IFNs are cytokines with important roles in antiviral immunity and consist of three major families, designated as type I, II, and III [37]. Whilst type I and type III IFNs predominantly protect the host from pathogens, IFNγ being the sole member of type II IFNs, has additional well-documented functions in cancer immunosurveillance [38–41]. In the tumour microenvironment, sources of IFNγ include tumour-infiltrating NK cells, ILC1s, γδ T cells, CD8⁺ T cells, and Th1 cells [42–44]. IFNγ signaling is mediated by the IFNγ receptor (IFNGR), leading to downstream activation of Janus kinase 1 (JAK1) and JAK2, respectively, and subsequent phosphorylation of the transcription factor signal transducer and activator of transcription 1 (STAT1) [45]. Nuclear translocation of the phosphorylated STAT1 homodimer then ensues, allowing subsequent binding to the corresponding gamma-activated sequence (GAS) DNA element and induction of IFN-stimulated gene (ISG) expression (Figure 2).

IFNγ signaling in cancer cells directly results in apoptosis through the induction of an array of pro-apoptotic genes including absent in melanoma 2 (AIM2) gene, IFN-induced transmembrane protein 2 (IFITM2) gene, inositol hexakisphosphate kinase 2 (IP6K2) gene, ISG12a, ISG54, phorbol-12-myristate-13-acetate-induced protein 1 (NOXA) gene, and neuronal precursor cell-expressed developmentally down-regulated protein 8 (NEDD8) ultimate buster 1 (NUB1), which ultimately results in a reduction in antiapoptotic B-cell lymphoma-2 (BCL-2) proteins and downstream activation of caspase 3 leading to cell death [46, 47], and IFNγ has also been shown to promote cell death via induction of nitric oxide (NO) or reactive oxygen species (ROS)-mediated pathways [48]. Conversely, IFNγ binding to receptors on anti-tumour type 1 immune cells typically results in activation and induction of effector function [49]. Accordingly, in mouse models of colorectal cancer (CRC), heterozygous genetic deletion of IFNγ (Ifng^+/-), homozygous deletion of the corresponding receptor (Ifngr1^-/-), or antibody-mediated neutralisation of IFNγ, increased adenocarcinoma development [24, 41, 50]. In humans, an IFNG-associated type 1 immune signature is associated with improved prognosis in human CRC patients [51], and similarly predicts prolonged patient survival across multiple other cancer types such as breast, liver, and ovarian cancers [52–54]. Mechanistically, IFNγ induces the expression of cytokines IL-1β, IL-6, IL-12, IL-18, IL-23, TNFα, and proinflammatory factors such as ROS and NO by M1 macrophages, thereby contributing to M1 macrophage-mediated anti-tumour immunity [55, 56]. Both IFNγ and TNFα promote M1 macrophage polarisation [57, 58], and induction of M1 macrophages is associated with reduced tumour burden in mouse CRC [59]. Similarly in humans, increased tumour M1 macrophage infiltration is associated with improved survival in multiple cancer types including ovarian [60], lung [61], and hepatocellular carcinoma [62].

Furthermore, IFNγ stimulates major histocompatibility complex class II (MHC-II) expression on professional antigen-presenting cells (pAPCs) such as dendritic cells (DCs) and macrophages through the transcriptional coactivator MHC class II transactivator (CIITA) [63], and together with M1 macrophage-derived IL-12, promotes T cell activation and subsequent polarisation towards a Th1 phenotype [64, 65]. As mentioned, Th1 cells are potent producers of IFNγ thus providing positive feedback, and both Th1 cells and IFNγ play critical roles in CD8⁺ cytotoxic T cell activation, the latter directly while the former indirectly through licensing XC chemokine receptor 1⁺ (XCR1⁺) DCs [43, 66]. XCR1⁺ DCs are specialised in cross-presentation of antigens, which is critical in efficient CD8⁺ T cell activation [67]. Whilst CD8⁺ T cells, like Th1 cells, are potent producers of IFNγ, they additionally exert direct cytotoxicity on tumour cells through Fas ligand (FasL) or perforin and granzyme which can be further enhanced in response to IFNγ signalling [68]. Cytotoxic CD8⁺ T cells have well-established roles in eliminating cancer cells in vivo [69], and are the most potent anti-tumour immune cell type known against human cancer [70].

Various other mechanisms of IFNγ-mediated anti-tumour immunity have been proposed (Figure 2), for example by decreasing cancer stem cells in the 4T1 mouse model of breast cancer [43, 71]. IFNγ induces cancer cell senescence and arrest via activating p16^Ink4a/p19^Arf and p21^Cip1 thereby inhibiting tumorigenesis in models of pancreatic islet cancer and Burkitt lymphoma [72]. Furthermore, in a heterotypic xenograft mouse model of human gallbladder cancer, treatment with IFNγ reduced tumorigenesis by inhibiting angiogenesis. This is via suppression of M2 macrophage vascular endothelial growth factor (VEGF) production and highlights the therapeutic potential of IFNγ [73]. Overall, IFNγ is capable of inducing anti-tumour immunity via direct action on tumour cells or indirectly through stimulating type 1 immunity.

IL-12 is part of the IL-12 family of cytokines which includes IL-23, IL-27, IL-35, and IL-39, each with very distinct functions [74, 75]. For instance, whilst IL-12 is crucial in Th1 polarisation through the induction of STAT4 and T-box expressed in T cells (Tbet) thereby eliciting anti-tumour immunity [65], IL-23 participates in Th17 immune responses [76] while IL-35 is a recently characterised cytokine with immunosuppressive and pro-tumorigenic properties [77]. IL-12 is produced by the pAPCs macrophages and DCs, and signalling is mediated through binding to the IL-12 receptor (IL-12R), a heterodimer consisting of IL-12Rβ1 and IL-12Rβ2, on target cells [78]. Mechanistically, IL-12 exerts potent anti-tumoural functions by coordinating the aforementioned anti-tumour type 1 response and suppresses alternative T cell fates by inhibiting Th2, Th17, and regulatory T (Treg) cell differentiation [34, 35]. IL-12-induced Th1 cells in turn activate NK cells and CD8⁺ T cells, all of which are potent IFNγ producers [79]. IFNγ can further enhance IL-12 production by pAPCs thereby creating a positive feedback loop [80]. Mice genetically deficient in functional IL-12 are more susceptible to skin cancer and sarcoma development [81, 82], while treatment with recombinant IL-12 delayed tumorigenesis, reduced peritoneal seeding, and prolonged survival in mice with metastatic ovarian cancer [83]. IL-12 can also induce IFNγ-independent CRC tumour rejection via activation of CD8⁺ T cells, a process that is dependent on CD4⁺ T cells and granulocyte-macrophage colony-stimulating factor (GM-CSF) [84]. Similarly, in humans, polymorphisms in the IL-12 genes IL-12B (encoding the p40 subunit of IL-12, A > C, rs3212227), are associated with increased risk of breast cancer when harbouring the A allele [odds ratio (OR) = 1.68, 95% confidence interval (CI) 1.09–2.59] [85], while the IL-12A rs568408 GA/AA variant is associated with increased risk of cervical cancer (OR = 1.43, 95% CI 1.06–1.93) [86].

On the other hand, unlike IL-12 and IFNγ with predominant anti-tumoural properties, the role of TNFα in tumorigenesis is context-dependent. Whilst initially discovered as an anti-tumoural cytokine capable of inducing necrosis of sarcomas in mice [87], more recent studies have uncovered pro-tumoural aspects of TNFα [36]. TNFα is produced by macrophages (similar to IL-12), and additionally by NK cells and activated T cells. TNFα elicits antigen-independent killing of tumour cells by CD8⁺ T cells and NK cells [88], and can directly induce target cell death through signalling [89, 90]. Conversely, genetic deficiency of TNFα enhanced tumour burden in a mouse model of skin cancer [91], and in humans elevated TNFα expression is associated with breast cancer recurrence [92] and poor prognosis in ovarian cancer patients [93]. TNFα may promote tumours through various proposed mechanisms. TNFα may promote breast cancer cell stemness through the nuclear factor kappa B (NF-κB) pathway [94, 95], or through inducing dedifferentiation of melanoma cells resulting in downregulation of tumour antigens thereby promoting immunoevasion [96]. Furthermore, TNFα may directly promote Treg activation through the TNF receptor type 2 (TNFR2) thereby suppressing anti-tumour immunity [97].

IL-4, the prototypical type 2 cytokine, exerts its function through binding to IL-4R, which is comprised of the IL-4Rα subunit and the common gamma chain (γc, Figure 3) [128]. Early studies of IL-4 soon after its discovery found that overexpression of IL-4 by tumour cell lines through genetic approaches led to tumour rejection after in vivo implantation, consistent with an anti-tumoural role [129–131]. This was observed in a wide variety of tumour cell line cancer models such as renal cell tumour [129], CRC [131], myeloma, and breast cancer [130]. However, the anti-tumoural functions of IL-4 may only occur at supraphysiological levels, as most subsequent studies found that endogenously expressed IL-4 instead exert a predominantly pro-tumoural role. In a mouse model of colitis-associated cancer, genetic IL-4-deficiency (thus removing endogenous IL-4 signals) reduced tumour burden [132], and genetic deletion of STAT6, the downstream signaling mediator for IL-4, similarly reduced tumorigenesis in a model of adenomatous polyposis coli (APC)-mutation-mediated CRC [133]. IL-4 may promote tumorigenesis through multiple proposed mechanisms, including through the induction of Th2 cells which produce other pro-tumoural cytokines such as IL-13 [24]. IL-4-signalling induces alternative activation of macrophages to an M2 tumour-associated macrophage (TAM) phenotype with well-documented pro-tumoural functions in the tumour microenvironment [59, 134]. Tumour cells express elevated levels of IL-4R, and IL-4 signaling reduces cancer cell apoptosis through the upregulation of anti-apoptotic genes [135] while also directly promoting tumour cell proliferation [136]. Together with IL-13, IL-4 may also promote MDSCs to inhibit anti-tumour immunity as shown recently where MDSC-mediated CD8⁺ T cell suppression is critically dependent on IL-4 and IL-13 signalling [24].

Like IL-4, IL-13-signalling is mediated by STAT6, through binding to IL-13R which is comprised of an IL-4Rα subunit shared with the IL-4R, coupled with an IL-13Rα1 subunit that provides specificity (Figure 3) [128]. IL-13 similarly promotes M2 TAM polarisation and MDSC-mediated T cell suppression thereby contributing to tumorigenesis [29, 59, 134]. In a model of sporadic CRC, genetic IL-13 deficiency significantly reduced tumour burden, and adoptive transfer of IL-13⁺ ILC2s compared to IL-13^– ILC2s led to enhanced MDSC activation and downstream suppression of anti-tumoural IFNγ⁺ CD8⁺ T cells and Th1 cells [24]. Importantly, when culturing MDSCs with ILC2-derived factors which enhance MDSC-mediated T cell suppression, neutralisation of ILC2-derived IL-4 and IL-13 reprogrammed MDSCs to an anti-tumoural phenotype characterised by enhanced CD8⁺ T cell activation and IFNγ⁺ and granzyme expression. Similarly in a study on bladder cancer, detection of IL-13 in patient urine samples correlated with increased ILC2 and MDSCs prevalence and reduced T cells [137]. IL-13 may also directly promote cancer cell proliferation and metastasis through an alternative IL-13Rα2 (Figure 3) for example in pancreatic cancer [138, 139]. In humans, IL-13R expression is increased in a broad range of solid tumours including CRC, glioblastoma, breast cancer, and pancreatic cancer, and is associated with poor prognosis [140].

IL-5 contributes to type 2 immunity mainly through eosinophils. Eosinophil development is IL-5-dependent, and genetic ablation of the IL-5R subunit IL-5Rα in mice results in defective eosinophils [141]. Likewise, human eosinophil progenitors also express IL-5Rα [142]. Furthermore, IL-5 directly recruits and activates mature eosinophils, and both IL-5 and eosinophils have been implicated in cancer [143, 144]. The majority of animal and human studies seem to suggest an anti-tumoural role of IL-5 and eosinophils, suggesting them as the dominant effectors of anti-tumoural responses by type 2 immunity. Eosinophils can directly induce cancer cell lysis through the release of cytotoxic granules in a human CRC cell line model [145], or indirectly through promoting CD8⁺ T cell-mediated tumour rejection in mouse melanoma [146]. In human patients, increased eosinophils have been associated with improved prognosis across several cancer types including CRC, melanoma, and breast cancer [147–150].

Nevertheless, IL-5 may have pro-tumorigenic roles similar to other type 2 cytokines such as IL-4 and IL-13. A recent study found that IL-5-activated eosinophils upregulated glycolysis, which resulted in depletion of glucose in the tumour microenvironment [151]. This resulted in increased cancer metastasis to the lungs, due to impaired NK cell effector function which is critically dependent on glucose. Similarly, others have found IL-5 and eosinophils to promote lung metastasis by CRC MC38 cells, however instead through the recruitment of Tregs in response to the eosinophil-derived chemokine C-C motif chemokine 22 (CCL22) [144]. In this model, genetic deficiency or antibody-mediated neutralisation of IL-5 reduced metastasis, while adoptive transfer of eosinophils enhanced tumour spread, consistent with a pro-tumoural role. Therefore, the role of IL-5 and eosinophils in tumour development is likely context-dependent and may vary depending on the specific cancer subtype. Indeed, other prognostic studies in humans have found eosinophils to be associated with poor survival in patients with Hodgekin’s lymphoma, leukaemia, and cervical cancer [152–154].

IL-9 is produced by ILC2s and Th9 cells, and like IL-5, has been shown to either promote or inhibit tumorigenesis. In nude mice, exogenous IL-9 treatment inhibited gastric cancer growth [155]. Interestingly, in this model, IL-9 treatment is associated with a reduction in serum IL-4, indicating potential opposing functions. In a mouse model of colitis-associated cancer, IL-9 suppressed tumour growth through CD8⁺ T cell activation [156]. Similarly, findings were observed in melanoma where IL-9 supports anti-tumoural CD4⁺CD8⁺ double-positive T cells directly inducing their proliferation and preventing apoptosis via signaling through the IL-9R [157]. Overexpression of IL-9 by the CT26 CRC cell line genetically led to enhanced CD4⁺ and CD8⁺ T cell recruitment resulting in heightened serum INFγ and tumour lysis [158]. Therefore, although considered as part of type 2 immunity, studies are consistent with IL-9 inducing anti-tumour immunity through cross-activating type 1 immune INFγ producing cells such as CD8⁺ T cells and Th1 cells [155, 158]. Conversely, some studies have found that IL-9 may instead promote tumorigenesis as seen in a heterotypic CT26 implant model of CRC where genetic IL-9 deficiency reduced tumour growth [159]. A potential explanation may be that IL-9 is also a prototypical mast cell growth factor and may exert tumour-promoting properties through mast cells [160]. Mast cells are associated with poor prognosis in a broad range of human cancer types, including melanoma, pancreatic, and prostate cancer [161–163], and can promote cancer angiogenesis through tryptase-mediated extracellular matrix remodelling [164]. Therefore, the role of IL-9 in tumorigenesis is likely dependent on the specific immune cell landscape of the tumour microenvironment depending on the cancer type.

IL-25 and IL-33 are well-established type 2 immune-related cytokines with recently discovered roles in cancer immunity. Both are potent inducers of type 2 immunity with some degree of overlap in function, despite belonging to distinct cytokine families [165]. Specifically, IL-33 and IL-25 directly activate ILC2s and stimulate the release of type 2 cytokines, and the differential sensitivities of tissue-specific ILC2s to these cytokines underlie their functional differences and relative importance in inducing type 2 immunity at different organs [166]. For example, ILC2s in the intestines express high levels of IL-25R, while expression of IL-33R component suppression of tumorigenicity 2 (ST2) is minimal [167]. IL-33 also has additional roles in stimulating Tregs and Th17 cells, the latter likely in a mast cell-dependent manner [168–171]. These may underlie the differential roles of IL-33 and IL-25 in different cancer types depending on the tissue site.

The roles of IL-33 and IL-25 in CRC have recently been reviewed [31]. Briefly, IL-25 is produced by rare tuft cells in the intestines and may promote CRC tumorigenesis through directly inducing tumour stemness or indirectly via stimulation of ILC2s which activates MDSCs [24, 172, 173]. Critically, therapeutic blockade of the IL-25-ILC2-MDSC axis increased IFNγ⁺ expressing CD8⁺ T cells and Th1 cells and significantly reduced CRC burden. In human CRC patients, increased tumour IL-25 expression is associated with poor prognosis, indicating that blocking IL-25 signalling may be a potential novel therapeutic option. Conversely, others have found that IL-25 treatment can reduce subcutaneous tumour growth across a broad range of implanted cancer cell lines in immunodeficient mice [174], most likely through directly inducing apoptosis as shown in breast cancer cells [175]. Therefore, any therapeutic attempts utilising IL-25 to treat cancer will need to be directed to tumour cells and must be used with caution given the potential to elicit pro-tumoural immune responses in the tumour microenvironment.

IL-33 may similarly promote or inhibit cancer in a context-dependent manner. Genetic deficiency of ST2 enhanced NK cell effector function resulting in reduced 4T1 breast cancer growth and metastasis in mice [176]. Others have found IL-33 to stimulate ILC2-mediated NK cell suppression in lung cancer [151]. Similarly, IL-33 administration increased tumour-infiltrating IL-13⁺ ILC2s, MDSCs, and Tregs, and correlated with reduced NK cell cytotoxicity and accelerated tumour growth and metastasis to the lung and liver [177]. Mechanistically, IL-33 can promote metastasis via the induction of desmoplastic reactions by activating cancer-associated fibroblasts [178]. Furthermore, both IL-33 and IL-25 have been shown to promote angiogenesis which also contributes to metastasis, and genetic deletion of ST2 reduced VEGF expression and tumour burden in mouse breast cancer [179–181]. IL-33 can also promote tumorigenesis through stimulating mast cells and Tregs, as seen in mouse models of CRC [182, 183].

Currently, there are no clinical trials related to IL-33 or IL-25 for cancer treatment. Nevertheless, given their overarching effects on tumour immunity, acting upstream of the main effector type 2 cytokines IL-4, IL-5, IL-9, and IL-13 and T cells, the prospect of therapeutically targeting these cytokines is promising. For example, MDSCs have been shown to exert therapeutic resistance across all cancer treatment modalities including surgery, chemotherapy, radiotherapy, and immunotherapy across a broad range of cancer types [184–188]. Combination therapies through concomitant IL-25-signalling blockade may effectively deplete MDSCs by inhibiting the IL-25-ILC2-MDSC axis thereby potentiating existing cancer therapies [24].