2.1 Components of the TME

Single-cell-based technologies have enabled a better dissection of the TME through precise monitoring of cell sub-populations and spatial architecture, thus revealing the heterogeneous and complex nature of the TME [10, 12]. Other than the cancer cells that form the bulk of the tumor, the other predominant populations include the immune cells constituting the tumor-associated macrophages (TAMs), natural killer (NK) and dendritic cells (DCs), and T and B lymphocytes. The blood and lymphatic ECs, complex collagen fibers and glycoproteins form the ECM, while the cancer-associated fibroblasts (CAFs) and mesenchymal stem cells further assist in ECM remodeling and even chemoresistance [13, 14]. The unique signatures of its cellular components, the associated signaling and the diversity of the TME have been targeted in cancer therapy.

The contribution of the immune system in the modulation of cancer has recently gained importance. Other than tumor heterogeneity, the ‘immune system’ forms a crucial aspect of the complex architecture of the TME, and their modulation can be leveraged to overcome the persistent problems of therapy failure and resistance. Cytotoxic CD8⁺ T cells infiltrate in the TME to get primed by antigen-presenting cells (APCs), macrophages, B cells and DCs to modulate cytotoxic effector T (T_eff) cell response. For instance, DCs secrete the chemoattractants chemokine (C-X-C motif) ligand 9 (CXCL9) and 10 (CXCL10) to facilitate the infiltration of CD8⁺ T_eff cells in the TME and for T cell cytotoxic activity [15]. Instead, their engagement in inhibitory crosstalk, such as with the PD-1/ programmed death-ligand 1 (PD-L1) signaling axis, leads to immunosuppression in the TME [16]. Other immune cells facilitating a pro-tumorigenic response in the TME include the immunomodulatory regulatory T (T_reg) cells and the myeloid suppressor (MDSCs) that result in an immunosuppressive TME [17] and therapy failure. However, the presence of NK cells in the TME is believed to be anti-tumorigenic, resulting from the release of cytokines and chemoattractants such as X-C motif chemokine ligand 1 (XCL1), chemokine ligand 5 (CCL5) or Fms-related tyrosine kinase 3 ligand (FLT3LG) and leading to APC accumulation in the TME [18-20].

CAFs were initially considered a homogeneous population. However, recent studies indicated that CAFs consist of several types of stromal cells that differ in their origin, functions, number, and phenotype [21-24]. Thus, CAFs can either lead to cancer progression or inhibit cancer growth, depending on their nature. They secrete several growth factors, such as fibroblast growth factor (FGF) and C-X-C motif chemokine ligand 12 (CXCL12), to promote angiogenesis via the VEGF, stromal cell-derived factor 1 (SDF-1) and TGF-ß signaling [25, 26]. CAFs recruit and polarize immune cells such as macrophages, neutrophils, T cells and DCs to a pro-tumorigenic phenotype by secreting several cytokines, chemokines, and other effector molecules such as Interleukin 6 (IL-6), and 8 (IL-8), TGF-β, CXCL12, CCL2, SDF-1, VEGF, Indoleamine-pyrrole 2,3-dioxygenase (IDO) and tryptophan 2,3-dioxygenase (TDO2) [27]. However, certain CAFs (Slit Guidance Ligand 2 [Slit2]⁺ and cluster of differentiation 146 [CD146]⁺ CAFs) were shown to have anti-tumorigenic effects such as tumor suppression and increased tumor chemosensitivity [28]. Furthermore, downregulating the paracrine signaling of fibroblasts, such as the platelet-derived growth factor (PDGF)/ PDGF receptor (PDGFR) signaling pathway and hepatocyte growth factor (HGF)/ mesenchymal-epithelial transition factor (MET) signaling pathway, was shown to promote chemosensitivity in CAFs [29, 30].

The TME modulates ECs to induce an angiogenic response because of the high nutritional demand of the tumor [31]. Hypoxia-inducible factor 1-alpha (HIF-1α) activation in tumor cells promotes the secretion of proangiogenic factors, such as VEGF, FGF-2, and PDGF, and stimulates angiogenesis [32, 33]. ECs are targeted indirectly for tumor therapy by inhibiting angiogenesis via neutralizing antibodies for VEGF or inhibitors of VEGF activity [34]. In a clinical trial, bevacizumab, a monoclonal antibody against circulating VEGF-A, combined with chemotherapy was shown to improve the overall and progression-free survival of colorectal cancer patients compared with chemotherapy alone [35]. However, tumor cells become resistant after long-term treatment, and these anti-angiogenic drugs were shown to promote vasoinvasion leading to metastasis and reduced lifespan in mice [36]. Thus, it suggests the need for targeting the multiple oncogenic interactions in the TME rather than individual cell types and molecules.

ECM acts as a scaffold and plays a tumor-suppressing role in healthy tissues; however, it is modified in tumor tissue to possess a tumor-promoting role [37]. The ECM components underlying tumor-promoting activity, such as fibronectin and its splice variants, crosslinked collagen I and tenascin-C, are induced in the TME [37, 38] and interact with integrins to influence tumor cell migration, proliferation and cellular signaling [39]. Various strategies have been developed to target aberrant ECM components to develop novel treatments, including fresolimumab (to inhibit collagen synthesis), collagenases and matrix metalloproteinases (MMPs) (to promote collagen degradation), 4-methylumbelliferone (to inhibit hyaluronic acid synthesis), hyaluronidase (to promote hyaluronic acid degradation), Vitaxin and Volociximab (to target integrin and inhibit angiogenesis) [40]. Moreover, Provenzano et al. found that systemic administration of pegvorhyaluronidase alfa (PEGPH20), an enzyme against hyaluronic acid, reduced stromal hyaluronic acid, normalized interstitial fluid pressures, re-expanded the microvasculature and led to tumor suppression in pancreatic ductal adenocarcinomas murine models [41].

Thus, different components of the TME secrete growth factors, components of the ECM, cytokines and extracellular molecules that are essential for signaling between cells in the TME and systemically. Thus, it would be critical to identify strategies for identifying key vulnerabilities and targeting them to alleviate the immune suppression prevalent in most TMEs.

2.2 TME-based cancer therapy

In recent times, combining therapies, especially those inducing the engagement of immune cells, have been the primary focus of TME-based therapeutics. In this section, we review several facets of the TME that have been targeted for therapy. Table 1 shows the clinical trials targeting different components of the TME.

Tumors have a high demand for rich vasculature to keep up with the high nutrient and glucose demand for their progression. However, most tumors are poorly vascularized and hypoxic, which advances cancer progression and chemoresistance [70, 71]. Additionally, poor tumor vasculature increases interstitial fluid pressure, protecting the tumor core from cancer therapeutics via the bloodstream [72, 73]. To block neoangiogenesis and mitigate hypoxia, VEGF antagonists have been used as a line of therapy [74]. However, its success depends on other microenvironmental factors. For example, anti-VEGF treatment is ineffective in obese mice due to increased IL-6 and FGF-2 expression by the adipocytes in the TME, while co-targeting these improve the anti-VEGF therapy response [75].

Tumor-specific ECM varies among different cancer types. This specificity is exploited for therapies. For example, breast cancer samples have a remarkably high fibrillar collagen content, leading to a poor prognosis [76, 77]. Thus, probes targeted to tumor-specific collagen help detect tumors and micrometastases [78]. Conjugation or recombinant fusion of therapeutic agents such as cetuximab or lumican to a collagen-binding domain peptide can increase efficacy and safety [79, 80]. Similar to collagen, TME also displays a unique matrix of fibrin and fibronectin. Specific antibodies (such as L19) against fibronectin to improve tumor response in the clinical trials of patients with glioblastoma by localizing interleukin 2 (IL-2) or interleukin 12 (IL-12) in the tumor, leading to increased infiltration of cytotoxic T cells [81-83].

The abundance of TAMs in TME can be therapeutically harnessed if they can be polarized to their anti-tumor phenotype. Histone deacetylase inhibitor, TMP195, polarizes TAMs toward an anti-tumor phenotype and increases the efficacy of carboplatin, paclitaxel, and anti-PD-1 therapies [84]. Moreover, the phagocytic role of TAMs can be harnessed to concentrate cytotoxic drugs in the TME [85, 86]. Miller et al. found a reduction in liver metastases due to enhanced delivery of nano-encapsulated platinum in the TME via TAMs in a breast cancer mouse model [87]. On a similar note, immunosuppressive myeloid cells can be differentiated/polarized for anti-tumor phenotypes [88, 89]. For this, myeloid progenitor differentiation has been manipulated via a bone marrow homing nanoparticle therapeutically containing the immunostimulatory muramyl tripeptide. This peptide shows anti-tumor effects by epigenetic reprograming of the multipotent progenitor cells in the bone marrow, which overcomes the immunosuppressive TME [55]. Moreover, a prodrug of 6-diazo-5-oxo-l-norleucine (DON) that blocks glutamine metabolism in myeloid precursor cells can differentiate monocytes into anti-tumor TAMs, leading to tumor regression in mouse models [56].

Stimulating a patient's dominant immune system can have long-lasting effects against cancer. Targeting the immunosuppressive TGF-β signaling in T cells by PLGA nanoparticles enhances antitumor immunity in mice. These nanoparticles bind to T cells and release a TGF-βR1 inhibitor, SD-208, that stimulates the potency of CD8⁺ T_eff cells while inhibiting the inhibitory Treg cells in the TME [90]. CAR-T cell therapy by engineering T cells to express synthetic receptors that recognize tumor-associated antigens in a major histocompatibility complex (MHC)-independent way has shown huge success in hematological malignancies. Although CAR-T cell therapy in solid tumors is challenging since CAR-T cells cannot penetrate solid tumors through vascular endothelium, prostate-specific membrane antigen (PSMA)-directed CAR-T cells have shown success in clinical trials against prostate cancer [57]. CD133-CAR-T therapy in colorectal, pancreatic and hepatocellular carcinoma has shown anti-metastatic potential in a phase I clinical trial [91]. Moreover, therapeutic cancer vaccines have shown promising results in cancer immunotherapy by amplifying tumor-specific T-cell responses. They can be categorized as cellular, viral, or molecular vaccines [92]. Viral vaccines using a heterologous prime-boost strategy to amplify T-cell responses have been successful in prostate cancer. Here, the delivery of a tumor antigen by a viral vector is boosted by a subsequent delivery of the same antigen by another vector [93]. Also, peptide-based vaccines that deliver long immunogenic antigens to DCs have been shown to induce CD4⁺ T and CD8⁺ T cell cytotoxic activities and improve survival in patients with HPV-16-positive cervical cancer and newly diagnosed glioblastoma (NCT02455557) [59].

DCs form a nexus between the adaptive and innate immune responses. DCs present antigen to T cells along with upregulation of co-stimulatory molecules and cytokine production. The inactivity of DCs in the TME to perform these functions hampers immune response to tumors. DC deficiency in the TME can be caused by several factors. Tumor-derived exosomes are known to inhibit DC differentiation by releasing immunosuppressive factors such as IL-6 and TGF-β [94, 95]. Moreover, high lipid accumulation in DCs facilitated by tumor cells can also decrease the secretion profile and reduce antigen-presenting capacity, and high amounts of hyaluronic acid in the ECM affect DC maturation [96]. Also, hypoxia in the TME inhibits DC maturation and function by inducing VEGF signaling caused by the binding of hypoxia-induced VEGF to its receptors on DC membranes [97]. Combinational DC-based therapy, such as DC-based vaccine or granulocyte-macrophage colony-stimulating factor (GM-CSF), which stimulates DC differentiation, activation and migration, along with immune-checkpoint blockade, has shown success in clinics [98]. Immune checkpoint blockers pembrolizumab and nivolumab are among the most frequently used blockers, along with US FDA-approved PD-1/PD-L1 inhibitors [96].

CAFs are unique in expressing their cell surface markers. For example, depleting fibroblast activation protein (FAP)⁺ CAFs via a FAP vaccine decreases collagen density and improves chemoresistance in mouse models [99]. However, CAF depletion has also been shown to reduce infiltrating immune cells, leading to tumor progression [100]. CAF depletion in the TME leads to a shift from T helper 2 (Th2) to T helper 1 (Th1) to promote inflammation, accompanied by the up-regulation of IL-2 and IL-7 and downregulation of TAMs, Tregs and MDSCs [101]. Moreover, CAF depletion using diphtheria toxin-based immunotherapy reduced cancer growth by increasing CD8⁺ T_eff-cell infiltration to the TME [102]. Another approach to target CAFs is to inhibit their tumor-promoting functions. For example, CAFs are known to activate the transcription of homeobox (HOX) transcript antisense RNA through paracrine TGF-β1, which leads to epithelial-mesenchymal transition and thus promotes breast cancer metastasis. Moreover, inhibiting TGF-β1 was found to significantly inhibit CAF-induced tumor growth and lung metastasis in an MDA-MB-231 orthotopic tumor transplantation nude mouse model [103]. Furthermore, inhibiting the proliferation of CAFs by co-administering IPI-926, which inhibits the hedgehog signaling, with the chemotherapeutic drug gemcitabine increasing tumor drug sensitivity in the mouse models of pancreatic ductal adenocarcinoma [104].

Most current therapies directed against the TME target TAMs, tumor vasculature, DCs, ECM, T cells and CAFs. As each of these cell types functions uniquely to modulate the TME, it is important to analyze them and identify critical nodes that could be targeted to inhibit TME support to tumor cells.

3.1 Protein kinase C (PKC) signaling

The extent to which PKC isoform activation or inactivation affects TME components, including stroma and immune systems, determines their promotion or suppressor functions on tumor growth. PKC is a family of structurally related serine/threonine kinases that functions as the transducer of signals from a variety of molecules ranging from hormones (adrenaline, angiotensin), growth factors (insulin, epidermal growth factor), cytokines (Tumour necrosis factor α (TNF-α), IL-1β and IL-6) and neurotransmitters (i.e., dopamine, endorphins) to regulate cell survival, proliferation, differentiation, apoptosis, adhesion and malignant transformation [105, 110, 111]. The binding of ligands to their receptors can activate phospholipase C, leading to the upregulation of cytosolic concentrations of the activators of PKC signaling, diacylglycerol (DAG) and Ca⁺⁺ [112, 113]. The activation of PKC can upregulate several molecular pathways, including Akt, signal transducer and activator of transcription 3 (STAT3), nuclear factor-κB (NF-κB) and apoptotic pathways, to regulate tumorigenesis and metastasis [112]. PKC alpha, a PKC isoform, showed antitumor activity by inducing the polarization of TAMs within the TME [114]. Additionally, the protein levels of PKC alpha, beta and epsilon were found to be downregulated in cancers such as colon cancer [115, 116]. PKC theta, another isoform of PKC, showed tumor-suppressive effects by inducing immune suppression within the TME by controlling CTLA4-mediated regulatory T-cell function [117-119]. Conversely, phorbol esters, the naturally occurring activating ligands of PKC, showed tumor-promoting functions, suggesting that PKC could be an oncogene [111, 120, 121]. Moreover, within the TME, PKC beta, another PKC isoform, is a well-documented effector of the VEGF signaling that promotes angiogenesis and is required for invasiveness in certain tumors, such as pancreatic tumors [122-124]. Therefore, it is likely that different isoforms of PKCs or the same isoform within different contexts may act as tumor promotors or tumor suppressors in a context-dependent manner.

Nonetheless, combining existing therapies with novel molecules to modulate the dysregulated PKC signaling in cancer could be promising. Bryostatins, which are PKC activators, were shown to protect against phorbol ester-induced tumors [125]. Epoxytiglianes, another class of PKC activators, showed efficacy in preclinical mouse models and clinical mast cell tumors in canine models [126-128]. Another activator of PKC, tigilanol tiglate, has been approved for use in canine mast cell tumors [129]. Besides activators, CGP 41251, an inhibitor of PKC, has also shown anti-tumor activity and was found to reverse multidrug resistance when combined with adriamycin [130].

Although preclinical activation has led to the identification of complex PKC functions, their translation in clinical trials is impeded by a lack of mechanistic insights and robust pathological markers. This understanding is required to clinically address the action or inaction of the isoforms and to reveal the gain or loss of function of isoforms required for optimum efficiency of therapeutic interventions.

3.2 Notch signaling

Recent evidence indicates that a distinct population within tumors can express distinct Notch ligands or paralogues, which can both activate and inhibit tumor development in different cancers since the outcome of Notch signaling is highly changeable depending on the context. For instance, the presence of Notch ligand delta-like canonical Notch ligand 1 (DLL1) on DCs interacts with the NOTCH2 receptor on tumor cells to promote DC immunosuppressive function. However, jagged canonical Notch ligand 2 (JAG2) on DCs plays negative tumor-promoting roles Moreover, Notch mutations have been suggested to serve as predictive biomarkers for immune checkpoint therapy in various cancers. Thus, although not yet clinically successful, an integrative analysis with newer perspectives holds the potential for clinical developments [131].

Notch signaling plays a key role in determining cell fate and regulating embryonic as well as tumor angiogenesis [132, 133]. Notch receptors are heterodimeric, single-pass transmembrane receptors that interact with either one of their membrane-bound ligands (Jagged1, Jagged2, and Delta-like ligands Dll1, Dll3 and Dll4) to stimulate the expression of target genes [133]. In solid tumors’ TME, the activation of Notch signaling generally promotes oncogenesis [134]. For example, Notch ligand Dll4 is upregulated in tumor samples from clear cell renal cell carcinoma patients [135], and inhibiting Dll4 leads to the disruption of tumor vasculature within the TME [136, 137]. However, Notch signaling functions as a tumor suppressor in the malignancies of myeloid origin, such as acute myelogenous leukemia and chronic myeloid leukemia [138, 139]. Moreover, a chronic blockade of Notch signaling leads to vascular tumors of the liver, skin, ovary, testes and colon in mice [140, 141].

Demcizumab and enoticumab (Dll4 targeting antibodies) alone or in combination with existing antitumor drugs, and Rovalpituzumab (anti-Dll3 targeting antibody) work by suppressing cancer stem cells and angiogenesis in the TME and have progressed to randomized phase II trials [142-144]. Brontictuzumab (anti-Notch1 receptor antibody) and Tarextumab (anti-Notch2/3 receptor antibody) were tested in relapsed or refractory tumors, small-cell lung cancer and pancreatic ductal adenocarcinoma, but clinical trials were discontinued owing to a low response rate in patients [145-147]. MEDI3622 (anti-TACE antibody) showed promising preclinical activity in human colorectal adenocarcinoma progression [148, 149]. Inhibiting γ-secretase using small molecules like PF-03084014 and BMS-906024 has shown promising results in triple-negative breast cancer, desmoid tumors and pancreatic ductal adenocarcinoma, and related phase II and III trials are currently underway [150-153]. Small molecule and peptide inhibitors of the intracellular domain of the Notch receptor (NICD)-transcriptional complex assembly have been developed, and phase I/IIa clinical trials are being conducted on one such inhibitor, CB-103 (NCT03422679). Other categories of Notch inhibitors, such as IMR-1 and PRI-724, have been found to be effective in inhibiting the Notch-transcriptional complex in in vitro model of triple-negative breast cancer [154].

Although extensively studied in the past several decades, Notch signaling therapeutics have failed clinical expectations. The major shortcomings are high cytotoxicity, shown by pan-NOTCH inhibitors, and low affinity of antibody-drug conjugates (ADC). To combat these issues, it was suggested to investigate novel isoform-specific drugs with high ADC affinity. Moreover, combination therapies targeting chemoresistance, endocrine resistance and radio-resistance also hold promise.

3.3 TGF-β signaling

TGF-β therapeutics have recently shown great promise with the use of TGF-β-neutralizing antibodies and ligand traps, which inhibit the binding of TGF-β with its receptors [155]. Moreover, dosing strategies to bypass cellular toxicity or specifically target TGF-β isoforms that are maximally linked to cancer progression hold promise in clinics to avoid the cytotoxicity of TGF-β inhibitors [155].

Transforming growth factor-β (TGF-β) is a family of cytokines that intricately regulate embryonic development, tissue homeostasis and regeneration [156]. Moreover, they regulate various aspects of cancer cells, such as adhesion, differentiation, cell cycle progression and apoptosis [157, 158]. TGF-β signaling plays a biphasic role in TME and cancer progression. It acts as a tumor suppressor in the initial stages of malignancies by suppressing cell proliferation and inducing apoptosis [159]. However, cancer cells adapt to the protective TGF-β signaling and utilize its moonlighting functions to create a conducive TME by activating CAFs, promoting angiogenesis and ECM production, and suppressing anti-tumor immune responses [160-162].

Various strategies have been adopted to target the deregulated TGF-β signaling, including neutralizing antibodies to target either ligands or receptors, ligand traps, small-molecule inhibitors, and antisense oligonucleotides (ASOs). Fresolimumab, a human monoclonal antibody against TGF-β in a phase 1 clinical trial (NCT00356460), demonstrated preliminary evidence of anti-tumor activity in malignant melanoma and renal cell carcinoma [42]. LY3022859, an antibody against TGF-β receptor 2, showed survival benefits in mouse models [163]. TGF-β ligand traps are chimeric fusion proteins designed to prevent TGF-β from binding to its receptors. AVID200, a ligand trap for TGF-β1 and TGF-β3, showed anti-tumor activity in mouse models [164] and feasibility in clinics in patients with advanced solid-state tumors in a phase 1 clinical trial (NCT03834662) [165]. Moreover, various small molecule inhibitors of TGF-β, including galunisertib (LY2157299) [166-168], vactosertib (TEW-7197) [169-171] and LY3200882 [172], have shown specific antitumor activity. Lastly, phase 1 clinical studies with trabedersen (AP12009), Lucanix (belagenpumatucel-L), or with ASOs against TGF-β2 mRNA, showed better survival in glioblastoma, melanoma, pancreatic cancer or colorectal cancer patients [173, 174].

The pleiotropic effects of TGF-β in tissue homeostasis have rendered exploiting their pro-tumorigenic properties difficult, as targeting TGF-β can systemically affect healthy tissues and tumor cells, thus safety concerns. Thus, a better understanding of the molecular mechanisms of TGF-β in their regulation of normal and cancerous cells is required. Moreover, stratifying patients based on biomarkers who may benefit from TGF-β targeting is essential.

3.4 ER stress response pathways

The presence of hypoxia [175, 176], reduced nutrient availability [177, 178], accumulation of reactive oxygen species [179, 180] and a decreased pH within the TME [180, 181] contribute to a chronic upregulation of ER stress in the TME, impacting the fate and survival of cancer cells. Additionally, oncogenic transformation contributes to the constitutive activation of ER stress sensors such as inositol-requiring enzyme 1α (IRE1α), leading to persistent activation of ER stress by cancer cells [182]. Such chronic activation of ER stress modulates the TME by reducing the surface expression of major histocompatibility complex class 1 and 2 proteins, thereby impeding immune recognition of the cancer cells by NK cells [183, 184]. Therefore, reducing the ER stress load could be utilized to target TME for more favorable treatment outcomes.

Therapies modulating ER stress response pathways by inhibiting IRE1α, PRKR-like ER kinase (PERK) and molecular chaperone binding-immunoglobulin protein (BiP) have shown success in various preclinical and clinical studies. IRE1α inhibitors KIRA8 or AMG-18, STF083010, MKC8866 and MKC3946 reduced tumor growth in various cancers, including multiple myeloma [185-187], melanoma [188], breast cancer and prostate cancer [189]. Similarly, PERK inhibitors GSK2606414 and GSK2656157 possess antitumor activity and can reactivate T cell function in the TME in mouse embryonic fibroblasts; however, these compounds had adverse effects in mouse models, hindering their progress toward clinical trials [190-193]. Further, suppression of BiP signaling by KP1339 (also known as IT-139) or HA15 in glioblastoma, bladder or breast cancer cell lines has been shown to enhance the response to anti-cancer therapy [194, 195].

3.5 Modulation of the TME by lactate bioavailability

Targeting lactate transporters, such as solute carrier family 16 member 1 (SLC16A1) and member 7 (SLC16A7), to reduce lactate levels in tumors has shown huge preclinical success. Another approach to inhibit the conversion of pyruvate to lactate through targeting lactate dehydrogenase A (LDHA) is useful in inhibiting the oncometabolic lactase functions and revoking T cell- and NK cell-mediated immunosuppression in various cancers [108].

Due to elevated glucose uptake rates, cancer tissues have high levels of metabolic by-products like lactate that can be partly attributed to the accelerated metabolism of cancer stem cells and other constituents of the TME [2, 196]. Recent reports suggest the role of proton-coupled lactate efflux in maintaining an acidic phenotype in the TME, thereby promoting angiogenesis [197, 198], cell invasion [198] and metastasis [199-202]. Additionally, high lactate concentration in the TME inhibits the maturation of monocytes to dendritic cells [203, 204] and reduces cytokine production and cytotoxic activity by T cells and NK cells [205-207], thereby contributing to suppressed immune recognition of cancer cells.

Suppression of lactate efflux by cancer cells using the small molecule inhibitor AZ3965, α-cyano-4-hydroxycinnamate (CHC), has shown promising results in preclinical models of Burkitt lymphoma, breast cancer, gastric cancer, small cell lung cancer and glioblastoma [208, 209]. Moreover, pharmacological inhibition of lactate dehydrogenase-A, a key gene involved in lactate synthesis using N-hydroxyindoles and galloflavin or genetic ablation of LDHA, reduced tumorigenesis in non-small cell lung cancer[210], pancreatic ductal adenocarcinoma [211, 212] and cervical cancer cells [212, 213].

Despite understanding the pathogenic roles of lactate, how its targeting affects host anti-tumor immunity and synergizes with other anti-cancer immune therapies still needs to be explored since lactate also partly modulates the metabolism in the TME as an energy source, signaling molecule, and as a key tumor immunosuppressive factor [202]. Moreover, identifying more potent lactate transporter and LDHA inhibitors warrants improving anti-cancer therapies.

3.6 Metabolic reprogramming of the TME

Metabolic reprogramming is an essential hallmark of cancer characterized by the ability of cancer cells to reprogram the metabolism of non-cancerous cells, specifically immune cells, to go against their nature, help in tumor progression, and better adapt to the limited availability of nutrients in the TME [214-216]. This reprogramming, via regulating metabolic enzymes’ activities, helps enrich the TME with nutrients and plays a causal role in tumor progression [217, 218].

The metabolic niche of the TME is regulated by four key factors: 1) intrinsic metabolism of the tumor cells, 2) tumor-non-tumor cell interaction, 3) location and heterogeneity of the tumor, and 4) metabolic homeostasis of the body [219]. Regarding intrinsic metabolism, several studies have demonstrated that tumor cells drive aerobic glycolysis and promote cell proliferation via fueling mitochondrial metabolism [220, 221]. Moreover, apart from glucose and lactate, tumor cells use fatty acids, proteins and amino acids as fuel [222, 223]. For example, glutamate is converted into aspartate in cells with a dysfunctional electron transport chain to promote proliferation, allowing tumor cells to adapt rapidly to the substrates available in their TME niche [224, 225].

T cells provide a natural defense against cancer cells as they can specifically kill tumor cells by recognizing tumor-specific antigens. However, glycolytic tumors have very low T cell infiltration and proliferation as activated T cells and tumor cells compete for glucose in the TME [226]. Limiting glucose levels leads to cellular competition; this impairs T-cell function via decreased mTOR signaling. Reduction in mTOR activity diminishes IFNγ transcripts in T cells, mitigating Th1 CD4⁺ T cell differentiation [227-229]. Furthermore, tumor cells compete with T cells for amino acids [230]. For example, glutamine is required for T-cell function and differentiation, and tumor cells use glutamine to activate STAT3 to promote cell proliferation [231].

The polarization of TAMs towards either the M1 (antitumoral) or M2 (protumoral) phenotype is dictated by cellular metabolism and thus regulates its response to tumor cells. Increased glycolysis by tumor cells leads to the formation of TAMs with low glycolytic potential, which promotes metastasis [232]. Moreover, tumor cells produce lactate which induces the M2 phenotype of TAMs via stabilizing HIF-1α and activating G-protein-coupled receptor 132 [210, 233, 234]. Glutamine metabolism in TAMs also promotes an M2 phenotype via the production of α-ketoglutarate, which aids in fatty acid oxidation and epigenetic activation of M2 genes [235].

Tumor location plays a key role in TME modulation because different organs and tissues have different proteomic and metabolic signatures. These differences determine the metabolite dependencies of tumor cells. Moreover, the perfusion level, tissue function and cell-type composition within the same organ also contribute to metabolic heterogeneity. For example, blood vessel proximity distinctly defines metabolic niches. Additionally, a significant correlation has been found between glycolysis, mitochondrial metabolism and local oxygen concentrations in a study conducted on human melanoma and head and neck cancers [236]. Glucose tracing studies stipulate that tumor cells in highly perfused locations rely mainly on glucose, whereas those in less perfused regions depend on other carbon sources [237]. Furthermore, solid tumors, being metabolically heterogeneous, show glutamine-depleted cores, which promote histone hypermethylation, resulting in the reduced expression of differentiation-related genes and cancer cell dedifferentiation [238, 239].

Although cellular metabolism and its role in the TME are well explored, the role of systemic nutrient levels in characterizing the metabolic environment of the TME is still elusive. Recent studies indicate that dietary restrictions and hormonal modulation affect local metabolism [240]. For example, dietary restriction of serine-glycine is beneficial in tumors lacking p53 because of their inability to counteract reactive oxygen species (ROS)-associated oxidative stress [241, 242]. Moreover, the gut microbiome also produces specific microbial metabolites that can be altered by dietary restrictions, thus affecting tumor cell metabolism [243].

Metabolic enzymes were previously considered to catalyze their specific reactions, and their roles were strictly limited to regulating metabolic pathways. However, research in the past decades indicated that these enzymes have a moonlighting function in phosphorylating various proteins that regulate many pathways ranging from cell-cycle progression and proliferation to apoptosis, autophagy and T-cell activation. Some of these enzymes are pyruvate kinase M2, phosphoenolpyruvate carboxykinase 1, acylglycerol kinase, hexokinase and phosphoglycerate kinase 1 [244-247]. Moreover, the metabolic products of these enzymes play a crucial role in regulating gene expression [245, 248]. Interestingly, these enzymes perform their non-canonical functions via protein-protein interactions and regulate many central signaling pathways and functions of several organelles, such as the nucleus, ER and mitochondria [249]. Thus, unraveling the moonlighting functions of metabolic enzymes that help in tumor progression helps us to better understand the TME dynamics and can be exploited to develop better therapeutic interventions [250].

Several challenges hinder the targeting of the pro-tumorigenic metabolic profile of the TME. Targeting the tumor cells’ proliferative profile also affects normal cell metabolism at the systemic level. Combining these drugs with other cancer hallmarks, such as immunity, could increase the therapeutic window for targeting oncometabolites. Another prevalent strategy is to inhibit enzymes mutated in cancer, such as the isocitrate dehydrogenase 1 (IDH) 1/2 mutation-induced oncometabolite D-2-Hydroxyglutarate (D2HG) in glioblastoma (GBM) and acute myeloid leukemia (AML). IDH inhibitors have shown clinical success in AML and are under investigation with combination therapies in GBM. Thus, targeting tumor-specific metabolites in the optimum therapeutic window and exploiting cancer-specific vulnerabilities hold great potential in metabolomics [251].

3.7 cGAS-STING signaling in the TME

Research on the cGAS-STING signaling promoting tumor progression is emerging in the field of cancer. The Cancer Genome Atlas (TCGA) database categorizes 18 different malignant tumor types. Researchers have observed differences in the expression of essential genes in the cGAS-STING signaling mechanism between normal and malignant tissues, along with MB21D1-encoding cGAS, transmembrane protein 173 (TMEM-173)-encoding STING, TANK-binding kinase 1 (TBK-1) and interferon regulatory factor 3 (IRF-3). Comprehensive research and recent evidence proved that these four genes were significantly elevated in almost all cancer models, suggesting that cGAS-STING signal transduction might be stimulated in all cancer types [252, 253]. In some cancer models, extremely invasive tumors can ambiguously depend on and utilize the cGAS-STING pathway to modulate tumorigenesis with significant implications for cancer therapy [254]. NF-κB regulates cell proliferation, apoptosis and survival of normal cells, thus acting as a crucial stimulator of the inflammatory response. Additionally, NF-κB promotes the development of inflammation, tumors and immune dysfunction [255]. Chromosomal instability induces chronic inflammatory signals by constantly activating the cGAS-STING signaling mechanism, which downstream NF-κB function and consecutively increases metastatic cancer cell progression [256] (Figure 1).

Moreover, TCGA dataset analysis revealed that the STING expression level in cancer is negatively correlated with infiltrating immune cells in various tumor models, demonstrating that significant upregulation of the cGAS-STING signaling mechanism predicts a poor prognosis in cancer patients [252]. It has been shown that various tumor cells can specifically advance the accumulation of astrocyte-gap junctions to enhance brain metastasis by expressing protocadherin 7 (PCDH7), composed of connexin 43 (Cx43)[257]. These junction carriers pass to the cGAMP, from cancer cells to adjacent astrocytes, to stimulate STING by triggering IRF-3 and TBK-1 to generate TNF and IFN-α. Similarly, as paracrine signals, these factors further activate the NF-κB and STAT-1 pathway in metastatic brain cells, thus promoting brain metastasis and resistance in lung and breast cancer therapy [257].

3.8 Siglec signaling in the TME

Tumor cells express an abnormal quantity of sialic acids on their cell surface [258]. Most sialic acids belong to negatively charged disaccharides masking the chains of glycan on glycoproteins and glycolipids [258, 259] and are known as sialoglycans. On tumor cells, sialoglycans are involved in tumor cell-to-cell interactions within the TME, and it was also suggested that sialoglycans with negatively charged moieties could form an invisible cover by protecting tumor cells from immune recognition [260]. A vaccination trial with autologous sialidase treatment in human breast and melanoma cancer revealed that the elimination of sialic acid elicited a robust antitumor and an immune response [261]. Studies over the past decades have suggested a clear function of sialoglycans in tumors via immune evasion, not only by shielding of antigens but also by sialic acids show robust immunomodulatory properties as sugar moieties [260, 262, 263]. Sialoglycans form different ligands for sialic acid binding proteins, i.e., Factor H (FH), to evade complemented activation and are proposed as a mediator of the selectin-independent adhesion of lymphocyte trafficking, plays an essential role in metastasis [258, 264, 265]. Sialic acid-binding immunoglobulin-like lectins (Siglecs) have specific interactions with the immunomodulatory properties of sialoglycans. In mammals, siglecs are divided into two groups: the structurally preserved siglecs- (1, 2, 4 to 15) and the CD33-related siglecs (3, 5 to 11, 14 to 16) [266, 267]. Tumor cells with abnormal expression of sialic acid upregulate siglec family expression to infiltrate immune cells in the TME. The function of inhibitory siglecs closely resembles the PD-1 immune checkpoint function (Figure 1) [266, 267]. TME also promotes abnormal sialylation in tumor cells and modulates the expression of siglec on infiltrating immune cells [268]. Future studies could explore the approach to target the dysregulation of the sialoglycan-siglec axis in cancer, which could contribute to shaping the immunosuppressive TME, composing an obstacle to overcome towards effective immunotherapy in cancers.

4.1 Checkpoint signaling mechanisms

Several inhibitory immune receptors have been identified and studied intensively in the TME immune population, such as PD-1, CTLA4, TIM-3, LAG3 and B and T lymphocyte attenuator (BTLA), and are termed “immune checkpoints”, which essentially indicate a molecule that acts as a guard of immune responses against tumors or pathogens. The immuno-suppressive functions of the immune checkpoints usually rely upon ligand-receptor interaction. Recent studies using advanced technologies, like mass cytometry (CyTOF) and single-cell RNA-sequencing (scRNA-seq), followed by functional studies, have shown a dynamic and diverse immune landscape that facilitates the understanding of tumor heterogeneity in various cancers, including various tumor stages and genetic backgrounds. In the TME, exhausted T cells exhibit reduced effector function and increased expression of immune checkpoints (e.g., PD-1, CTLA4, TIM-3, LAG3, BTLA) [269]. In this section, we summarize several well-researched immune checkpoint receptor signaling mechanisms.

4.2 Immunotherapies targeting immune checkpoint signaling

In recent years, the TME has appeared as a promising approach for targeting several cancers. Here we discuss recent research targeting immune checkpoint signaling by blockade antibodies, ICIs, or various tumor-suppressive blocking peptides, nanoparticles, and CAR-T cell therapy to target the receptor-ligand intercommunication and attenuate T cell function to kill tumor cell progression. Many studies or clinical trials have conquered and are authorized for clinical translation [335]. Nonetheless, the overall survival response for these immunotherapies is not enough and needs a more comprehensive study [335]. Blocking antibodies PD-1 or PD-L1 is the most widely used in various cancer models as immunotherapy. T cell-targeted immune modulators are used as monotherapy or combination treatment with chemotherapies for almost all cancer types [336, 337]. Active clinical trials on different immune checkpoint proteins in several cancer models are shown in Figure 2 and Table 2. Recently small molecules have been studied and shown to have the potential to emerge into immune checkpoint inhibitors with nanomolar binding affinity to PD-L1. Conceivable data revealed abolished PD-1 and PD-L1 interactions, improved T-cell activity, and enhanced antitumor immunity in colon, breast, pancreatic, and kidney cancer models [338-341]. Recently, glioblastoma mouse models revealed the inhibition of intratumoral immune-suppressive microglia or macrophages into the TME and increased CD8⁺ T cell infiltration, activation, and cytotoxicity into the tumor tissue and synergizing with anti-PD-1 combination therapy [342, 343]. In contrast, to intensify the T cells’ function and minimize the adverse effects of using CTLA4 immunotherapies, ICI treatment using combinations of anti-PD-1 (nivolumab) and anti-CTLA (ipilimumab) have shown to boost CD8⁺ T cell activation, proliferation, enhance the T_eff cell memory (CD8⁺) and produce interferon-γ and granzyme-B in malignant pleural mesothelioma (MPM) patients [344]. Recently, the US FDA-approved PD-L1 inhibitors atezolizumab and durvalumab were combined with carboplatin plus etoposide (CP/ET) to target and kill aggressively growing tumors in advanced SCLC patients and reported an overall increase in survival [345, 346]. Recent data from several clinical trials are encouraging combination therapy with a TIM-3 directed antibody (TSR-022) and PD-1 together with chemotherapy for patients with NSCLC [347]. The researcher explained that the combination regimen was well tolerated across multiple dosing levels and that the responses showed the strategy is worth pursuing. Anti-LAG3 monoclonal antibody relatlimab was also evaluated in several advanced solid tumors such as head and neck squamous cell carcinomas (HNSCCs), melanoma, NSCLC, renal cell carcinoma and bladder cancer, and the trial evaluated the efficacy of relatlimab, mono-immunotherapy or in a combination regimen with an anti-PD-1 antibody nivolumab [348]. LAG3 may synergize with other inhibitory molecules (PD-1, CTLA4) to improve the inhibitory activity of Treg cells, leading to APC-induced immune tolerance [327]. Several ongoing clinical trials investigating the blockade of TIM-3 inhibitor (sabatolimab) with or without PD-1 inhibitors in advanced solid tumors concluded that the doses for combination therapy of sabatolimab and spartalizumab were under tolerance and showed anti-cancer activity with survival benefits [349]. Moreover, CAR-T cell therapy is emerging to be a progressive new pillar in immune cell therapy for cancer, which has yielded remarkable clinical responses in patients with B-cell leukemia. However, many challenges remain to be addressed to overcome its ineffectiveness in treating solid tumors and hematological malignancies [350]. The great potential of CAR-T cell therapy at the beginning or earlier during the treatment course was unraveled, and the strategy revealed higher success rates and reduced toxicity associated with anticancer treatments [351]. Early administration of the therapy may also provide access to a higher proportion of naïve unexposed T-cell population, which is beneficial to facilitate the production of CAR-T cells. Finally, immunotherapeutic development of targeting immune checkpoint signaling molecules has promptly increased over the decade. The development of new biomarkers or identification of new targeting mechanisms with combinational strategies and novel nano-drug delivery strategies could significantly boost immunotherapy benefits in an effort to improve the quality of life for cancer patients by enhancing overall survival and eventually eliminate cancer.

4.3 Immunosuppressive chemokine cell signaling into the TME

The chemokine gradient determines the composition of the TME. Pro-tumor chemokine gradient attracts immune suppressive immune cells and excludes effector immune cells. On the other hand, the antitumor chemokine gradient favors the migration of effector immune cells. Here, we discuss key chemokine signaling axes in tumors and immune crosstalk therapeutically targeted to make the TME less immune suppressive.

4.4 Therapeutics targeting chemokine signaling in the TME

5.1 Spheroids

Spheroids are the most widely used 3D tumor models that precisely model in vitro conditions of a TME, including cellular heterogeneity, cell-cell interactions, cell-ECM interactions, signaling pathways, and gene expression profiles [376, 377]. Spheroids are created by a forced aggregation of cells, and those larger than 500 μm mimic the TME of micrometastases (microscopic, aggregated cancer cells that have escaped the original tumor) and avascular tumors owing to the presence of a gradient of oxygen and nutrients [378]. Moreover, this gradient leads to the formation of concentric zones in the spheroids with a center containing an anoxic core with necrotic cells, a middle zone containing hypoxic cells, and an outer zone of highly proliferative cells, recapitulating the properties of an in vitro tumor [379].

Spheroids can be associated with microfluidic or scaffold-based procedures to generate physiologically more relevant tumor models [377]. They are also advantageous over 2D cell culture models in studying the role of ECM stiffness on tumor growth when combined with biomaterials like hydrogel by mimicking cell-cell/cell-ECM interactions in the TME [380]. Spheroids can also be fine-tuned by varying the composition of tumor cells, CAFs, and immune cells to obtain an immunosuppressive TME to study the roles of various immune cells like monocytes, NK cells, or T cells in tumors [381]. The spheroid model of human breast cancer played a key role in discovering the dependence of angiogenesis (sprout formation) on VEGF and FGF growth factors [382]. Spheroid 3D models, therefore, represent a powerful model system for studying tumor biology and drug testing. Nonetheless, there are certain limitations, including 1) poor uniformity in spheroid size or morphology, 2) low throughput, and 3) difficulties in retrieving cells for analyses, etc., in the 3D spheroid models that limit reproducibility while using these models [377].

5.2 Organoids

Organoids are small 3D self-organized tissue cultures derived from progenitor stem cells that maintain natural cancer cell heterogeneity in vitro, including genetic and phenotypic features of the tissue from which the culture is derived [383, 384]. In contrast to the spheroids, where a forced aggregation of cancer cells forms a 3D structure, organoids follow a natural course of 3D tumor development based on the genetic programming of progenitor cells [377]. Therefore, organoids more closely resemble the in vitro development of tumors and can be cultured from patient-derived cells from biopsy samples. Since patient-derived organoids (PDOs) consist of only epithelial cells, to develop a TME in organoids, either of the two approaches for organoid culture can be used: reconstituted the TME (submerged Matrigel culture) model or native TME (air-liquid interface culture) model.

5.3 Microfluidic models

Microfluidic models consist of a network of microfluidic channels where tumor spheroids can be continuously perfused and grown in the presence of fine-tuned concentrations of growth factors or drugs [397, 398]. These models allow fine control over the mechanical forces, the orientation of the tissue interfaces, and chemical gradients, and they can be modified for high throughput screening [399, 400]. Moreover, microfluidic devices used in these models use microscale volumes, which can drastically reduce their operating costs compared to 3D TME models [401]. They have been extensively used for studying tumor-stroma interactions and the effects of growth factors or drugs in a biomimetic TME for various cancer types, including breast cancer, colorectal cancer, melanoma, and Merkel cell carcinoma [376, 402]. Specific cell-cell interaction can also be studied using microfluidic models, for example, in tumor cell and stromal interaction studies. Culturing organoids in the microfluidic device have helped us understand tumor-stroma interactions and their systemic effects [403, 404]. They have also been used to mimic microvascular functions and TME modeling [175]. Limitations of the microfluidic models include the requirement of specialized skills to perform the experiments, edge effects and high shear pressure in the device can compromise the reproducibility of the experimental outcomes, and the most used material, polydimethylsiloxane (PDMS), to fabricate microfluidic devices indiscriminately absorbs small molecules [397, 398, 405, 406].