Canonical TGF-β/SMAD signalling

Active TGF-β binds to TGF-βRI and TGF-βRII complex followed by TGF-βRI phosphorylation on specific serine and threonine residues (Ref. Reference Groppe22). Then, the receptor complex activates receptor regulated-SMADs (R-SMADs) through C-terminal serine residue phosphorylation (Refs Reference Zhang23, Reference Piek24). Phosphorylated R-SMADs (SMAD2, SMAD3) recruit SMAD4 that is translocated into cell nucleus and binds transcriptional complexes to regulate the transcription of target genes (Ref. Reference Matsuzaki25). After that, the SMAD2/3/4 complex is moved out of the nucleus by activated TGF-β receptors for cyclical use or degraded through ubiquitination (Refs Reference Xiao26, Reference Wang27). Alternatively, the activity of TGF-β signalling can be tempered by the inhibitory activity of SMAD7. Specifically, SMAD7 can block TGF-β-mediated transcriptional events via binding to DNA in the cell nucleus, form a complex with E3- ubiquitin ligases and translocate from cell nucleus to the plasma membrane for TGF-β receptor ubiquitination, or inhibit TGF-β signalling via blocking interactions between R-SMADs and the TGF-β receptor complex (Ref. Reference Hanyu28) (Fig. 1, Table 1).

Fig. 1.

Canonical and non-canonical TGF-β signalling pathways. TGF-β signalling is composed of both canonical and non-canonical pathways. The canonical pathway is SMAD-dependent, where active TGF-β binds heterotetrameric complexes of TGF-β type II (TGF-βRII) and type I (TGF-βRI) receptors on cell surface (with TGF-βRIII being the co-regulatory molecule) followed by activation of Smad2/3 and recruiting of Smad4 for cell nucleus translocation and target gene regulation. The non-canonical pathway is SMAD-independent, where TGF-β activates MAPKs, NF-κB, Ras, TRAF6, TAK1-p38/JNK and PI3K kinases, leading to various biological effects. Although TGF-β-induced PI3K/AKT/mTOR and MAPK-ERK signals are initiated from TGF-β receptor homodimers, TGF-β-induced MAPK-JNK and MAPK-p38 signallings are initiated from TGF-β receptor heterodimers. AKT, protein kinase; ERK, extracellular signal-regulated kinase; IKK-2, IκB kinase 2; ICD, intracellular domain; JAK, Janus kinase; JNK, c-JUN N-terminal kinase; MAPK, mitogen-activated protein kinase; MKK, MAP kinase kinases; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; PI3K, phosphatidylinositol 3-kinase; PS-1, membrane-embedded protease presenilin-1; STAT, signal transducers and activators of transcription; TAK1, TGF-β-activated kinase 1; TRAF6, tumour necrosis factor receptor-associated factor 6; TACE, TNF-α converting enzyme; TGF-β, transforming growth factor-β.

Table 1.

Canonical and non-canonical TGF-β signallings

Non-canonical TGF-β signalling

Although SMADs play a vital role in TGF-β signalling, TGF-β also transduces signals independent of SMADs that involves factors such as phosphoinositide 3-kinase/protein kinase (PI3K)/AKT/mTOR, extracellular signal-regulated kinase (ERK), mitogen-activated protein kinase (MAPK), TGF-β-activated kinase 1 (TAK1) that is an upstream factor of p38 MAPK, c-JUN N-terminal kinase (JNK) MAPK and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signallings, Janus kinase (JAK)/signal transducers and activators of transcription 3 (STAT3) signalling, TGF-βRI intracellular domain signalling. TGF-βRs can trigger PI3K/AKT signalling via directly interacting with PI3K (Ref. Reference Yi, Shin and Arteaga29) that activates mammalian target of rapamycin complex 2 (mTORC2) (Ref. Reference Lamouille30). TGF-βRI can activate the Ras GTPase by phosphorylating the adaptor protein ShcA that leads to the sequential activation of c-Raf, MAP kinase/ERK kinase (MEK) and ERK1/2 kinases (Ref. Reference Lee31). Upon ligand binding, TGF-βRs activate TAK1 via interacting with the tumour necrosis factor receptor-associated factor 6 (TRAF6), an E3 ubiquitin ligase of TAK1 (Ref. Reference Yamashita32); activated TAK1 can initiate p38 MAPK and JNK MAPK signallings via activating MAP kinase kinases (MKKs), where MKK3 and MKK6 are upstream regulators of p38 and MKK4 activates JNK (Ref. Reference Lin33). P38 signalling can also be triggered by TRAF4-mediated TAK1 polyubiquitylation (Ref. Reference Zhang34). Besides, TRAF6-mediated TAK1 polyubiquitylation can also activate NF-κB signalling via activating IκB kinase 2 (IKK-2) (Ref. Reference Hamidi35). Moreover, TGF-β also promotes JAK1/2 phosphorylation that activates STAT3 signalling (Refs Reference Dees38, Reference Tang39). Interestingly, TNF-α converting enzyme (TACE) and membrane-embedded protease presenilin-1 (PS-1) can cleave TGF-βRI and generate an intracellular domain (ICD) that enters the nucleus to mediate gene transcription (Refs Reference Liu36, Reference Gudey37) (Fig. 1, Table 1).

TGF-β in cell proliferation

TGF-β suppresses the growth of normal and premalignant epithelial cells (Ref. Reference Nagaraj and Datta43) (Fig. 2). TGF-β inhibits epithelial cell proliferation by causing cell cycle arrest at the G1 phase that involves three transcriptional events, that is, inducing cyclin-dependent kinase (CDK) inhibitors (Ref. Reference Zhang, Alexander and Wang44), suppressing c-Myc expression (Refs Reference Chen, Kang and Massague45, Reference Seoane46, Reference Seoane, Le and Massague47) and attenuating inhibitor of DNA-binding 1 (ID1), ID2 and ID3 (Refs Reference Je48, Reference Selesniemi, Albers and Brown49). In the cell cycle, the G1/S transition requires the activation of CDKs including CDK2, CDK4 and CDK6, where CDK2 binds cyclin E and CDK4/6 pairs with cyclin D to phosphorylate their substrates (Ref. Reference Zhang, Alexander and Wang44). There are several CDK inhibitors such as P16INK4A, P15INK4B, P21CIP1 and P27KIP1, which control cell cycle progression by pairing with CDK/cyclin complexes (Refs Reference Karimian, Ahmadi and Yousefi50, Reference Abbastabar51, Reference Zhang52). TGF-β downregulates the expression of c-Myc, a known oncogene that enables uncontrolled cell proliferation (Ref. Reference Chen, Kang and Massague45), with the aid of transcription factors P107, E2F4/E2F5 and C/EBP (Refs Reference Chen53, Reference Frederick54). Interestingly, c-Myc inhibition not only directly affects cell cycle progression, but also promotes the expression of several important CDK inhibitors in response to the same TGF-β signal (Refs Reference Seoane46, Reference Seoane, Le and Massague47).

The functionalities of TGF-β on cancer cells change as tumour microenvironment (TME) undergoes drastic changes during cancer progression (Ref. Reference Nagaraj and Datta43). Most cancer cells become insensitive to TGF-β-mediated growth inhibition, which attributes to the loss of function of TGF-βRI/II or SMAD2/4. For example, loss of 18q21 chromosome that harbours SMAD4 was reported in pancreatic and colorectal cancers (Ref. Reference Liu55). Genetic alterations were not observed in some cancers, implicative of other mechanisms for acquiring resistance to TGF-β-mediated growth inhibition (Ref. Reference Papageorgis56). Some tumour cells evade growth restriction by activating Myc/SMAD3 and PI3K/AKT pathways or suppressing p15/p21/WAF1/CIP1 signalling (Ref. Reference Zhou57). The P38/MAPK pathway also participates in the process of circumventing TGF-β-induced growth restrictions, with the specific molecular mechanism remaining to be explored (Ref. Reference Bakkebo58).

TGF-β in programmed cell death

TGF-β and apoptosis. TGF-β has been reported as an apoptosis inducer in most studies (Ref. Reference Zhang, Alexander and Wang44) (Fig. 2). In SMAD-dependent signallings, several apoptotic target genes are regulated by SMAD transcriptional complexes such as death-related protein kinases (DAPK), Src homology inositol phosphatase (SHIP) and TGF-β-induced early gene 1 (TIEG1) (Ref. Reference Yang59). For example, TGF-β upregulates the expression of DAPK towards elevated levels of caspase activity and apoptosis in hepatocytes (Ref. Reference Jang60), enhances SHIP transcription to trigger apoptosis through decreasing Akt phosphorylation in haematopoietic cells (Ref. Reference Valderrama-Carvajal61) and induces TIEG1 expression to promote apoptosis via suppressing Bcl2 in epithelial Mv1Lu cells (Ref. Reference Chalaux62) or inducing oxidative stress in an in vivo mouse model (Ref. Reference Kammoun63). In SMAD-independent pathways, TGF-β-mediated apoptosis involves many other genes, such as Bcl-2 family members, BCL-XL and X-linked inhibitor of apoptosis that are anti-apoptotic, and caspase 3/8 and Bcl-2-interacting killer that are pro-apoptotic (Refs Reference Francis64, Reference Huang65). TGF-β-induced reactive oxygen species (ROS) promote apoptosis via regulating various Bcl-2 family members such as Bcl-2 modifying factor and Bcl-2 interacting mediator in a series of mouse cancer cell lines including AML12 (liver cancer), NMuMG (breast cancer), CMT170 (pulmonary system adenocarcinoma) and A20 (lymphoma) (Ref. Reference Ramjaun66). TGF-β promotes apoptosis via inhibiting the expression of Krüppel-like factor 5 that cooperates with SOX4 during oncogenesis in pancreatic ducal adenocarcinoma cells (Ref. Reference David67).

The effect of TGF-β on tumour growth is double-edged, as it prevents apoptosis under certain conditions despite its functionality as an apoptosis inducer, with the specific mechanism remaining to be investigated (Refs Reference Ehata68, Reference Schlapbach69). In particular, TGF-β induces the expression of Dec1, an apoptosis inhibitor, in mouse mammary carcinoma cells (Ref. Reference Ehata68), and elevates c-FLIP level that prevents Fas-triggered activation of caspase-8 and caspase-3 for cell apoptosis evasion in microglia (Ref. Reference Schlapbach69).

TGF-β and other programmed cell death. Autophagy is an evolutionarily conserved pathway through which cytoplasmic substrates, such as damaged organelles and cytoplasmic proteins, are delivered to lysosomes for degradation (Ref. Reference Wu, Won and Rubinsztein70). It has been reported that TGF-β can activate autophagy in hepatocellular carcinoma and mammary carcinoma cells (Ref. Reference Kiyono71), increase the expression of autophagy-related genes such as Beclin-1, Atg5, Atg7 and DAPK in hepatoma cells (Ref. Reference Suzuki, Kiyono and Miyazono72), trigger autophagosome accumulation and convert microtubule-associated protein 1 light chain 3 (LC3) to its lipidated form, LC3-II (participating in autophagosome formation), in HuH7 human hepatocellular carcinoma cells (Ref. Reference Suzuki, Kiyono and Miyazono72). In addition, silencing SMAD2/3/4 or DAPK, or inhibiting JNK all suppress TGF-β-induced autophagy (Ref. Reference Suzuki, Kiyono and Miyazono72), suggestive of the involvement of both SMAD-dependent and SMAD-independent TGF-β pathways.

Ferroptosis is an iron-dependent form of regulated cell death driven by excessive lipid peroxidation (Ref. Reference Chen73), and TGF-β1 represses SLC7A11 (a marker inhibiting ferroptosis) expression via activating SMAD3 and enhancing lipid peroxidation in hepatocellular carcinoma cells (Ref. Reference Kim74).

Pyroptosis is an inflammatory form of programmed cell death that is critical for amplifying protective immunity during infection (Ref. Reference Jorgensen and Miao75), the impact of which on tumours varies with the type and stage of transformed cells (Ref. Reference Xia76). The roles of TGF-β1 in pyroptosis vary with the cancer type and developmental stage. For instance, TGF-β1 induces pyroptosis in colorectal cancers (Ref. Reference Jiang77) but inhibits pyroptosis in triple negative breast tumours (Ref. Reference Tamura78).

Studies on TGF-β and non-apoptotic cell death are still in the nascent stage. With our incremental knowledge on the roles of TGF-β in controlling the life/death switch of cancer cells, and growing number of novel forms of programmed cell death events being discovered, we foresee increased attention in this field towards improved harness over TGF-β signalling and cancer control.

TGF-β in tumour angiogenesis

Tumour growth and metastasis require a rich blood supply to provide nutrients and oxygen, leading to angiogenesis once the tumour size reaches more than 1–2 mm in diameter (Ref. Reference Carmeliet and Jain86). Tumour cells secrete growth factors such as TGF-β to induce tumour angiogenesis towards accelerated cancer development (Ref. Reference Roberts87) (Fig. 2). Endothelial cells with higher cell proliferation, progression and migration rate play a vital role in angiogenesis (Ref. Reference Tian and Schiemann88). TGF-β induces pro-angiogenic growth factors such as vascular endothelial growth factor (VEGF) and connective tissue growth factor to directly stimulate EC towards tumour angiogenesis and accelerated cancer progression (Refs Reference Shimo89, Reference Fang90). TGF-β can significantly increase the expression of VEGF and promote tumour angiogenesis in RAS-transformed epithelial tumours (Ref. Reference Rak91). TGF-β and VEGF form a feedback loop through the semaphorin 3A/Neuropilin-1-expressing monocytes (NEM) axis to avoid over-activated TGF-β signalling (Ref. Reference Groppa92). Moreover, TGF-β signalling can increase the expression of miR29a and angiogenic regulatory enzymes such as MMPs (MMP-2, MMP-9, MMP-10, etc.) that contribute to neovascularisation, or inhibit tissue inhibitors of metalloproteinases (Refs Reference Wang93, Reference Miyazono, Ehata and Koinuma94). Besides, TGF-β promotes tumour vasculature by elevating the pericyte–endothelium association, where tumour-associated fibroblasts enhance tumour blood capillaries through TGF-β-dependent signalling (Ref. Reference Zonneville95).

TGF-β in tumour-associated inflammation

Tumour-associated inflammation is cancer-promotive (Ref. Reference Candido and Hagemann118) by secreting numerous cytokines including TGF-β (Ref. Reference Hanahan and Weinberg41) (Fig. 2). TGF-β1 inhibits the expansion of CD8⁺ cytotoxic T cells and promotes the development of both pro-inflammatory CD4⁺ T helper (Th) 17 cells and regulatory T cells (Tregs) (Ref. Reference Landskron119).

TGF-β in immune evasion

The immune system is the main defense system that recognises and clears foreign antigens including tumour cells (Ref. Reference Chen and Ten Dijke120). However, tumour cells develop varied strategies to evade the immune surveillance including TGF-β-induced immune suppression (Fig. 2). TGF-β may inhibit the proliferation and functionality of various innate and adaptive immune cells such as T cells, B cells, dendritic cells (DCs), natural killer (NK) cells, macrophages, neutrophils, monocytes and granulocytes (Ref. Reference Sanjabi, Oh and Li121) (Table 2).

Table 2.

Effects of TGF-β on immune cells

DC, dendritic cell; NK, natural killer; TAM, tumour-associated macrophage; TAN, tumour-associated neutrophil; MDSC, myeloid-derived suppressor cell; Th, helper T cell; Treg, regulatory T cell.

In innate immunity, DCs are antigen-presenting cells capable of inducing adaptive T cell response and sensitising the anti-tumour immunity (Ref. Reference Diamond133), where TGF-β leads to systemic immunosuppression by downregulating key factors in DC maturation or activation (Ref. Reference Papaspyridonos125). NK cells suppress tumour progression during the anti-cancer immunity by directly recognising tumour cells and initiating cytotoxic reactions (Ref. Reference Caligiuri134). TGF-β1 suppresses NK cells by reducing the production of interleukin-15 (IL-15) and repressing the expression of natural killer group 2 member D (NKG2D), a major NK cell-activated receptor (Refs Reference Friese128, Reference Lazarova and Steinle129). TGF-β can convert tumour-associated macrophages (TAMs) from the tumour-suppressive (M1) to the cancer-promotive (M2) phenotype (Ref. Reference Gong130) according to markers expressed on cells. Similarly, TGF-β polarises tumour-associated neutrophils (TANs) from a cancer-suppressive N1 to a tumour-promotive N2 phenotype, contributing to tumour growth via immunosuppression (Ref. Reference Fridlender131). TGF-β can lower the production of ROS and nitric oxide (NO) that are essential for the cytotoxicity of macrophages against cancer cells under certain conditions (Ref. Reference Siegert135). In addition to affecting TAM/TAN polarisation, TGF-β regulates immune cell functionalities by, for example, increasing PD-L1 in DCs towards T cell anergy against hepatocellular carcinoma cells (Ref. Reference Song126), and driving myeloid suppressor cell differentiation towards immunosuppression if coupled with IL-6 (Ref. Reference Eriksson127). Besides, TGF-β regulates the generation of natural and induced Tregs by, for example, transforming human thymocytes into natural Treg cells ex vivo in combination with elevated IL-2 secretion and ectopic FOXP3 expression (Ref. Reference Gallego-Valle132).

In adaptive immunity, immune-mediated tumour elimination requires cytotoxic CD8⁺ T cells that kill malignant cells by producing perforin, granzyme A, granzyme B, IFN-c and Fas ligand (FASL) (Ref. Reference Wu and Lanier136), and provides long-term immunity against key tumour epitopes (Ref. Reference Grosser137). TGF-β represses CD8⁺ T cells through both direct inhibition on CD8⁺ T cells and indirect effects on accessory cells such as Tregs that regulate CD8⁺ T cell functions (Ref. Reference Derynck, Turley and Akhurst3). For example, CD4⁺ and CD8⁺ T lymphocytes are more effective in clearing thymoma and melanoma cells in TGF-βRII knockout mice than the wild-type mice (Ref. Reference Gorelik and Flavell138). In addition, TGF-β inhibits the proliferation and activation of CD8⁺ cytotoxic T-cells by regulating the generation of Tregs (Ref. Reference Fu122), suppresses the proliferation of CD4⁺ Th1/Th2 cells through reducing the expression of their master transcription factors (Ref. Reference Ludviksson123) and induces Th17 lineage development via modulating an IL-12 cytokine family member IL-23 (Ref. Reference Mangan124).

Receptor kinase inhibitors

Several TGF-β receptor kinase inhibitors have been designed to interact with the ATP binding site of the catalytic domain and inhibit the ATP-dependent kinase activity of TGF-β receptor towards the blockage of the downstream signalling cascade (Ref. Reference Huynh, Hipolito and Ten Dijke159). Inhibitors including galunisertib (LY2157299) for prostate cancer treatment (NCT02452008, phase II, recruiting, Eli Lilly and Company), LY3200882 for treating solid tumours (NCT02937272, phase I, active, Eli Lilly and Company) and vactosertib (TEW-7197) for curing myeloproliferative neoplasm (NCT04103645, phase II, recruiting, MedPacto) have reached the stage of clinical trials (Refs Reference Jin160, Reference Fujiwara161, Reference Yap162). However, these inhibitors have poor pharmacokinetic and pharmacodynamic properties that do not generate desirable therapeutic effects, as well as poor target (TGF-β receptors) specificity that also inhibit several other TGF-β-related proteins or kinases such as activin, nodal, myostatin and MAPK-p38 (Refs Reference Inman163, Reference Yingling164).

Ligand traps

Ligand traps are soluble and have a high-affinity to TGF-β receptors and, thereby, can block the binding of TGF-β to its receptors (Refs Reference Yang165, Reference Hinck and O'Connor-McCourt166). Ligand traps comprise Fc-stabilised dimers of TGF-ΒRII extracellular domains by design for sequestering TGF-β1 and TGF-β3, which prevent them from binding to TGF-βRII (Ref. Reference Lan167). These agents were firstly used to inhibit tumour dissemination in mice carrying a genetically engineered TGF-β ligand trap (Ref. Reference Yang165). Distinct TGF-β traps have been developed that differ in their ligand binding specificities (Refs Reference Zwaagstra168, Reference Qin169). For example, AVID200 (Formation Biologics), an engineered TGF-β ligand trap that is 1000 times more selective in neutralising TGF-β1 and TGF-β3 than TGF-β2 (Ref. Reference Kim158), has been predominantly applied in haematopoiesis, bone marrow fibrosis and myelofibrosis (Ref. Reference Tremblay and Mascarenhas170). The efficacy of AVID200 is being evaluated in patients carrying malignant solid tumours (NCT03834662, phase I, active but not recruiting, Formation Biologics). M7824 (MSB0011359C, EMD Serono/GlaxoSmithKline/Merck KGaA) is a bifunctional fusion protein targeting PD-L1 and TGF-β that shows encouraging efficacy and manageable tolerability in platinum-treated non-small cell lung cancer (NSCLC) patients in a clinical trial (NCT02517398, phase I, completed, EMD Serono Research & Development Institute, Inc.) (Ref. Reference Strauss171). Luspatercept (Luspatercept-aamt, Acceleron), a recombinant fusion protein that binds TGF-β ligands towards reduced SMAD2/3 signalling, decreases the severity of anaemia in patients with lower-risk myelodysplastic syndromes (Ref. Reference Fenaux172). However, because of the pleiotropic activities of TGF-β and lack of appropriate biomarkers, patient selection criteria and optimal dosing regimens of these ligand traps are yet to be determined (Ref. Reference Kim158).

Neutralising antibodies

Antibodies are used to antagonise TGF-β activity and block the binding of TGF-β ligand to its cognate receptor (Ref. Reference Kim158). Fresolimumab (GC1008, Genzyme) is a human anti-TGFβ IgG4k mAb that shows promising therapeutic efficacy against malignant melanoma and renal cell carcinoma in a clinical trial (NCT00356460, phase I, completed, Sanofi) (Ref. Reference Morris173). LY3022859 is an anti-TGF-βRII mAb that fails in a trial composed of 14 advanced cancer patients (NCT01646203, phase I, completed, Eli Lilly and Company) because of the onset of uncontrolled cytokine-release syndrome at the higher doses examined prior to the observation of any therapeutic benefits (Ref. Reference Tolcher10). There also exist several other antibodies being examined in pre-clinical or clinical trials. For instance, 264RAD (AstraZeneca, pre-clinical) reduces pharyngeal tumour cell growth and metastasis in vivo (Ref. Reference Eberlein174), 1D11 (Genzyme Corp., Sanofi, pre-clinical) suppresses breast cancer metastasis in vivo (Ref. Reference Nam175), the clinical efficacy of SAR439459 is currently being evaluated in malignant solid neoplasm (NCT04729725, phase I, active, Sanofi) (Ref. Reference Bedinger176) and NIS793 is being examined in metastatic pancreatic ductal adenocarcinoma (NCT04390763, phase II, recruiting, Novartis Pharmaceuticals) (Ref. Reference Greco177). Unfortunately, some antibodies show serious side effects such as uncontrolled cytokine-release syndrome at a high dose (Ref. Reference Derynck, Turley and Akhurst3).

Antisense oligonucleotides

ASOs can be used to halt cancer progression by specifically targeting genes that are difficult to inhibit by small molecules or mAbs (Ref. Reference Gleave and Monia178). However, ASO-based therapeutics are normally instable, where nanoparticles and chemical modifications have been developed to improve their delivery and stability (Refs Reference Dias and Stein179, Reference Tibbitt, Dahlman and Langer180). Trabedersen (AP12009, Antisense Pharma GmbH/Isarna) is a TGF-β2-specific phosphorothioate ASO that has been under clinical development for treating patients carrying refractory high-grade glioma, with the median survival after recurrence exceeding that after receiving chemotherapy without observing life-threatening toxicities (Ref. Reference Hau181). A next-generation TGF-β2-selective ASO, ISTH0047 (Isarna Therapeutics GmbH), has been reported to inhibit lung metastasis in syngeneic orthotopic renal cell carcinoma and breast cancer (Ref. Reference Huber-Ruano182).

Combinatorial strategies involving TGF-β and immunotherapy

The suppressive role of TGF-β in human innate and adaptive immune systems has led to the feasibility of combining therapeutics against TGF-β with anticancer immunotherapies (Ref. Reference Arteaga185). For instance, TGF-β attenuates tumour response to PD-L1 blockade by contributing to T cell removal, and drive immune evasion in genetically reconstituted colon cancer metastasis (Refs Reference Mariathasan186, Reference Tauriello187), suggestive of the potential synergy between TGF-β inhibition and immunotherapy. As an investigational product candidate, SRK181-mIgG1 (Scholar Rock), a potent and highly selective inhibitor of the TGF-β1 isoform, is developed to overcome drug resistance to checkpoint inhibitors such as anti-PD-(L)1 mAbs for treating bladder cancers (Ref. Reference Martin188). Another example is the anti-TGFβ1/2 antibody NIS793 that is being tested in a dose-escalation trial together with the anti-PD-1 antibody spartalizumab in patients carrying advanced-stage cancers (NCT02947165, phase I/Ib, completed, Novartis Pharmaceuticals). The safety and efficacy of vactosertib (TEW-7197), an inhibitor of TGF-β/SMAD signalling, is evaluated in combination with durvalumab (Imfinzi, Medimmune, AstraZeneca), a mAbs blocking PD-L1, in a phase Ib/IIa clinical trial comprised of 60 metastatic NSCLC and urothelial carcinoma patients (NCT03732274, phase Ib/IIa, active but not recruiting, MedPacto).

Combinatorial strategies involving TGF-β and oncolytic viruses

By altering the phenotype of immune-excluded or immune-desert tumours towards an immune-infiltrated phenotype, perturbating TGF-β signalling (largely through inhibition) is expected to enhance the therapeutic efficacy of oncolytic viruses in treating solid tumours (Ref. Reference Groeneveldt189).

Herpes simplex virus (HSV) is commonly used in oncolytic virotherapy and has been combined with approaches interrogating TGF-β signalling for cancer treatment. For instance, HSV1716 is combined with TGF-β receptor 1 (TGF-βR1) inhibitor A8301 to treat syngeneic models of murine rhabdomyosarcoma, where TGF-β inhibition augments the immunotherapeutic efficacy of HSV1716 (Ref. Reference Hutzen190). HSV1, combined with TGF-βR1 inhibitors SB431542, LY2109761, LY2157299, shows remarkable efficacy in arresting recurrent glioblastoma via an inhibitor-triggered MAPK/JNK blockage (Ref. Reference Esaki191). On the other hand, HSV synergises with TGF-β in glioblastoma treatment, where TGF-β pretreatment sensitises cancer cells to the oncolytic efficacy of HSV by attenuating the innate immune responses mediated by NK cells and macrophages (Ref. Reference Han192).

Another frequently reported oncolytic virus used to create synergies with modalities modulating TGF-β signalling is adenovirus. A selective expression of a soluble form of TGF-β receptor II fused with Fc (sTGFβRIIFc) is combined with an adenoviral vector (namely Ad.sTbetaRFc) to treat cancers that results in over 85% reduced breast cancer progression in vivo (Ref. Reference Seth193), and significantly suppresses prostate tumour burden and bone metastasis (Ref. Reference Hu194). Another combinatorial strategy that relies on adenovirus-based oncolytic virus expressing sTGFβRIIFc, called TAd.sTβRFc, shows comparable efficacy in inhibiting prostate cancer metastasis with Ad.sTbetaRFc (Ref. Reference Hu194). A clinical trial featuring AdAPT-001 (NCT04673942, phase I, recruiting, EpicentRx), a type 5 adenovirus that carries a TGF-β trap transgene, is ongoing, where AdAPT-001 was shown safe and well-tolerated at clinically relevant doses in tumour-bearing mice (Ref. Reference Larson195).

Combinatorial strategies involving TGF-β and chemo-/radiotherapy

TGF-β signalling promotes acquired chemotherapy resistance as a result of induced EMT in cancer cells (Ref. Reference Zhang196), and plays key roles in p53/ATM-mediated DNA mismatch repair as a protective mechanism against radiation-induced damage (Ref. Reference Ewan197). Thus, TGF-β inhibition has been coupled with chemo-/radiotherapies towards enhanced therapeutic responses and activated TME (Refs Reference Zhang196, Reference Wang198). The TGF-β/SMAD inhibitor LY3200882 has been shown as a safe monotherapy and combinatorial therapy once coupled with gemcitabine and nab-paclitaxel in pancreatic cancer patients (NCT02937272, phase I, active but not recruiting, Eli Lilly and Company) (Ref. Reference Yap162). However, coupling TGF-β inhibitors with radiotherapy does not achieve desirable therapeutic outcome. For example, a clinical trial fails in combining focal radiotherapy and TGF-β blockage via fresolimumab (GC1008, Genzyme) in the treatment of patients carrying metastatic breast cancers despite the observed overall survival extension among patients receiving a high dose fresolimumab (NCT01401062, phase II, completed, Weill Medical College of Cornell University) (Ref. Reference Formenti199).