Next-generation immunotherapy for solid tumors: Combination immunotherapy with crosstalk blockade of TGFβ and PD-11/PD-L1

Abstract

Introduction:

In solid tumor immunotherapy, less than 20% of patients respond to anti-programmed cell death 1 (PD-1)/ programmed cell death 1 ligand 1 (PD-L1) agents. The role of transforming growth factor β (TGFβ) in diverse immunity is well-established; however, systemic blockade of TGFβ is associated with toxicity. Accumulating evidence suggests the role of crosstalk between TGFβ and PD-1/PD-L1 pathways.

Areas covered:

We focus on TGFβ and PD-1/PD-L1 signaling pathway crosstalk and the determinant role of TGFβ in the resistance of immune checkpoint blockade. We provide the rationale for combination anti-TGFβ and anti-PD-1/PD-L1 therapies for solid tumors and discuss the current status of dual blockade therapy in preclinical and clinical studies.

Expert opinion:

The heterogeneity of tumor microenvironment across solid tumors complicates patient selection, treatment regimens, and response and toxicity assessment for investigation of dual blockade agents. However, clinical knowledge from single-agent studies provides infrastructure to translate dual blockade therapies. Dual TGFβ and PD-1/PD-L1 blockade results in enhanced T-cell infiltration into tumors, a primary requisite for successful immunotherapy. A bifunctional fusion protein specifically targets TGFβ in the tumor microenvironment, avoiding systemic toxicity, and prevents interaction of PD-1+ cytotoxic cells with PD-L1+ tumor cells.

1. Introduction

1.1. TGFβ signaling pathway

Transforming growth factor β (TGFβ)—a member of a TGFβ superfamily consisting of more than 30 cytokines [1]—mediates biological processes in tumorigenesis and tumor regression [2]. TGFβ acts as a tumor suppressor during the early stages of cancer, and as a tumor enhancer during the advanced stages of cancers [3]. TGFβ was known to stimulate the epithelial-to-mesenchymal transition, contributing to chemoresistance and metastasis [4]. In the innate immune system, TGFβ forms a complex with transcription factors, including either TGFβ-activated kinase 1 (also known as MAP3K7) binding protein 1 (TAB1), TGFβ-activated kinase 1 (TAK1), or nuclear factor-κB (NF-κB), to alter the function of NK cells, dendritic cells, and macrophages [5]. Tumor-secreted TGFβ negatively regulates the function of CD8+ T cells, CD4+ T cells, and the production of cytotoxic effector cytokines in the adaptive immune response [6].

TGFβ signaling pathway is regulated by TGFβ isoforms 1, 2, and 3 [6]. Among these three isoforms, TGFβ1 is the most abundant and ubiquitously expressed [6]. Under physiological conditions, TGFβ is hindered by the latency-associated peptide and latent TGF-β–binding protein forming a large latent complex. TGFβ is released from its precursor large latent complex by proteolytic cleavage [2]. The bioactive TGFβs subsequently bind to TGFβ receptor II (TGFβRII), followed by the dimerization of TGFβ receptor I (TGFβRI). The ligation of TGFβs and its receptors induces the phosphorylation of the SMAD2/3 protein, which forms a heteromeric complex with SMAD4. The complex SMAD2/3 and SMAD4 proteins are translocated into the nucleus and induce the transcription of TGFβ-related genes. In addition, TGFβ signaling pathway can transmit signaling through other factors, including extracellular signal-regulated kinase (ERK), p38 mitogen-activated protein kinase (p38 MAPK), JUN N-terminal kinase (JNK), NF-κB, TAK1, and tumor necrosis (TNF) receptor-associated factor 4 (TRAF4) or TNF receptor-associated factor (TRAF6) [6].

1.2. PD-1/PD-L1 signaling pathway

Programmed cell death 1 (PD-1; also known as CD279) [7]—expressed during T-cell activation and maturation—plays a substantial role in maintaining peripheral tolerance [8] and countering active signals [9] through TCR and CD28. PD-1 ligates to programmed cell death 1 ligand 1 (PD-L1; also known as CD274 and B7-H1) and programmed cell death 1 ligand 2 (PD-L2; also known as CD273 and B7-DC) [10]. This ligation results in signaling pathways downstream of PD-1 contributing to the regulation of T-cell function and differentiation and the fine-tuning of T-cell fate [11]. In the proximal signaling, PD-1 was found to inhibit the phosphorylation of lymphocyte-specific protein tyrosine kinase (Lck)/ zeta-chain-associated protein kinase 70 (ZAP70) and protein kinase C-theta (PKCθ) [12], and act as the linker for activation of T cells (LAT) docking on the T-cell membrane. PD-1 regulates the distal signaling by suppressing phosphorylation of Akt, Erk, and Ras [13] in a SHP1/2-dependent manner [14], resulting in not only the attenuation of cytotoxic cytokines transcription and cell cycle but also reprogramming of T-cell metabolism [15] and consequently cell exhaustion and apoptosis survival.

Both cancerous cells and pathogens exploit pathways to escape T-cell–mediated immune responses by PD-1 expression [16]. So far, as many as ten transcription factors and chaperones have validated involvement in Pdcd1 promoter, including transcription factors that are both acute–—nuclear factor of activated T cells 1 (NFATc1) [17], activator protein 1 (AP-1) [18], Notch [19], T-box expressed in T cells (T-bet) [20], B lymphocyte-induced maturation protein-1 (Blimp-1) [21], nuclear factor kappa light chain enhancer of activated B cells (NF-κB) [22], B-cell lymphoma (Bcl6) [23]—and chronic—Forkhead box protein O1 (FoxO1) [24], signal transducer and activator of transcription (STAT)3, STAT4, interferon-stimulated gene factor 3 (ISGF3) [25]. Tumor-intrinsic PD-L1 expression is induced by interferons (IFNs) [26], including type I and II IFNs, TNFα and γ-chain cytokines [27]. Upon antigen exposure, PD-1 induction is linked to prolonged T-cell receptor (TCR) stimulation.

2. The clinical efficacy of TGFβ- and PD-1/PD-L1-targeting therapies

2.1. TGFβ-targeting therapies

Among TGFβ-targeting strategies, at the ligand level, clinical studies investigated the inhibitory effect of antisense oligonucleotides (AONs) on TGFβ synthesis (Figure 1). One of the AONs investigated was AP12009 (Trabedersen)—an 18-mer phosphorothioate AON designed to specifically run complementary to the human TGFβ2 mRNA [28] to treat patients with glioblastoma, melanoma, pancreatic and colorectal cancers (NCT00844064, NCT00431561, and NCT00761280) [28,29]. Better tumor reduction was shown with AP12009 compared to standard chemotherapy treatment, with a favorable safety profile [4,29,30]. Two out of 29 patients developed treatment-related adverse events (TRAEs) grade 3 in phase I and II trials of AP12009 [31,32]. The limitations of AONs are its non-specific/off-target effects, RNA-binding affinity, and delivery method to the target tissue [31,33]. Belagenpumatucel-L (LucanixTM) is another allogeneic vaccine [34] investigated in two clinical studies (NCT01058785, NCT00676507) for the treatment of patients with lung cancer. It contains four different TGFβ antisense oligo-expressing plasmid transfected, irradiated NSCLC cell lines, of which two are adenocarcinoma, one is large cell carcinoma, and one is squamous cell carcinoma. A phase I trial [35] (NCT00676507) involving patients with advanced glioma showed better median survival of the responding patients receiving belagenpumatucel-L (78 vs 47 weeks for patients who were treated conventionally). Two follow-up phase II trials involving patients with NSCLC suggested tolerance and survival advantage [36,37]. Additionally, a phase III trial involving patients with stage III/IV NSCLC who had stable disease after frontline therapy did not demonstrate a dramatic increase in survival, indicating belagenpumatucel-L as a maintenance treatment [38]. However, patients who received belagenpumatucel-L and were randomized within 12 weeks of the completion of frontline chemotherapy resulted in an encouraging increase in overall survival, suggesting further and confirmatory studies of this vaccine.

Figure 1. Schematic representation of TGFβ and PD-1/PD-L1 targeting monotherapies in clinical trials.

(A) Anti-TGFβ–signaling drugs are characterized into three approaches: antisense oligonucleotides blocking TGFβ mRNA translation, monoclonal antibodies neutralizing TGFβ molecules, and receptor kinase inhibitors blocking intracellular signals. (B) Anti–PD-1/PD-L1 drugs are classified into two levels: ligand level disrupting PD-L1–PD-1 interaction, and intracellular level suppressing downstream signal transduction of PD-1. Abbreviations: Transforming growth factor β (TGFβ); programmed cell death 1(PD-1); programmed cell death 1 ligand 1 (PD-L1)

Monoclonal antibodies and soluble TGFβ receptor II were employed to neutralize excessive TGFβ from tumors and fibroblasts [4]. Genzyme developed three monoclonal antibodies, including lerdelimumab, metelimumab, and fresolimumab (GC1008); the development of the two former antibodies was stopped in 2005 after an unsuccessful trial [39,40]. A human derivative of the 1D11 antibody, GC1008, can trap all three forms of TGFβ ligands [4,6]. Five clinical trials investigated GC10008 for the treatment of patients with primary brain tumors, pleural malignant mesothelioma, advanced renal cell carcinoma, melanoma, and breast cancer (NCT01472731, NCT01112293, NCT00923169, NCT00356460, NCT01401062) [6]. Phase I trials with fresolimumab showed partial response in patients with malignant melanoma [41]. In addition, fresolimumab with radiotherapy exhibited a better overall survival of patients with breast cancer [6,42]. AVID200 is an innovative, computationally designed receptor that specifically neutralizes TGFβ1/3 but not TGFβ2 [43,44]. AVID200 was found to kill 4T1 tumor cells in a syngeneic triple-negative breast cancer model in a dose-dependent manner, with a better killing efficacy than that of 1D11 neutralizing antibody. No treatment-related cardiotoxicity was recorded at doses up to 30mg/kg in a non-human primate study [44]. An ongoing phase I trial is investigating the safety and efficacy of AVID200 for the treatment of patients with advanced malignant solid tumors [6]. Another agent that is being investigated in an early-stage clinical trial is BETA-PRIME, a TGF-β trap fusion protein.

TGFβRI kinase inhibitors or monoclonal antibodies against TGFβ receptors are an attractive approach to interrupt the intracellular signal transduction of TGFβ [40]. Pfizer developed a TGFβRI kinase inhibitor, called PF-06952229, and this inhibitor is currently in a phase I (NCT03685591) trial to treat patients with advanced breast cancer and prostate cancer [6]. Yan Wu and his group generated a new anti-TGFβRII antibody, called anti-TGFβRII monoclonal antibody (IMC-TR1; LY3022859); its murine derivative exhibited a great tumor reduction in breast and colon cancer models [45]. A phase I trial evaluating the safety and tolerability of IMC-TR1 in patients with advanced solid tumor reported 8 out of 12 patients developing TRAEs, with four patients experiencing TRAEs grade ≥3 [46]. A decoy receptor, TGFRII dominant-negative receptor, is under investigation that binds to TGFβRII and, consequently, prevents the TGFβ intracellular signal [47–49].

2.2. PD-1/PD-L1-targeting monotherapies

PD-1 and PD-L1 immune checkpoint blockade have been revolutionary in immunotherapy, and both PD-1 and PD-L1 inhibitors have shown significant clinical efficacy in treating patients with solid tumors as well as liquid tumors [50].

2.2.1. Anti–PD-L1 therapy

Among PD-L1 inhibitors, atezolizumab is a widely investigated antibody approved by the U.S. Food and Drug Administration (FDA). Atezolizumab’s structure allows it to competitively bind to PD-L1, preventing the interaction of PD-1 and PD-L1 and ultimately preventing T-cell exhaustion [51]. A meta-analysis of 14 clinical trials with more than 3000 patients reported that 6% of patients had a complete response and 16% had a partial response. The median overall survival rate was one year, with manageable toxicities and 0.17% treatment-related death [52].

Another potential anti–PD-L1 monoclonal antibody is avelumab, being studied in more than 200 clinical trials globally. Avelumab is another FDA-approved antibody developed by Merck and Pfizer [52]. The unique feature of avelumab is the interaction of its Fc region with the Fcɣ receptor III (CD16) of NK cells, which mediates the antibody-dependent cell-mediated cytotoxicity (ADCC) [53]. The global and open-label JAVELIN clinical trial program investigated avelumab’s safety and efficacy in 1738 patients across 15 different cancer types. Avelumab showed an overall response rate of 3–56% in treated patients with refractory metastatic breast cancer and Merkel cell carcinoma, respectively [54]. Among the patients, 10.2% experienced grade ≥ 3 TRAEs and 0.2% experienced avelumab-related death [55].

AstraZeneca developed a potent PD-L1 nanobody, called envafolimab or KN035, which is administered by subcutaneous injection. The crystal structure of this single domain antibody KN035 enhances a more durable binding to PD-L1 with high specificity, resulting in tumor regression in the animal model [56]. There have been eight clinical trials evaluating the efficacy, safety, and tolerability profile of envafolimab in patients with solid tumors, including in patients with hepatocellular carcinoma and HER2-positive breast cancer. Among these trials, three trials in the US, China, and Japan reported a manageable side effect and promising anti-tumor efficacy in patients with advanced solid tumors [57–59].

Durvalumab is a human IgG1 monoclonal PD-L1 antibody developed by Medimmune/AstraZeneca that is administered intravenously. The first report of durvalumab for the treatment of patients with solid tumors was in a study of patients with stage III non-small cell lung cancer (NSCLC) [60]. Durvalumab was also investigated for the treatment of patients with other solid tumors, including head and neck squamous carcinoma [61], pleural mesothelioma [62], hepatocellular carcinoma [63], triple-negative breast cancer [61], and urothelial carcinoma [64]. Following the PACIFIC trial results, durvalumab was approved as an adjuvant therapy for patients with stage III, unresectable NSCLC after standard chemoradiotherapy [60]. In combination with either cisplatin or carboplatin and etoposide, durvalumab has been used as a first-line treatment for adults with extensive-stage NSCLC [65].

2.2.2. Anti–PD-1 therapy

In an analysis of 19 randomized clinical trials of patients with solid tumors, both anti–PD-1 and anti–PD-L1 therapies showed an acceptable safety profile. However, the overall survival of patients treated with anti–PD-1 is significantly higher than that of patients treated with anti–PD-L1 [66]. For example, a humanized monoclonal IgG4 antibody for PD-1 developed by BeiGene, called tislelizumab, was modified to minimize the engagement with Fc-γ receptor (FcγR) to specifically avoid macrophage-mediated, antibody-dependent cellular phagocytosis and improve antitumor efficacy [67]. In 2019, tislelizumab was approved for the treatment of patients with relapsed or refractory classical Hodgkin’s lymphoma. Zhou et al. reported that the treatment of tislelizumab prolonged overall survival in patients by 5–7 months, with improved progression-free survival and overall response rates in patients with locally advanced or metastatic NSCLC compared to treatment with docetaxel [68], regardless of PD-L1 expression. Similarly, in another report, tislelizumab was reported to improve survival and increase quality-adjusted life years for pretreated patients with advanced NSCLC compared to treatment with docetaxel [69]. As there are several publications summarizing anti–PD-1therapies, we refer to these publications [70–76].

2.3. Targeting TGFβ in and PD-1/PD-L1 in chimeric antigen receptor (CAR) T-cell therapy

CAR T-cell therapy has shown complete responses in patients with hematological malignancies, but benefit is limited in those with solid tumors due to the immunosuppressive microenvironment [77–80]. Therefore, as one of the most important regulators in TME, TGFβ-targeting has become a potential strategy for improving clinical outcomes in patients treated with CAR T-cell therapy. Recently, multiple preclinical studies have demonstrated enhanced antitumor efficacy of CAR T cells by blocking TGF-β signaling, such as knockout of TGFβRII [81] and overexpression of dominant negative TGFβRII [48]. A recent phase I clinical trial reported prostate-specific membrane antigen (PSMA)–targeting CAR T cells that overexpressed dominant, negative TGFβRII for treating patients with prostate cancer (NCT03089203). Early results of this study showed that PSMA-CAR T cells expressing dominant negative TGFβRII were safe in patients with metastatic castration-resistant prostate cancer. However, mortality was reported in one patient with excessive proliferation of CAR T cells directly related to the site of CAR insertion [82]. Additional clinical trials are investigating the efficacy of TGFβRII knockout (NCT04976218) and the overexpression of dominant, negative TGFβRII (NCT05166070) in CAR T cells to improve therapy for patients with solid tumors. A clinical trial has been initiated that is investigating the use of TGF-β CAR T cells in combination with GPC3 CAR T cells in patients with hepatocellular carcinoma (NCT03198546).

Upregulated PD-1 in CAR T cells following antigen encounter results in CAR T-cell hypofunction [78,83]. Investigating the treatment of patients with refractory/relapsed diffuse large B-cell lymphoma, a phase Ib trial (NCT04381741) showed the feasibility, controllable toxicities, and marked response rate of anti–PD-1 antibody tislelizumab in combination with CD19 CAR T cells co-expressing IL-7 and CCL19. In a first-in-human, phase I study of patients with malignant pleural mesothelioma treated with regionally delivered, mesothelin-targeted CAR T-cell therapy, CAR T-cell therapy followed by pembrolizumab administration showed to be feasible and safe and enhance CAR T-cell persistence [84]. In addition to the ongoing clinical trials investigating combination immunotherapy with CAR T cells and PD-1 blockade antibody (NCT04134325, NCT03874897, NCT04581473, NCT04995003, NCT04577326), phase I clinical trials for solid tumors (NCT04489862, NCT05373147, NCT04503980) are investigating CAR T cells that are modified to constitutively secrete PD-1 nanobodies (αPD-1). Interestingly, a potential therapeutic effect in patients with advanced refractory ovarian cancer was demonstrated in a case report from a clinical trial (NCT03615313) investigating combination αPD-1–mesoCAR T-cell therapy with anti-angiogenic drug apatinib [85].

An alternative strategy to block the PD-1 axis in CAR T cells is to disrupt cell-intrinsic PD-1 expression via CRISPR-Cas9; the safety and efficacy of PD-1 knockout CAR T cells is currently being evaluated for the treatment of patients with cancer. A PD-1 knockout, anti-MUC1, CAR T-cell clinical trial (NCT03525782) in NSCLC reported that all patients’ symptoms improved after infusion and no cytokine release syndrome was noted. Circulating CAR T cells gradually declined after infusion and decreased to approximately 20% in four months after one cycle treatment, indicating the necessity of further cycles. Wang et al. generated PD-1–and TCR-deficient, mesothelin-specific CAR T (MPTK-CAR T) cells and evaluated them in a dose-escalation study in patients with solid tumors (NCT03545815). No dose-limiting toxicity or serious adverse events were observed in any of the 15 patients. However, rather than TCR-negative CAR T cells, TCR-positive CAR T cells were predominantly detected in effusion or peripheral blood of three patients after infusion [86]. Wang et al. further reported that the expansion and persistence of CAR T cells with PD-1 disruption is not improved significantly, even in the setting of natural TCR and lymphodepletion (NCT03747965). PD-1 knockout T cells were susceptible to T-cell exhaustion and lacked long-term durability [87]. Therefore, further, well-designed clinical studies are required to fully understand the therapeutic potential of PD-1 knockout with CAR T-cell therapy. Utilizing PD-1 dominant negative receptor (PD1-DNR) could help T cells block the PD-L1/2 inhibitory signal while maintaining endogenous PD-1 function [83]; a clinical trial of PD1-DNR CAR T cells is ongoing in patients with mesothelioma (NCT04577326). Other combinations of gene and cellular therapies are being explored to promote anti-tumor efficacy in immunosuppressive solid tumors [88].

3. Rationale for combination therapy

3.1. Crosstalk of TGFβ and PD-1/PD-L1 signaling pathway

The molecular mechanisms that regulate PD-1 induction on different T-cell subsets have been elucidated with TGFβ as a key mediator in the regulation of PD-1 expression (Figure 2). In 2014, it was proposed that it is possible that endogenous TGFβ, at least partially, contributes to the increase of PD-1 though TCR activation [89]. By examining SMAD2/3-deficient T cells from SMAD2^−/− and SMAD3^−/− mice, it was found that TGFβ1 enhances antigen-induced PD-1 expression downstream of antigen-presenting via a SMAD3-dependent manner, rather than SMAD2 both in vitro and in TILs in vivo. A similar discovery was illustrated in a transplantation model [90]; this model showed that the blockade of TGFβ downregulated PD-1 and PD-L1 expression in tissue-resident CD8+ lymphocytes, and that neutralizing TGFβRII signal abrogated graft rejection as well as PD-1/PD-L1 antagonization. Beside SMADs-involved canonical pathway, TGFβ also takes part in mediating NF-κB and NFATc1 translocation via TRAF4/6 induced TAK1 activation [91–94], which may impact PD-1 transcription in a SMAD-independent non-canonical manner (Figure 2). These studies highlight the molecular mechanism and functionality of TGFβ-mediated tumor microenvironment remodeling via PD-1. In adoptive T-cell therapy, TGFβ inhibition represses the expression of PD-1 upon antigen-stimulation on CAR-T cells [81,95]. Since TGF-β was identified as a critical indicator of immune checkpoint blockade (ICB) resistance [96], it is believed that targeting TGFβ would not only diminish the inhibitory function of TGFβ itself but also control the hyper-induction of PD-1, ultimately leading to an enhanced cytotoxic T lymphocyte (CTL) function and improved anti-tumor response in ICB treatment.

Figure 2. Interplay of TGFβ and PD-1/PD-L1 signaling upon TCR/CD28 activation in T-cell.

(1) Upon antigen engagement, TCR/CD3 chain is phosphorylated, as well as CD8 and CD28. Lck and Zap70 are recruited to ITAM site and activated, resulting in the phosphorylation of downstream signaling, mainly including PKCθ, PI3K and Ras. Finally, antigen presentation leads to the nuclear translocation of NFATc1 and NF-κB, thereby activating T-cell proliferation and functionality as well as Pdc1 transcription. (2) PD-1/PD-L1 ligation directly suppresses TCR signaling by the attenuation of Lck, PI3K/Akt and Ras/Erk at multiple stages, resulting in inhibition of anti-tumor cytotoxicity, accelerating T-cell exhaustion and apoptosis. Moreover, PD-1 signaling polarizes naïve T-cell differentiation and promotes Treg proliferation, leading to an excessive production of TGFβ. (3) PD-1 signaling enhances the expression of itself by countering the process of FoxO1 phosphorylation via PI3k/Akt targeting, thus preventing FoxO1 sequestration and enabling Pdcd1 transactivation. (4) TGFβ signaling initiates the transcription of Pdcd1 by canonical pathway (SMAD2/3 & SMAD4) and contributes to PD-1 expression via non-canonical pathway (TRAF6-TAK1). All the above shows a sophisticated pattern illustrating that TGFβ and PD-1/PD-L1 signaling promote each other in a positive feedback manner to impair the anti-tumor CTL response. Abbreviations: Transforming growth factor β (TGFβ); programmed cell death 1(PD-1); programmed cell death 1 ligand 1 (PD-L1); major histocompatibility complex (MHC); T-cell receptors (TCR)

Additionally, PD-1/PD-L1 ligation is also involved in TGFβ production. Up-regulated PD-1 signaling upon PD-L1 ligation or TCR engagement promote naïve T-cell polarization to regulatory phenotypes and function, resulting in TGFβ level augmentation [97] (Figure 2). A “PD-1-FoxO1-PD-1” positive feedback loop in CTL during lymphocytic choriomeningitis virus (LCMV) infection was described [24]. First, FoxO1 directly binds to Pdcd1 promotor and promotes the expression of PD-1 in exhausted CD8+ T Cells, indicating FoxO1 role in PD-1 induction for the terminal differentiation of exhausted CTL. Furthermore, impaired mTOR activation results in enhanced FoxO1 activity and blockade of PD-1 in vivo increases mTOR activation. Since PD-L1 triggers PD-1 signaling AKT, mTOR, and S6 activation, it can be concluded that PD-1 maintains FoxO1 activation and facilitates its interaction with the PD-1 promoter by suppression of mTOR/S6 and following FoxO1 phosphorylation. In this model, the level of PD-1 increases even though the nuclear translocation of NFATc1 and NF-kB, which are downstream of TCR-CD28 stimulation, would be proximally inhibited by PD-1/PD-L1.

3.2. Rationale of combination therapy

Multiple preclinical studies have shown that TGFβ in the tumor immune microenvironment (TME) limits the sensitivity of the tumor to ICBs [98]. Vanpouille-Box and his colleagues indicated that TGFβ impaired the CD8+ effector T-cell function, therefore reducing anti-tumor responses induced by radiotherapy. Powles group examined tumors from a phase II trial (IMvigor210) of atezolizumab-treated patients identifying stromal TGFβ as the key to immune response deficiency. An elevated F-TBRS score (gene signature representative of TGFβ signaling in fibroblasts) on 47% of the population with immune exclusion correlated with stable and progressive disease. Using a model that recapitulated the immune desert, the authors found that dual blockade of anti-TGFβ and anti-PDL1 antibodies increased tumor infiltrating CD8+ T cells and reduced tumor burden in EMT6 and MC38 murine models [99]. In alignment with the study, Eduard Batlle and his group mentioned that only a few mice responded to galunisertib after 14 days [100]. They found that the immune resistance was due to the upregulation of PD-1 in galunisertib-activated T cells. Furthermore, anti–PD-L1 treatment alone only showed a limited response in metastatic mouse models. In contrast, co-treatment of galunisertib and anti–PD-L1 antibody enhanced IFNγ levels and T-bet expression in CD4+ T cells, as well as the production of granzyme B in CD8+ effector cells. Based on these results, it was concluded that inhibition of TGFβ induced changes the tumor from that of immune-excluded phenotypes to inflamed phenotypes [98]. Furthermore, Yi Zhang and his group found that myeloid-derived suppressor cells secreted TGFβ, which upregulated PD-1 levels in CD8+ T cells, leading to PD-1/PD-L1 inhibitor resistance [101]. Lastly, a phase II trial in advanced hepatocellular carcinoma reported high baseline plasma TGFβ levels (≥200 pg/mL), associated with poor outcomes in pembrolizumab-treated patients [102].

Growing evidence implicates that the ICB response greatly depends on PD-/PD-L1 levels on immune cells and TGFβ levels on fibroblasts and stromal cells in the TME [98,99], underscoring the need for additional TGFβ inhibition combined with PD-1/PD-L1 blockade to counteract the ICB resistance, thereby improving anti-tumor efficacy.

4. Current landscape of dual blockade of TGFβ and PD-1/PD-L1 therapy on solid tumors

Herein, we describe dual blockade of TGFβ and PD-1/PD-L1 stratified into four groups, which include newly designed fusion proteins, anti-PD-1/PD-L1 monoclonal antibody (mAb) with anti-TGFβ mAb, anti-PD-1/PD-L1 antibody with TGFβR inhibitors, and other approaches (Figure 3).

Figure 3. An overview of dual blockade therapies in current clinical trials.

Combination therapies of TGFβ and PD-1/PD-L1 are under investigation in clinical trials including: TGFβ antibody + PD-1 antibody, TGFβ antibody + PD-L1 antibody, TGFβRI kinase inhibitor + PD-1 antibody, TGFβRI kinase inhibitor + PD-L1 antibody. Abbreviations: Transforming growth factor β (TGFβ); TGFβ receptor (TGFβR); programmed cell death 1(PD-1); programmed cell death 1 ligand 1 (PD-L1)

4.1. Fusion proteins

4.1.1. M7824

A bifunctional fusion protein, M7824, or bintrafusp alfa, was developed by EMD Serono and National Cancer Institute. M7824 acts as a dual trap that simultaneously blocks PD-L1-mediated-tumor cell-intrinsic and alleviates TGFβ-mediated-tumor cell-extrinsic immunosuppressive pathway [103]. M7824 contains a light chain identical to that of avelumab with a heavy chain generated by substituting three amino acids at the constant regions [104,105] (Table 1). M7824 decreased the tumor plasticity and resistance to chemotherapeutics was driven by TGFβ and PD-L1 [106,107].

Table 1.

Overview of preclinical trials of dual blockade anti-PD-1/PD-L1 & anti-TGFβ

Drug	Combination strategy	Tumor cells for in vivo model	References
Fusion protein
M7824	anti-PD-L1 + TGFβRII ecd	Human breast adenocarcinoma: MDA-MB-231	[108]
		Human NSCLC cell lines: H441, H460, HCC4006	[108]
		Murine breast carcinoma: EMT6 & 4T1	[105]
		Human NSCLC cell lines: A549, H460 & HCC4006, PC-9	[107]
		Human bladder cell lines: HTB-4, HTB-1 & HTB-5	[110]
		Murine breast carcinoma: EMT6 & MC38	[109]
		Human prostate carcinoma: PC3	[108]
		Human ovarian carcinoma: SKOV3	[108]
		Human esophageal squamous cell carcinoma: KYSE450, KYSE70, EC109	[101]
SHR-1701	Anti-PD-L1/TGF-βR fusion protein	Murine lung cancer cell: CMT167	[115]
LBL-015	anti-PD-1+TGFβRII fusion protein	Murine colon carcinoma: MC38	[119]
TST005	anti-PD-L1 + TGFβRII ecd	Murine breast carcinoma: EMT6 & MC38	[120]
a-PDL1-TGFβRII ecd	anti-PD-L1 + TGFβRII ecd	Human breast adenocarcinoma: MDA-MB-231	[122]
		Human melanoma tumor cells: A375, SK-MEL-5	[122]
Monotherapy combination
Anti-TGFβ Ab 1D11 & anti-PD-1 Ab BE0146	Anti-TGFβ Ab & anti-PD-1 Ab	Murine colon carcinoma: CT26	[123]
Anti-TGFβ Ab 1D11 & anti-PD-1 Ab RMP1–14		Murine breast carcinoma: 4T1	[124]
Anti-GARP:TGFβ Ab 58A2 & anti-PD-1 Ab RMP1–14	Anti- GARP:TGFβ Ab & anti-PD-1 Ab	Murine colon carcinoma: CT26 & MC38	[127]
Anti-TGFβ Ab 1D11 & anti-PD-L1 Ab Atezolizumab	Anti-TGFβ Ab & anti-PD-L1 Ab	Murine breast carcinoma: EMT6	[125]
		Murine colon carcinoma: MC38	[125]
Anti-TGFβ Ab 1D11 & anti-PD-L1 Ab 6E11	Anti-TGFβ Ab & anti-PD-L1 Ab	Murine breast carcinoma: EMT6	[99]
		Murine colon carcinoma: MC38	[99]
Galunisertib (LY2157299) & anti-PD-1 Ab RMP1–14	TGFβRI inhibitor & anti-PD-1 Ab	Aggressive pancreatic ductal adenocarcinoma (KRASmutTP53mut)	[129]
Vactosertib (TEW-7197) & anti-PD-1 Ab RMP1–14		BrafV600EPten−/−/MAF melanoma fibroblasts	[130]
Tranilast & anti-PD-1 Ab RMP1–14	TGFβ inhibitor & anti-PD-1 Ab	Murine breast carcinoma: 4T1	[132]
		Murine breast carcinoma: E0771	[132]
LY364947 & anti-PD-L1 Ab MIH5	TGFβRI inhibitor & anti-PD-L1 Ab	Murine breast carcinoma: EMT6	[131]
		Murine colon carcinoma: MC38	[131]
Galunisertib (LY2157299) & anti-PD-L1 Ab 10F.9G2		Orthotopic MTO140 isografts	[100]
SB-431542 & DPPA-1	TGFβRI inhibitor & a PD-1/PD-L1 disruptor	Human squamous cell carcinoma: EC109	[101]
rAd.sT-sTGFβRIIFc & anti-PD-1 Ab	TGFβRIIFc expression vector & anti-PD-1 Ab	Murine breast carcinoma: 4T1	[134]
		Murine renal adenocarcinoma: Renca	[134]

Transforming growth factor β (TGFβ); TGFβ receptor (TGFβR); programmed cell death 1(PD-1); programmed cell death 1 ligand 1 (PD-L1)

In vitro, M7824 attenuated the TGFβ-induced mesenchymalization by preventing mesenchymal gene dysregulation in tumor cells and reverting established epithelial-mesenchymal transition (EMT) human carcinoma cells. This result improves cell proliferation and restoration of tumor cell sensitivity to several chemotherapeutic drugs or radiation therapy. Similar to avelumab, M7824 can mediate the lysis of PD-L1-expressing tumor cells via the mechanism of antibody-dependent cellular cytotoxicity by employing NK cells collected from both healthy donors and cancer patients [53,107,108]. These results were confirmed in breast, cervical, and urothelial human tumor cells [105,108–110].

M7824 rescues NK cells from TGFβ-mediated immunosuppression and augment NK activation and its recruitment to the TME, thereby boosting NK-mediated cytotoxicity [105,108,109]. M7824 increased effector and effector memory T-cell densities in the spleen of EMT6-bearing mice and induced an active CD8+ T cell phenotype in the TME in the immunophenotypic analysis.

Data from a phase I, open label, dose-escalation study (NCT02517398) showed that M7824 had a manageable safety in patients with solid tumors [111]. Eighty percent of patients treated with M7824 had advanced NSCLC with an overall response rate (ORR) of 25% (500 mg ORR, 22.5%; 1200 mg ORR, 27.5%) across different PD-L1 levels. Pruritus was the most common TRAE observed in 18.8% of patients, with maculopapular rash observed in 17.5% and decreased appetite observed in 12.5%. Grade ≥3 TRAEs occurred in 20 patients (25.0%). There were no treatment-related deaths reported in this expansion cohort [112].

In a phase I study, 30 Asian patients with esophageal SCC received M7824 at 1200 mg every 2 weeks until disease progression, unacceptable toxicity, or withdrawal. At the time of cutoff in August 2018, the investigator confirmed ORR was 20% (95% CI, 7.7–38.6); the median survival was 11.9 months, 23.3% patients developed TRAEs grade 3 or 4, and there were no treatment related deaths [113].

There are multiple ongoing trials that combine M7824 with another immunotherapy, chemotherapy, or radiation therapy. A phase I/II study is actively recruiting for combined immunotherapy in patients with prostate cancer, with the combined immunotherapy consisting of BN-Brachyury vaccine, M7824, an IL-15 agonist (ALT-803), and an IDO1 inhibitor (Epacadostat) [114]. Another phase I trial (NCT03524170) was designed to combine M7824 and radiation therapy in treating patients with metastatic hormone receptor positive, HER2-negative breast cancer (Table 2).

Table 2.

Overview of clinical trials of dual blockade anti-PD-1/PD-L1 & anti-TGFβ

Drug	Combination Strategy	Type of Cancer	Phase	NCT Number
Fusion protein
M7824	anti–PD-L1 + TGFβRII ecd	Head & Neck	1, 2	NCT04220775
			1, 2	NCT04247282
		Lung	3	NCT03631706
			2	NCT03840902
			1, 2	NCT03840915
			1, 2	NCT04297748
			1, 2	NCT03554473
			2	NCT04560686
			2	NCT04396535
		Breast	1	NCT03620201
			1	NCT03524170
			1	NCT04296942
			1	NCT03579472
			2	NCT04489940
			1	NCT04756505
		Solid Tumors	1	NCT02699515
			1	NCT02517398
			1, 2	NCT04574583
			1, 2	NCT04708470
		Biliary Tract/ Cholangiocarcinoma	2	NCT04727541
		Biliary Tract/ Cholangiocarcinoma/ Gallbladder	2, 3	NCT04066491
			2	NCT03833661
		Cholangiocarcinoma	1	NCT04708067
		Pancreatic	1, 2	NCT03451773
			1, 2	NCT04327986
		Neoplasms/ Skin Neoplasms	N/A	NCT04267861
		Advanced Kaposi Sarcoma	1, 2	NCT04303117
		Prostate	2	NCT03315871
			1, 2	NCT04633252
		Prostate/ Solid Tumors	1, 2	NCT03493945
		Colorectal	1, 2	NCT03436563
			2	NCT04491955
		Cervical/ Anal/ Oropharyngeal/ Vulvar/ Vaginal/ Penile/ Rectal	1, 2	NCT04287868
			1, 2	NCT04432597
		Cervical	2	NCT04246489
			1	NCT04551950
		Cervical/ Oropharyngeal/ Anal/ Vaginal/ Penile	2	NCT03427411
		Urothelial/ Bladder/ Genitourinary/ Urogenital	1	NCT04235777
		Thymic	2	NCT04417660
		Gastric	1, 2	NCT04835896
		Mesothelioma/ Lung	2	NCT05005429
		Ovarian	1	NCT05145569
		Olfactory/ Nasal	2	NCT05012098
		Esophageal	2	NCT04595149
		Breast / Melanoma / Hematopoietic and Lymphoid / Lung	1, 2	NCT04789668
SHR-1701	anti–PD-L1 + TGFβRII ecd	Advanced Malignancies	1	NCT05048134
			1	NCT05198817
		Solid Tumors	1	NCT04324814
			1	NCT03774979
			1	NCT03710265
			1, 2	NCT04856774
			1, 2	NCT04679038
		Solid Tumors/ Lymphoma	1, 2	NCT04407741
		Nasopharyngeal Carcinoma	1	NCT04282070
			1, 2	NCT05020925
		Breast	2	NCT04355858
		Melanoma	2	NCT05106023
		Head & Neck	2	NCT04650633
		Pancreatic	1, 2	NCT04624217
		Lung	2	NCT04937972
			2	NCT04974957
			2	NCT04884009
			2	NCT04580498
			2	NCT05177497
			2	NCT04699968
			3	NCT05132413
			2	NCT04560244
		Rectal	2	NCT05300269
		Colorectal	2, 3	NCT04856787
		Gastric/ Gastroesophageal	2, 3	NCT05149807
			3	NCT04950322
		Cervical	3	NCT05179239
LBL-015	PD-1Ab +TGFβRII bifunctional protein	Solid tumor	1, 2	NCT05107011
TST005	PD-L1 mAb + TGF-β trap	Solid tumor	1	NCT04958434
Monotherapy combination
Vactosertib (TEW-7197) & Pembrolizumab (MK-3475)	TGFβRI inhibitor & anti–PD-1Ab	Colorectal/ Gastric	1, 2	NCT03724851
Vactosertib (TEW-7197) & Pembrolizumab (MK-3475)	TGFβRI inhibitor & anti–PD-1Ab	Lung	2	NCT04515979
Vactosertib (TEW-7197) & Pembrolizumab (MK-3475)	TGFβRI inhibitor & anti–PD-1Ab	Colorectal/ Liver	2	NCT03844750
Vactosertib (TEW-7197) & Durvalumab (MEDI4736)	TGFβRI inhibitor & anti–PD-L1 Ab	Lung	1, 2	NCT03732274
Vactosertib (TEW-7197) & Durvalumab (MEDI4736)	TGFβRI inhibitor & anti–PD-L1 Ab	Stomach	2	NCT04893252
Vactosertib (TEW-7197) & Durvalumab (MEDI4736)	TGFβRI inhibitor & anti–PD-L1 Ab	Urothelial Carcinoma	2	NCT04064190
Galunisertib (LY2157299) & Nivolumab (MDX1106)	TGFβRI inhibitor & anti–PD-1Ab	Solid Tumor/ Lung/ Hepatocellular Carcinoma	1, 2	NCT02423343
Galunisertib (LY2157299) & Durvalumab (MEDI4736)	TGFβRI inhibitor & anti–PD-L1 Ab	Pancreatic	2	NCT02734160
NIS793 & Spartalizumab (PDR001)	Anti-TGFβ Ab & anti–PD-1Ab	Pancreatic	2	NCT04390763
NIS793 & Spartalizumab (PDR001)	Anti-TGFβ Ab & anti–PD-1Ab	Breast/ Lung/ Hepatocellular/ Colorectal/ Pancreatic/ Renal	1	NCT02947165
NIS793 & Tislelizumab (VDT482)	Anti-TGFβ Ab & anti–PD-1Ab	Colorectal	2	NCT04952753
SAR439459 & Cemiplimab (REGN-2810)	Anti-TGFβ Ab & anti–PD-1Ab	Malignant Solid Tumor	1	NCT03192345
			1	NCT04729725
SAR439459 & Atezolizumab (Tecentriq)	Anti-TGFβ Ab & anti–PD-L1 Ab	Liver	1, 2	NCT04524871
LY3200882 & LY3300054	TGFβRI inhibitor & anti–PD-L1 Ab	Solid Tumors	1	NCT02937272
LY3200882 & Pembrolizumab (MK-3475)	TGFβRI inhibitor & anti–PD-1Ab	Advanced Cancer	1, 2	NCT04158700
SRK-181 & anti-PD-1/PD-L1 Ab	TGFβ1 inhibitor & anti-PD-1/PD-L1 Ab	Solid Tumors	1	NCT04291079
BCA101 & Pembrolizumab (MK-3475)	EGFR/TGFB fusion Ab & anti-PD-1	Head & Neck/ Anal/ Colorectal/ Lung/ Ovarian/ Pancreas/ Skin	1	NCT0442954

Transforming growth factor β (TGFβ); TGFβ receptor (TGFβR); programmed cell death 1(PD-1); programmed cell death 1 ligand 1 (PD-L1)

Trial data obtained from Clinical Trials database (https://clinicaltrials.gov/ct2/home).

However, data from a phase III trial (NCT03631706) did not indicate great improvement of M7824 as a first-line treatment drug in participants with PD-L1 expressing advanced NSCLC compared to pembrolizumab. The report showed that in the M7824 group, the median progression-free survival was 7.0 months (vs 11.1 months for pembrolizumab), the median overall survival was 21.1 months (vs 22.1 months for pembrolizumab), and 90 of 151 patients (59.6%) developed serious TRAEs [vs 61 of 152 patients (40.1%) for pembrolizumab] (Table 2). Furthermore, another phase II-III trial (NCT04066491) demonstrated that M7824 did not ameliorate overall survival in chemotherapy- and immunotherapy-naïve participants with locally advanced or metastatic biliary tract cancer in combination with the current standard of care (gemcitabine plus cisplatin) (11.5 months vs 11.5 for M7824 vs placebo, respectively). Fifty-eight of 146 patients (39.7%) developed serious TRAEs when in combination with M7824 (vs 24.2% with placebo).

4.1.2. SHR-1701

Jiangsu HengRui Medicine, a biotech company, has been developing a new fusion protein called SHR-1701. The concept of this fusion protein is akin to that of M7824. SHR-1701 is designed to improve anti-tumor response by concurrently inhibiting TGFβ and PD-L1- mediated immunosuppression. A preclinical study by Yuan Liu’s group found that SHR-1701, not anti-PD-1/PD-L1 monotherapy, could reverse the imbalance of CD8+ T cells and Tregs and elicit predominant protection against tumors [115]. At present, there are more than 20 clinical trials assessing the safety, tolerability, pharmacokinetics, and clinical activity of SHR-1701 for the treatment of patients with advanced solid tumors, breast cancer, and nasopharyngeal carcinoma (Table 2). In a dose escalation phase I study, 49 patients with advanced solid tumors showed an ORR at 17.8% (95% CI, 8.0%−32.1%) with 8 patients achieved PR and 18.4% patients who had TRAEs greater than grade 3 [116]. Data from a multicenter phase I study (NCT03774979.) showed that 27 patients with advanced NSCLC with EGFR mutations had an ORR of 16.7% (95% CI, 4.7%−37.4%) and disease control rate of 50.0% (95% CI, 29.1%−70.9%). Two out of 27 patients (7.4%) developed grade 3 TRAEs (anemia, hypokalemia, and asthenia) [117]. In an expanded study of NTC03774979, as of February 2021, 11 out of 32 patients with squamous cell carcinoma achieved PR and 11 had SD. The ORR was 15.6% (95% CI, 5.3–32.8) and no grade 4 or 5 of TRAES were reported [118]. SHR-1701 showed a manageable safety profile but further investigations are needed.

4.1.3. LBL-015

LBL-015 is a tetravalent bispecific fusion antibody composed of a high affinity anti–PD-1monoclonal antibody and a TGFβRII fusion protein. It was designed as 2:2 antibody-cytokine fusion format using Mab fusion platform enhancing both in vitro and in vivo antitumor efficacy in a dose dependent manner [119]. A multi-center phase I and II study (XXX) was initiated in 2021 to evaluate the safety, tolerability, and pharmacokinetics of LBL-015.

4.1.4. TST005

Bispecific antibody TST005 consists of a truncated extracellular domain of TGFβRII and a humanized anti–PD-L1 IgG1 antibody (AM4B6) with ablated Fc immune effector function. TST1005 has high PL-L1 binding activity which facilitates the TGFβ trap to high PD-L1 tumor cells, thereby minimizing the toxicities of systemic of TGFβ signaling. Following promising preclinical studies, a clinical trial evaluation of TST005 in patients with advanced solid tumors is initiated (NCT04958434) [120].

4.1.5. BR102

Similar to the aforementioned fusion proteins, BR102 comprises an anti–PD-L1 antibody (GS636) that was fused at the C terminus of TGFβRII ectodomain. This structure allows the fusion protein for TGFβ suppression and PD-1/PD-L1 immunosuppression in the TME. It is reported that the N-glycosylation modifications of BR102 were tightly involved in the regulation of proteolytic activity. To avoid the proteolytic degradation, three motifs were mutated. The BR102 mutant is more resistant to proteolytic degradation and showed an additive antitumor effect in a MC38/hPD-L1 bearing model [121].

4.1.6. a-PDL1-TGFβRII ecd

Similar to BR102, Ravi et al. developed Y-traps, which consist of antibodies targeting PD-L1, fused to a TGFβ receptor II ectodomain (a-PDL1-TGFβRII ecd) [122]. They significantly decreased the frequency of Foxp3-expressing Tregs while greatly elevating the percentage of tumor-reactive IFNγ+CD8+ cells; they showed superior antitumor efficacy compared to atezolizumab in melanoma and TNBC tumor xenograft mice models by counteracting TGFβ-mediated Tregs differentiation.

4.2. Combination of anti–PD-1/PD-L1 mAb and anti-TGFβ mAb

Several groups studied the feasibility of antibody combination strategies against solid tumor in mouse models (Table 1, Figure 4); among which, a pan-isoform anti-mouse TGFβ antibody 1D11 was widely used.

Figure 4. Classification of dual anti-PD-1/PD-L1 and anti-TGFβ blockade strategies in both pre-clinal and clinical evaluations.

(1) combination of anti–PD-L1 mAb (6E11, MIH5, 10F.9G2) and anti-TGFβ mAb (1D11). (2) combination of anti–PD-1 mAb (RMP1–14, BE0146, Spartalizumab, Cemiplimab) and anti-TGFβ mAb (1D11, SAR439459, NIS793). (3) Fusion antibody: Anti-PDL1 mAb linked to two extracellular domains of TGFβR II molecules (M7824, SHR-1701). (4) combination of anti–PD-L1 mAb (Durvalumab,178G7, LY3300054) and TGFβRI inhibitor (Vactosertib, LY3200882). (5) combination of anti–PD-1 mAb (Pembrolizumab, RMP1–14) and TGFβR inhibitor (Galunisertib, SRK-181, LY3200882). Abbreviations: Transforming growth factor β (TGFβ); TGFβ receptor (TGFβR); programmed cell death 1(PD-1); programmed cell death 1 ligand 1 (PD-L1); monoclonal antibody (mAb)

Masaki Terabe and co-authors [123] reported that the neutralization of only TGF-β isotypes 1 and 2 would be sufficient to enhance the efficacy of anti–PD-1checkpoint blockade immunotherapy. The lack of response to ICB therapy was associated with active signature of TGFβ signaling in cancer-associated fibroblasts [99]. Dual TGFβ and PD-L1 blocking antibodies facilitated T-cell penetration into tumor tissue and anti-tumor response. Combination of 1D11 and PD-1 blocking mAb RMP1–14 led to improvements in tumor rejection and mouse survival with radiation therapy [124]. The single-cell transcriptomics revealed upregulation of the immune response genes in macrophages and downregulation of extracellular matrix genes in fibroblasts following dual blockade [125].

Two anti-human antibodies-based phase I clinical trials are now under recruitment. The first trial involves the combination of NIS793 (anti-TGFβ mAb) and Spartalizumab (PDR001, anti–PD-1mAb) against multiple solid tumors, including breast, lung, liver, colon, pancreatic and renal tumors (NCT02947165). By December 2020, 60 patients with advanced solid tumors received NIS793 in combination with PDR001 in dose expansion. TRAEs 3/4 were observed in 11% of patients, with rash (3%) being the most common TRAE [126]. The other trial involves the combination of SAR439459 (anti-TGFβ mAb) and Cemiplimab (REGN-2810, anti–PD-1mAb), which also targets malignant solid tumors. A preliminary biomarker and pharmacodynamic of a phase I clinical study (NCT03192345) displayed an increase in T- and NK-cell proliferation and enhanced production of proinflammatory cytokines/chemokines in patients treated with SAR439459 and/or cemiplimab [127].

To be mentioned, Sophie Lucas and her group developed a mAbs 58A2 that binds not TGFβ, but glycoprotein A repetitions predominant (GARP):TGF-β1 complexes. Along with anti–PD-1 clone RMP1–14, this combination exerts anti-tumor activity via multiple mechanisms; not only does it increase T-cell infiltration and function, but also promotes the densification and normalization of intratumoral blood vasculature [128].

4.3. Combinatorial anti-PD-1/PD-L1 antibody and TGFβR inhibitors

Another anti-TGFβ strategy under development is the use of small molecule inhibitors, which specifically block the intracellular signal transduction downstream of TGFβ receptors. Compared to anti-TGFβ antibodies, chemical compounds may have some advantages, including higher permeability and distribution in vivo, lower cost, and easier storage.

Several TGFβR inhibitors that antagonize serine/threonine kinase have been tested so far (Table 1). The first TGFβR inhibitors is galunisertib (LY2157299) [129]. Dual blockade led to a reduction in the neoplastic phenotype, and improved survival in vivo KRAS^mutTP53^mut pancreatic ductal adenocarcinoma model with a significant increase in CD11b/GR1⁺ cells in the spleen [129]. Brent A. H’s lab [130] investigated another TGFβR inhibitor TEW-7197. The delayed inhibition of TGFβ during the PD-1 treatment enhanced tumor regression compared to a continuous co-inhibition of PD-1 and TGFβ. Similarly, LY364947 intraperitoneal injection delays KPC1 pancreatic tumor growth, and improves therapeutic efficacy of anti-PDL1 mAb MIH5 [131]. Interestingly, Tranilast (Rizaben), usually known as an antiallergic drug, was also validated as a TGF-β inhibitor. Tranilast has been shown to normalize the TME and reprogram macrophages by redirecting them to an M1 phenotype with anti–PD-1antibody RMP1–14 against 4T1 and E00771 tumor [132].

Galunisertib is administered with pembrolizumab (MK-3475), durvalumab (MEDI4736), or nivolumab (MDX1106) for immune treatment against lung, hepatocellular carcinoma, and pancreatic tumor, while vactosertib (TEW-7197) is chosen to combine with pembrolizumab (MK-3475) or durvalumab (MEDI4736) for lung, colorectal, gastric, and urothelial carcinoma immune therapy (Table 2). Additionally, the combination of LY3200882 (TGFβ inhibitors) and LY3300054 (anti–PD-L1 antibody) is also under investigation. In a phase I study, 42 patients with recurrent refractory metastatic pancreatic cancer received combination therapy of galunisertib and durvalumab. It is reported that the combination therapy had a manageable safety profile with 3 out of 42 patients having had TRAEs greater than grade 3 [133]. Noticeably, one phase I/II trial using LY3200882 and Pembrolizumab (MK-3475) was withdrawn due to a strategic realignment and is not related to any concerns regarding safety or efficacy (NCT04158700) (Table 2).

4.4. Other combinations

Unique “blockers” to inhibit TGFβ and PD-1/PD-L1 were developed (Table 1). The first blocker is a short peptide “disrupter” [101]. Synergistic treatment with TGFβ receptor inhibitor SB-431542 with the peptide resulted in enhancing the ratio, function, and anti-tumor ability of MAGEA3-specific CD8+ T-cell response in patients with esophageal squamous cell carcinoma. Another approach is using an oncolytic adenovirus vector, which expresses soluble TGFβRII fused with a human IgG Fc fragment (rAd.sT-sTGFbRIIFc). As an oncolytic vector, intratumoral inoculation of this vector inhibits tumor growth and metastasis in vivo by down-regulating metastasis and reducing Th2 cytokines while enhancing Th1 reaction. When combined with anti–PD-1 and anti-CTLA-4 mAb, tumor burden was repressed significantly in 4T1 xenograft mode [134]. In addition, several investigational agents are being explored in combination with chemotherapy [72].

5. Immune-related toxicity in dual blockade of TGFβ & PD-1/PD-L1 therapy

Minimized and controllable toxicity is the first concern in drug evaluation, and TGFβ/PD-1 dual blockade drugs are no exception. First, clinical research based on data from over 800 patients has shown that most of the side effects in animal models that require attention have not been observed in humans. However, the concern remains if the housekeeping functions of TGFβ in mesenchymal tissues and wound repair keep their original level [135,136]. The various side effects occurring during anti-PD-1/PD-L1 treatment are obstacles before developing a successful dual therapy approach. If dual blockade therapy potentiates efficacy, in the event of toxicity, one possible solution is to decrease the dose of each individual drug when in combination as compared to monotherapy.

6. Conclusion

Studies on multiple TGFβ/PD-1 dual blockade therapies show a promising perspective of excellent anti-cancer efficacy across various solid tumor types. All of which contribute to a new view to remodel the “immune-excluded” tumors and improve the response rates of monotherapies. In conclusion, TGFβ/PD-1 dual blockade therapy could advance current immune checkpoint therapy. Further investigations are needed to examine mechanisms involving these two pathways and to understand pharmacokinetic and pharmacodynamic parameters of newly designed dual blockade drugs and the translational aspects in clinical evaluation.

7. Expert opinion

The efficacy of anti-PD-1/PD-L1 therapy is limited to a small cohort of patients with solid tumors, even for those who are responsive to therapy; resistance remains a limitation for solid tumor immunotherapy. Solid tumor–secreted TGFβ promotes an immunosuppressive TME. In addition, a lack of response to ICB therapy has been associated with TGFβ signaling in cancer-associated fibroblasts, due to an induction of T regulatory cells and the exclusion of infiltrated CD8+ T cells, leading to a dense and immunosuppressive stroma. TGFβ and PD-1/PD-L1 signaling synergize each other in a positive feedback manner to impair the anti-tumor CTL response.

Our review provides an overview of available mono- and dual-blockade therapies, as well as a comprehensive summary of crosstalk signaling between TGFβ and PD-1/PD-L1 pathways. The unique yet interdependent pathways of PD-1/PD-L1 and TGFβ immunosuppression provide rationale for utilizing combination therapy. Dual blockade results in enhanced T-cell infiltration, proliferation and rescue from exhaustion, as well as upregulated effector cytokine secretion. A combination of targeted antibodies or a bifunctional fusion protein can avoid potential systematic inhibition of a TGFβ pathway by specifically targeting proximal TGFβ in the TME, whereas PD-1+ cytotoxic cells do not interact with PD-L1+ tumor cells. To preserve the housekeeping functions of TGFβ in normal tissues, it is critical to target anti-TGFβ therapies to be active only in the tumor. Adoptive cell therapy provides such an opportunity, and genetic engineering gives the opportunity to selectively express blocking reagents within the tumor following antigen activation (activation-induced). An alternative approach is to provide rescue function from a decoy receptor expressed on adoptively transferred immune effector cell that can block either TGFβ or PD-L1/2. Employing CRISPR for PD-1, TGFβ, or both within adoptively transferred effector T cells has already shown to be promising in the clinical setting. Secreting TGFβ trapping (TGFβ-trap) molecules is another safe approach that is already in translation. A potential advantage of adoptive cell-limited dual blockade is the ability to enhance co-stimulation in exhaustion-resistant immune effector cells.

Dual blockade therapy in combination with cancer vaccines may be mutually beneficial and exert superior anti-tumor activity in both independent and complementary manners. Cancer vaccination can not only make TME more inflamed and less immunosuppressive by suppressing TGFβ signaling, which renders tumor cells more susceptible to anti–PD-1/PD-L1 therapy, but also collaborate with anti–PD-1/PD-L1 drug to modify the signature of tumor-infiltrating immune cells for better function. However, dual blockade may also strengthen cancer antigen presentation, remarkedly ameliorate activation, infiltration, and cytotoxicity of TILs in both immune-excluded and immune-desert models.

For administering blocking antibodies and fusion proteins, the pharmacokinetics and pharmacodynamic parameters of dual blockade therapy is not well understood. The heterogeneity of TME in each solid tumor combined with alterations in TME following prior lines of therapy complicates patient selection, treatment regimens, and response and toxicity assessment. However, insights into therapy efficacy and guide treatment regimens can be provided by serial evaluation of circulating T-cell phenotype, as well as serum measurement of soluble PD-L1, PD-1 and TGFβ. Preventing effector immune cell exhaustion or blocking immune suppression is useful majorly in tumors with immune effector cell infiltration. Pre-conditioning of the TME with radiation, chemotherapy, or both is being explored to overcome the immune desert that immunotherapy is currently facing. The success of dual blockade therapy in a specific solid tumor may not be immediately translatable to other solid tumors, necessitating further clinical trials that are focused on studying each solid tumor. However, a strong foundation to investigate and translate dual blockade therapies can be provided by the knowledge gained from interpreting the clinical results, toxicity management and correlative studies of large cohorts of patients who received FDA-approved ICB agents. Within the next decade, tumor-targeted, tunable, and selective dual blockade of PD-L1/PD-1 and TGFβ therapies can become an exciting option in the treatment of therapy-resistant solid tumors.

Article highlights.

1. PD-1/PD-L1-targeting monotherapies have achieved promising results in treating patients with selected solid tumors; efficacy is limited to less than 20% of patients, even in those who are responsive to therapy.

2. In addition to suppressing anti-tumor immune responses, TGFβ plays a crucial role in ICB-based therapy.

3. TGFβ and PD-1/PD-L1 signaling synergize each other in a positive feedback manner to impair the anti-tumor CTL response.

4. Dual blockade TGFβ and PD-1/PD-PD-L1 therapy improves anti-tumor efficacy and has manageable treatment-related adverse events.

Abbreviations

TGFβ

Transforming growth factor β

TAB1

TGFβ-activated kinase 1 binding protein 1

TAK1

TGFβ-activated kinase 1

NF-κB

nuclear factor-κB

TGFβR

TGFβ receptor

ERK

extracellular signal-regulated kinase

p38 MAPK

p38 mitogen-activated protein kinase

JNK

JUN N-terminal kinase

TNF

tumor necrosis

TRAF

TNF receptor-associated factor

PD-1

Programmed cell death 1

PD-L1

programmed cell death 1 ligand 1

PD-L2

programmed cell death 1 ligand 2

Lck

lymphocyte-specific protein tyrosine kinase

ZAP70

zeta-chain-associated protein kinase 70

PKCθ

protein kinase C-theta

LAT

linker for activation of T cells

NFATc1

nuclear factor of activated T cells 1

AP-1

activator protein 1

T-bet

T-box expressed in T cells

Blimp-1

B lymphocyte-induced maturation protein-1

NF-κB

nuclear factor kappa light chain enhancer of activated B cells

Bcl6

B-cell lymphoma

FoxO1

Forkhead box protein O1

STAT

signal transducer and activator of transcription

ISGF3

interferon-stimulated gene factor 3

IFNs

interferons

TCR

T-cell receptor

AONs

antisense oligonucleotides

TRAEs

treatment-related adverse events

IMC-TR1

anti-TGFβRII monoclonal antibody

FcγR

Fc-γ receptor

PSMA

prostate-specific membrane antigen

αPD-1

PD-1 nanobodies

GARP

glycoprotein A repetitions predominant

ICB

immune checkpoint blockade

CTL

cytotoxic T lymphocyte

PD1-DNR

PD-1 dominant negative receptor

TIL

tumor-infiltrating lymphocyte

LCMV

lymphocytic choriomeningitis virus

TME

tumor immune microenvironment

EMT

epithelial-mesenchymal transition

ORR

overall response rate