reverses the malignant biological behaviors of tumors by blocking the tumorigenic components of cells at the molecular level (Druker, 2003). At present, it has been clinically found that the combined application of ICB and anti-angiogenic drugs or targeted therapy have synergistic effects in malignancy, but the molecular mechanism of this combination remains unclear.

The representative ICB therapy, pembrolizumab, is a PD-1 inhibitor, which has been approved for the treatment of colorectal cancer (Powles et al., 2020), non-small cell lung cancer (NSCLC) (Mok et al., 2019) (Abbreviations and full names of tumors and targets was shown in Supplementary Table 1) and other cancer types. Lenvatinib is a novel molecular targeted therapeutic drug that can inhibit a variety of tyrosine kinases, including vascular endothelial growth factor receptor (VEGFRs) 1–3, fibroblast growth factor receptor (FGFRs) 1–4, platelet derived growth factor receptor (PDGFR), stem cell factor receptor (c-kit) and rearranged during transfection (RET) proto-oncogene (Suyama and Iwase, 2018; Al-Salama et al., 2019). It exhibits potent antitumor, antiangiogenic, and immunomodulatory effects. Lenvatinib has been approved for unresectable liver hepatocellular cancer (LIHC) (Kudo et al., 2018), thyroid cancer (Schlumberger et al., 2015), and advanced renal cell carcinoma (RCC) (Motzer et al., 2015).

Nonetheless, the low remission rates with ICB monotherapies and the development of drug resistance with molecular targeted therapies have urged researchers to explore the potential of combination therapies. Data from previous studies showed that the median overall survival (mOS) of patients with advanced LIHC was 12.9 months with pembrolizumab monotherapy (Zhu et al., 2018) and 13.6 months with lenvatinib monotherapy (Kudo et al., 2018), whereas the mOS of patients treated with pembrolizumab and lenvatinib combination therapy reached 22.0 months (Finn et al., 2020). However, the therapeutic effect of ICB combined with lenvatinib is still poorly understood in some cancer types, and few investigations have addressed the mechanisms underlying the synergy of this combination.

Therefore, in this study, we first conducted a meta-analysis to analyze the efficacy of ICB and lenvatinib combination therapy. Bioinformatics analyses were subsequently performed to comprehensively analyze the synergistic mechanisms of ICB combined with lenvatinib and potential efficacies at the transcriptome (RNA-seq), genome (gene mutations, copy-number alterations [CNA]), protein, and immune cell levels.

Lenvatinib is a multitargeting kinase inhibitor that effectively inhibits the activities of a broad spectrum of targets (Supplementary Table 3). It also exhibits unique characteristics in terms of its IC₅₀ value (Supplementary Table 4; Wang et al., 2004; Tian et al., 2011; Stjepanovic and Capdevila, 2014; Kudo, 2018; Lin et al., 2018; Xie et al., 2018) and type-V target-binding mode (Supplementary Table 5; Wang and Long, 2011; Cong-min et al., 2013; Watanabe Miyano et al., 2016; Mori et al., 2017; Roskoski, 2017; Apsel Winger et al., 2019; Kinoshita, 2019). These properties enable lenvatinib to significantly inhibit tumor growth in many types of malignant cancers. Clinical trials of lenvatinib are currently underway (Supplementary Table 6). ICB combined with antiangiogenic therapy is one of the most promising immunotherapeutic combinations. Findings from the present meta-analysis confirm that PD-L1/PD-1 inhibitors combined with lenvatinib result in high ORR, and DCR across patients with multiple different cancer types. However, the complementary and synergistic mechanisms between the two therapy types are currently poorly understood. In the present study we analyzed this topic in detail from a bioinformatics perspective.

The overexpression of lenvatinib target genes and inhibitory immune-checkpoint genes, the positive correlation between the expression levels of these genes, and extensive interactions among the target proteins have suggested the potential for combination therapy (Patel and Minn, 2018; Li et al., 2021). In our expression analysis, the target genes of both ICB and lenvatinib were highly expressed in CHOL, GBM, HNSC, KIRC, LIHC, SKCM, and STAD. This was consistent with the findings from our meta-analysis, which showed that the combination of ICB and lenvatinib achieved better therapeutic effects than either therapy alone in these cancers (Figure 1B). In PAAD and THYM, the target genes of both therapies were not highly expressed, suggesting that combination therapy might not be effective against these cancers, however, this still needs to be clarified in future clinical studies. Combined with a target gene correlation analysis, we found the lenvatinib target genes VEGFR1, VEGFR3, FGFR3, FGFR4, PDGFRA, and the ICB target gene PD-1 were all highly expressed in LIHC, and a relatively strong positive correlation was observed between PD-1 and FGFR3, FGFR4, or PDGFRA (r ≥ 0.15) (Supplementary Figure 3). Accordingly, in clinical trials conducted in unresectable LIHC, the efficacy of combination therapy was superior to that of monotherapy (Figure 1B; Kudo et al., 2018; Zhu et al., 2018; Finn et al., 2020). In THCA, there was an overall strong negative correlation between PD-1/PD-L1 and lenvatinib target genes. Data from different clinical trials in thyroid cancer have shown comparable ORRs with combination therapy (62.0%) and lenvatinib monotherapy (64.8%) (Figure 1B; Schlumberger et al., 2015; Haugen et al., 2020). However, whether lenvatinib monotherapy is inferior to combination therapy in THCA requires further verification in phase III randomized clinical trials. Therefore, it is evident that the expression levels and correlations among lenvatinib and ICB target genes and proteins are closely linked to clinical efficacy. Further studies are required to investigate whether these highly expressed and highly correlated targets can be used as biomarkers for predicting clinical efficacy.

Immune-checkpoint blockade and lenvatinib target genes in tumor-promoting pathways exhibit extensive driver alterations, which may explain the improved efficacy of combination therapy in across multiple cancer types (Le et al., 2015; Bielski et al., 2018). The total mutational loads of the lenvatinib and ICB target genes in pan-cancer are highly consistent with the tumor mutation burden and the degrees of immunogenicity of different types of cancer (Lawrence et al., 2013). The increased production of tumor neoantigens resulting from targeted therapy, together with the high mutational load, further increases the immunogenicity of a tumor. This promotes the extensive infiltration of immune cells and increases immune cell activity, both of which are key factors that enhance the clinical benefit of ICB (Cristescu et al., 2018; Chan et al., 2019). Therefore, cancers with a high number of driver alterations in lenvatinib target genes and a high total mutational load in lenvatinib and ICB target genes may be candidates for combination therapy, e.g., in SKCM, UCEC, LUSC, BLCA, LUAD, STAD, UCS, SARC, ESCA, OV, GBM, HNSC, and CHOL. Interestingly, the total mutational load of LIHC was not high, but lenvatinib targets exhibited extensive driver alterations and were generally highly expressed. Additionally, PD-1 displayed relatively strong positive correlations with FGFR3 and FGFR4 (r ≥ 0.15). These findings are consistent with the unsatisfactory therapeutic effects of single-agent immunotherapy and the enhanced efficacy of combination therapy observed in LIHC (Kudo et al., 2018; Zhu et al., 2018; Finn et al., 2020). In addition, PD-L1 and VEGFR2 in GBM, CTLA4 and PDGFRA in SKCM, as well as PD-L1 and FGFR2 in STAD were all highly expressed genes with driver alterations (Supplementary Table 7). These data may help to inform future explorations of dual-target therapeutic strategies or drug development. However, gene-expression levels, the presence of driver alterations, and the mutational load cannot precisely predict the efficacy of combination therapies for all cancers, as this is also affected by other factors, such as vascular anomalies (Tian et al., 2017), the abundance of immune cell infiltrates in the TME (Roelands et al., 2020), and the heterogeneity of tumor metabolism (Bergers and Fendt, 2021).

Pathway-enrichment analysis showed that the ICB and lenvatinib target genes were implicated in multiple pathways associated with tumor development and progression, angiogenesis, immunotherapy, and tumor metabolism. Additionally, complex interactions among the pathways were observed. The lenvatinib target genes were significantly enriched in multiple pathways associated with tumor development and progression, such as the RTK–RAS–MAPK signaling pathway, the PI-3 kinase signaling pathway, the JAK–STAT signaling pathway, the cell cycle pathway, and the FGFR–FGF signaling pathway. Blockade of these target genes should inhibit tumor cell proliferation or induce tumor cell apoptosis. Because the lenvatinib target genes are associated with signal-transduction pathways downstream of the TCR [e.g., the MAPK–MEK pathway (Ebert et al., 2016) and the PI3K-AKT-mTORpathway (Pollizzi and Powell, 2015)], targeted therapy can also modulate T cells and other immune cells. With regarding to pathways related to angiogenesis, the triple inhibition of VEGFR, FGFR, and PDGFR can inhibit angiogenesis and increase vascular permeability, which rebuilds the TME and increases the local infiltration of immune cells (Naoum et al., 2018; Cabanillas et al., 2019; Yi et al., 2019). Furthermore, the inhibition of FGFR1–4 can help overcome resistance to antiangiogenic drugs (Lee and Secord, 2014; Zheng et al., 2016). Among the immune-related pathways, T cell activation was regulated by KIT, PD-1, PD-L1, LAG3, and CTLA4; cell adhesion was upregulated by VEGFR2, RET, PD-1, PD-L1, and CTLA4; and leukocyte proliferation was regulated by KIT, PD-L1, and CTLA4 (red circles in Figure 4A). For tumor metabolism-related pathways, ICB and lenvatinib can inhibit tumor cell proliferation by downregulating glycolytic enzymes and reducing mTOR activity in tumor cells (Chang et al., 2015). It is evident that the antitumor effect of ICB combined with lenvatinib is modulated by several targets across multiple pathways.

In this study, we also explored the impact of ICB and lenvatinib target genes on the tumor immune microenvironment at the immune cell level. In most cancer types, the expression levels of lenvatinib target genes generally showed a negative correlation with the level of CD8⁺ T cell infiltration, suggesting that lenvatinib treatment may increase the degree of infiltration of CD8⁺ T cells. In ACC, KIRP, LIHC, OV, and THCA, the expression levels of lenvatinib target genes generally showed a negative correlation with MDSC infiltration levels, suggesting that lenvatinib treatment may reduce the degree of infiltration of immunosuppressive cells. The combined use of lenvatinib and ICB to promote CD8⁺ T cell infiltration and inhibit the number and function of Tregs, M2 macrophages, or MDSCs (in terms of IL10 or TGF-β secretion) has been confirmed in murine models of LIHC (Kimura et al., 2018; Kudo, 2019), COAD (Kato et al., 2019), and THCA (Gunda et al., 2019). Lenvatinib can also reduce tumor PD-L1 level to improve anti-PD-1 efficacy by blocking FGFR4 (Yi et al., 2021). The mechanisms of action of lenvatinib include (Figure 6): downregulating the expression of PD-1, CTLA-4, and TIM3 in T cells; blocking the binding of VEGFA and bFGF to their receptors, thereby inhibiting T-cell exhaustion (Deng et al., 2020); and destroying tumor cells, leading to the release of tumor antigens, which in turn promotes IFN-γ-mediated antigen presentation (Kato et al., 2019) and enhances the cytotoxicity of natural killer cells (Zhang et al., 2019). Lenvatinib also impedes angiogenesis and vascular anomalies by inhibiting the secretion of angiogenic factors, such as VEGF, FGF, and PDGF, thereby exerting an immunomodulatory effect (Matsuki et al., 2018). In summary, the synergistic antitumor effect of combined lenvatinib and ICB therapy is achieved by reversing the suppression of the body’s immune functions, increasing tumor immunogenicity, promoting tumor vascular normalization, as well as interact at the transcriptome, genome, and protein levels.

There were certain limitations to our study. Firstly, even though our study covered majority of cancer types, number of tumor types were excluded such as most hematologic cancers. Secondly, although our analyses included relatively large sample sets, larger sample sets are required for further exploration.

This meta-analysis confirmed that ICB combined with lenvatinib exerted greater clinical benefits than monotherapy, across multiple cancers. A multi-omics approach further identified complex and extensive networks connecting the ICB and lenvatinib targets by performing analyses at the gene, protein, and cellular levels, as well as correlation analysis between tumor and immune micro-environment. Most targets were mutually promotional and synergistic, but some of them displayed antagonistic effects. However, the relationships between some targets remained unclear. We found nine tumor types including LIHC, STAD, KIRC, UCEC, LUAD, SKCM, HNSC, CHOL, and GBM with ≥2 positive treatment-related molecular biological characteristics. These findings provide insights for the development of targeted therapy–immunotherapy combinations, and for guiding individualized precision-medicine approaches for cancer treatment.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

YL, QL, QD, and JJ conceptualized and designed the manuscript. YL, QD, YW, MW, MH, HL, and QL acquired, analyzed, and interpreted the data. YL wrote and edited the manuscript. YL, JJ, QD, YW, MW, MH, HL, and QL reviewed and revised the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.