Cytokine CCL5 and receptor CCR5 axis in glioblastoma multiforme

Brain tumours originate from various types of cells, of which gliomas are most frequent. Recent epidemiological data in UK confirmed that glioblastoma (GB) is also the most common among glial tumours with 5–7 cases per 100.000 individuals yearly, and represents 50% of all gliomas.1 The World Health Organisation (WHO) distinguishes four grades of glioma, of which GB is the most aggressive, invasive, and lethal among all types of brain tumours. According to the standard histological classification, GB originates from neoplastic glial cells, also called astrocytes, either via the de novo pathway without clinical or histologic evidence of a less malignant precursor lesions (primary glioblastoma) or via the progressive pathway through development from a low-grade astrocytoma, progressing to anaplastic astrocytoma into diffuse glioblastoma (secondary glioblastoma). The major marker of secondary glioblastoma is mutated isocitrate dehydrogenase (IDH1)2, although also expressed in the proneural subtype of primary glioblastoma. Regardless of the origin, GB is characterized by histological features, such as necrosis, vascular proliferation and pleomorphism.3 Contrary to most tumour types, irradiation and chemotherapy have proven to be ineffective to impair GB progression in longer term, demonstrating its remarkable therapeutic resistance.4 Commonly used chemotherapeutic is temozolomide (TMZ), showing the highest effectiveness in GB.3, 5 However, only in about 55–60% of patients with methylated, i.e. silenced gene coding for O6-methylguanine deoxyribonucleic acid (DNA) methyltransferase (MGMT), thereby lowering the enzyme expression, the responsiveness to TMZ is more effective.6 There is thus continuous search for new, adjuvant therapeutics, including kinase inhibitors, anti-angiogenic agents and recently also immunotherapeutics, to increase average survival of glioblastoma patients.

There are other reasons for GB therapy resistance, i.e. the presence of glioblastoma stem cells (GSCs), mostly due to their high DNA repair mechanisms expression.7 The heterogeneity and plasticity of these cells that carry the genetic fingerprint of the developing glioma, has been recognised as one of the additional obstacles for their resistance. GSCs, similar to normal neural cells, which are precursors of glial and neural cells, express the characteristic stemness genes (e.g. CD133, Sox2, Nanog, Olig 4, Notch, etc.), in addition to selected oncogenes and tumour suppressor genes.8 The initial transcriptome analyses by Philips et al9 and Verhaak et al.10, set the basis for the Cancer Genome Atlas (TCGA), defining four different glioblastoma subtypes: the proneural (PN), mesenchymal (MES), neural (N), and classical (CL) by their major genomic characteristics, which are: PDGFRA/IDH1 (in PN), NF1/TP53 (in MES) and epidermal growth factor receptor EGFR abnormalities i.e. amplification of the epidermal growth factor (EGF) (in N) and mutation in EGFRvIII/PTEN (in CL). These GB subtypes differ significantly in survival rate, being shorter in MES subtype. However, mixed subtypes are observed in the single tumour, giving rise to intra-tumour heterogeneity.11

Solid tumour progression is not only relaying on the genetic and epigenetic variations of cancerous cells acquired during their evolution, but also on how their homotypic and heterotypic interactions with the stromal cells of associated microenvironment are. The “tumour microenvironment” (TME) consists not only of local, resident cells being invaded by the tumour cells, but also of infiltrating host cells, e.g. bone-marrow and blood-derived mesenchymal stem cells (MSC) and haematopoietic stem cells (HSC) and their progenitors, e.g. mature lymphocytes, macrophages, etc. Very recently, Salmon et al.12 reviewed specific determinants that might influence tumours development and argue that unrevealing these selective interactions, mediating for example tumour immunity should facilitate development of immunotherapeutic precision strategies for patients with cancer.

In glioblastoma in addition to their autonomous (inter-tumour) heterogeneity11, the increasing attention is paid to their non-autonomous heterogeneity (intra-tumour heterogeneity), that is presenting an obstacle to a more informed treatment, as we still do not understand the ability of GB cells to manipulate and exploit these non-cancerous cells collectively termed “tumour stroma”. As stated by Broekman et al.13, almost all cell types in the GB stroma are affected: the tumour is able to stimulate angiogenesis and co-opt existing vasculature, suppress immune cell functions, disarm microglia and macrophages that should recognize and fight these “foreign elements” in the brain and coerce astrocytes into supporting tumour modification extracellular matrix (ECM) to facilitate invasion. GB cells recruit innate immune cells and change their phenotype to support tumour growth and suppress adaptive immune responses.14 The increasing understanding of how T cells access the brain and how the tumour tricks the immune response, offers new strategies for mobilizing an anti-immune tumour response. For example, GB cells extensive cross-talk with microglia and infiltrating macrophages, through the release of cytokines (see below), extracellular vesicles exchange and connecting nanotubes and microtubules, results in their support to malignancy, as reviewed by Matias et al.15 Among these cells, MSC homing to GB, have crucial effects on the microenvironment, either by their de-differentiation to other stromal cells, via paracrine effectors such as immunomodulatory cytokines16, or by direct interactions with GB cells.17

Thus, GB cells spreading to the brain involve multiple modes of communication with stromal cells as extensively reviewed by Matias et al.15 The aim of this review is to reveal the complex interactions of one of the most important cytokines, affecting glioblastoma progression i.e., chemokine (C-C motif) ligand 5 (CCL5) and its receptors.

As reviewed by Balkwill18, chemokines are chemotactic cytokines that cause directed migration of stromal cells, such as leukocytes, that are induced by inflammatory cytokines. Chemokine signalling results in transcription of target genes that are involved in cell invasion, motility, interactions with the extracellular matrix (ECM) and cell survival. Chemokine signalling can coordinate cell movement during inflammation, as well as the homeostatic transport of HSCs, MSCs, myeloids cells lymphocytes, dendritic cells and neutrophils, as well as cancer-associated fibroblasts (CAFs).19 Directed migration of cells that express the appropriate chemokine receptor, occurs along a chemical gradient of ligand - known as the chemokinereceptor axis — allowing cells to move towards high local concentrations of chemokines. More than 50 human chemokines and 20 chemokine receptors have been discovered so far. Cytokines, as pro-inflammatory mediators, when excessive, also play a role in causing chronic inflammation, for example induced by bacterial (e.g. by H. pylori) or after viral (Hepatitis B) infections, inducing immune suppression.20 Various cancers have a specific complex chemokine network that influences the immune-cell recruitment of inflammatory cells to the tumour milieu, but being neutralised by cancer immune suppression, and providing an immunological privilege that enables neoplasia, i.e. tumour cell growth, survival and migration, angiogenesis and metastasis.19,21

The cytokine CCL5, also classified as C-C motif chemokine 5, has been initially termed RANTES (“Regulated upon Activation Normal T cell Expressed and Secreted”), and as a potent chemokine, attracting leukocytes.22 Later, CCL5 was recognised as a versatile inflammatory mediator, expressed by breast cancer cells (BC), where along with CCL2 it promoted pro-malignant activities by attracting macrophages, T-cells and granulocytes, as well as mesenchymal stem cells and enhancing angiogenesis.23 CCL5 has been suggested as potential therapeutic target to impair the disease progression. In immune cells, CCL5 was reported initially as a HIV-suppressive factor and expressed mainly by CD8+ T cells.24 It binds to its cognate receptor C-C chemokine receptor type 5 (CCR5), which is (alongside C-X-C chemokine receptor type 4 (CXCR4)), an HIV entry co-receptor into CD4+ T cells.25 The Food and Drug Administration approved the first CCR5-based entry inhibitor, now called maraviroc (MRV) in 2009, and based on this, new drugs that promote CCR5 and CXCR4 internalization, independent of canonical cellular signalling, provided clinical benefits for HIV patients with minor side effects.

The mode of CCL5 action thus comprises binding not only to its cognate and the most studied interacting partner, the CCR5 membrane receptor, but also to other members of the G-protein coupled receptors (GPCR), such as C-C chemokine receptor type 1 (CCR1) and C-C chemokine receptor type 3 (CCR3). In addition to that, non-conventional or auxiliary receptors of CCL5 are CD4426 and GPR75 (Figure 1A).27 This is not unusual, as many chemokine receptors display promiscuous ligand binding, meaning they have more than one high-affinity ligand.28 Such variety of CCL5 interactions causes the activation of multiple pathways and gives the ligand a diverse range of not only physiological, but also pathological functions, including in cancer.29 In this vein, CCL5 has been shown to be highly expressed in a plethora of cancer types, such as colorectal30, lung31, prostate32, breast28,33,34 and cervical cancer.35 In addition, tissue or plasma CCL5 also serve as a marker of poor prognosis, as it is the case in cervical35, prostate32, gastric36, breast37, pancreatic cancer38, as well as in glioblastoma.39 The cells that express and secrete CCL5 in cancer, can either be cancerous cells themselves or stromal cells.40

Figure 1

The receptor CCR5, classified as C-C chemokine receptor 5, alternatively termed also CD195, is the main receptor through which CCL5 transduces signalisation.41,42 Structurally, CCR5 is a GPCR, as are many other chemokine receptors.41 This means that it has a N-terminal extracellular tail responsible for ligand binding, seven hydrophobic transmembrane regions, the six loops that connect them, and a C-terminal cytosolic tail. This is crucial for transducing the signal caused by ligand binding after its heterotrimeric G-proteins binding or through G-protein independent pathways.41

Similar to other chemokine receptors, CCR5 also acts redundantly for signalling pathways.28 High affinity ligands that bind to CCR5 are CCL5, as well as chemokine (C-C motif) ligand 3 (CCL3) (also known as MIP-1α) and chemokine (C-C motif) ligand 4 (CCL4) (also known as MIP-1β) (Figure 1B).42,43 As already mentioned, substantial research has been dedicated to the role of CCR5 in HIV infection; M-tropic or macrophage strain HIV-1 uses CD4 as its main receptor to bind to and enter CD4+ T cells, but for this it also needs co-receptors, CCR5 and CXCR4 acting as such. Small molecular inhibitors (maraviroc [MRV]) and the humanized monoclonal antibody (leronlimab)44,45 are CCR5 antagonists that inhibit HIV-1 virus from entering the T-cells.43 MRV binds to CCR5 and acts as an allosteric inverse agonist, locking CCR5 in an inactive conformation.42 However, recent research has focused more on cancer, as similarly to CCL5, CCR5 is overexpressed in many types of cancers, including breast28,33,34,46, prostate32 and in particular glioblastoma.47,48 Both maraviroc49 and leronlimab46 have been shown to potently block cancer metastasis in murine xenografts.

The canonical (conventional) way of signalling is through a hepta-helical chemokine receptor and adjacent G proteins, more specifically their Gα subunit and Gβγ dimer.41 Upon ligand binding, CCR5 activates Gαβγ trimer by causing guanosine exchange (GDP à GTP) and the dissociation of the membrane-bound Gα subunit from the Gβγ dimer.50 Activated Gα affects adenylyl cyclase, and subsequently cellular cyclic adenosine monophosphate (cAMP) levels that activates cytosolic protein kinase A (PKA). Gα together with Gβγ affect various targets (e.g. PLCβ), resulting in the production of secondary messengers, such as inositol-1,4,5-triphosphate (IP3), diacylglycerol (DAG) and increased cytosolic calcium concentration. Both Gα and Gβγ trigger calcium influx, therefore the direct interaction of chemokine and its receptor can be confirmed by a calcium mobilisation assay.38 Influx of calcium activates calcium-dependant pathways (NF-κB), as well pathways independent of G-proteins, all favouring malignancy in one way or the other.41

The CCL5/CCR5 axis has been recently reported as a mechanism of tumour progression in pancreatic38, gastric20 and breast cancer.33, 44 Noteworthy, in cancer, the CCL5-receptors signalling can favour cancer progression directly by affecting proliferation, migration and cell survival of cancer cells, or indirectly, by affecting tumour microenvironment, i.e. by recruiting pro-tumour and/or anti-inflammatory effector cells.20, 51 The state of the art in affecting the key hallmarks of glioblastoma progression, first described by Kouno et al.47 and recent reports on MES-GB subtypes29 and glioblastoma stem cells52, 53 will be discussed below.

Cancer cell proliferation is regulated by many pathways, one of them, most commonly mediated by CCL5/CCR5 signalling, is mammalian target of rapamycin mTOR) pathway. This pathway is convergent with the Phosphatidylinositol 3-kinase (PI3K) pathway, as they both activate the Akt (protein kinase B) pathway (Figure 2).54

Figure 2

The “mTOR is a kinase, encoded by mTOR gene MTOR and is the member of phosphatidylinositol 3 kinase (PI3K)-related kinase family that plays an important role in transcriptional activation, as it regulates the eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1). The role of 4E-BP1 is to sequester the eukaryotic translation initiation factor 4E (eIF4E), inhibiting translation. By inducing hyper phosphorylation of 4E-BP1, mTOR complex 1 (mTORC1) disables its eIF4E binding, enhancing the rate of translation.51 Binding to CCR5, CCL5 has been proven to activate the mTOR/4E-BP1 pathway, inducing the translation of a specific subset of mRNAs that have a long and highly structured 5’-UTR region, coding for cell survival and growth related onco-proteins, such as cyclin D1, c-Myc, and Dad-1.20 Indeed, CCL5/CCR5 signalling stimulated survival and proliferation of MCF-7 breast cancer cells through the mTOR/4E-BP1 pathway.55 Although similar has not directly been shown in glioblastoma, there are numerous reports on the role of the mTOR pathway in GB.56 Moreover, Khan et al.52 have shown that inhibition of mTOR complexes, mTORC1 and mTORC2, significantly increased the in vitro and in vivo sensitivity of glioblastoma stem cells to radiotherapy. The relevance of mTOR in GSCs, has been recently demonstrated by Mecca et al.54, by inhibition of mTORC1/2 in glioblastoma, causing persistent and dramatic reduction in p-Akt levels, which inhibited GSCs’ proliferation. Along these lines, Pan et al.29, demonstrated that CCL5/CCL5-receptor signalling in GB cells created an autocrine signalling circuit, important for high-grade glioma growth. Interestingly, they found that increased CCL5 expression was restricted to both human and mouse MES-GB, subtype characterized by NF1 protein (neurofibromin). This protein negatively regulated CCL5 expression through suppression of AKT/mTOR signalling. Zhao et al.56 also reported that GB cell proliferation is mediated through CCR5 signalling. Using BrdU incorporation in vitro in the GB cells U87 and U251, they determined that CCL5 stimulation significantly enhanced proliferation. In their experiments, CCL5/CCR5 axis activation triggered the PI3K/Akt pathway to promote proliferation, whereas PI3K inhibitors decreased Akt phosphorylation, which in turn decreased proliferation. However, both mTOR and PI3K are known to activate the Akt pathway, but the mutual relation of these two pathways is not clear. Further studies in GB are urgently needed due to the notions that most aggressive MES-GB and GSCs are affected by CCL5/CCR5 mediated treatment.