Introduction

Lung cancer leads the ranking of cancer-related mortality in both the USA and Europe ( ). Non-small cell lung cancer (NSCLC) shows predominance over small cell histology (SCLC), which accounts for only 15–20% of the incident cases ( ). The biologic behavior of the two histological subgroups is inherently different, with SCLC being characterized by intense proliferative activity, higher propensity to early dissemination and intrinsic chemosensitivity ( ). Approximately 70% of newly diagnosed cases of SCLC present with extensive disease and, despite the generally favorable response rates to frontline chemotherapy, the life expectancy for these patients does not normally exceed 12 months ( ).

NSCLC represents a heterogeneous aggregate of histologies including squamous cell, adenocarcinoma and large cell carcinoma. Recent advances in the genetic characterization of NSCLC have enriched such classification introducing molecularly defined tumor phenotypes. The presence of epithelial growth factor receptor (EGFR) activating mutations, for example, defines a subgroup of NSCLC affecting approximately 50% of Asian patients and 10% of non-Asians, that are more likely to respond to EGFR tyrosine kinase inhibitors. More recent work on novel targets such as anaplastic lymphoma kinase (ALK) rearrangements, which identify 2–7% of adenocarcinomas with favorable response rates to ALK inhibitors, has further strengthened the concept of molecular classification of NSCLC ( ).

In the context of such a molecularly composite disease, Tumor Node Metastasis (TNM) staging remains a solid guide to treatment allocation and holds prognostic significance. In fact, while patients presenting with either surgically unresectable (stage IIIB) or metastatic tumors (stage IV) will inevitably succumb from their disease, those with operable NSCLC may potentially hope for a cure, with 5-year survival probabilities ranging from 80% for stage IA, to 30% for stage IIIA ( ).

Regardless of histology, metastatic diffusion is viewed as the main factor that influences treatment failure in the curative setting ( ), reduces the chances of long-term survival as a result of organ failure and determines the symptomatic burden typically affecting patients with advanced cancer. Metastatic dissemination to the liver and the central nervous system (CNS) is fairly common in the evolution of bronchogenic carcinoma. However, the development of metastatic spread to these districts remains difficult to predict in the individual patient.

Intracranial diffusion is the most common pattern of treatment failure in patients who present with potentially curable NSCLC ( ), and eventually affects up to 55% of the patients with locally advanced disease ( ). Cerebral metastases have an invariably fatal clinical course and represent the direct cause of death in these patients ( ). Although a number of clinical predictors of metastatic spread to the CNS including nodal status and non-squamous histology have been identified, the precise molecular mechanisms contributing to the neurotropic potential of lung cancer cells have not been fully elucidated.

Organotropic Spread of Lung Carcinoma: Role of Chemokine Signaling Pathways

Metastatic spread is viewed as a multistep process involving the adaptive selection of a genetically unstable cancer cell clone towards the acquirement of an invasive phenotype capable of overcoming the intrinsic biological and physical barriers to cell diffusion. The presence of a positive selective pressure to overcome the presence of extracellular matrix components, elude the surveillance of the immune system, secure survival in the bloodstream is fundamental in driving the metastatic phenotype. Similarly, the ability to arrest and extravasate into the target organ and display independent growth capacity are equally important traits that the metastasizing cell is specifically selected to display ( ).

The concept of cancer cell metastatization as a programmed, multifactorial cascade of events is not novel in oncology, having been first hypothesized by Stephen Paget in 1889, who compared the predisposition of certain organs to become colonized by secondary growths to a “fertile soil” that can actively promote tumor dissemination and growth.

In the context of lung cancer, mechanistic research has started elucidating the laborious genetic reprogramming underlying metastatic spread, that is now known to involve the activation of discrete molecular pathways such as epithelial-to-mesenchymal transition (EMT), a complex signaling cascade by which epithelial cells progressively lose cell–cell adhesion, apical-basal polarity and increase their motility and invasive properties ( ).

Unequivocal evidence has shown that chemokine signaling pathways, originally discovered as mediators of directional migration of immune cells following tissue injury and inflammation, may participate in systemic diffusion of cancer by modulating the site-specific homing of metastasizing tumor cells ( ).

Chemokines constitute a superfamily of small cytokines with molecular weight ranging from 6 to 14 kD. Historically subclassified as C, CC, CXC or CX3C according to the structure of repeated cystein residue-rich motifs within the N-terminal region of the protein, chemokines were first characterized as mediators of the innate immunity, being secreted by a wide variety of cells paricipating in the acute phase reaction including endothelial cells, monocytes/macrophages and leukocytes ( ). Besides inducing chemotaxis in responsive cells following insult, chemokine subgroups are secreted either phasically or constitutively to participate in tissue homeostatic processes such as organ development, angiogenesis, and wound repair ( ).

More than 50 different chemokines have been identified, all of which mediate their biological functions by binding to a G-protein coupled trans-membrane receptor. Each receptor is characterized by seven hydrophobic trans-membrane domains, an N-terminus extracellular chemokine binding region and an intracellular, serine and threonine rich domain which mediates intracellular signaling by ligand-controlled phosphorylation ( ). Among the downstream molecular events triggered by chemokine receptor activation, conformational changes in the cytoskeleton and in cell adhesion molecules, such as integrins, facilitate cell motility and chemotaxis ( ).

Chemokine signaling pathways can act at various levels in the context of metastatic spread. Recent evidence suggests that the chemokine secretion signature of a given tissue combined with the selective expression of the matched receptors by tumor cell clones might explain the site-specific spread of metastatic foci observed in solid tumors ( ). Pioneer work in breast cancer demonstrated an elevated expression of chemokine receptors CXCR4 and CCR7 in human breast cancer cell lines and in both primary and secondary specimens and that – more interestingly – the expression of their respective ligands CXCL12 and CCL21 peaks in those organs that are more frequently colonized by secondary deposits ( ).

In lung cancer, elevated mRNA levels of the chemokine receptor CCR7 measured in primary NSCLC specimens were found to correlate with more advanced stage and predict for a significant lymphotropic behavior ( ). Similarly, the expression of chemokine receptor CXCR4, a highly prevalent molecular trait in NSCLC specimens and immortalized cell lines, guides chemotaxis and metastatic diffusion to those tissues where levels of its ligand CXCL12 are more abundant ( ).

Taking advantage of the notoriously pleiotropic and redundant nature of chemokine signaling, additional biological actions that are inherent to the establishment of metastatic cancer growth are at least in part sustained by an intricate network of interactions between chemokines and their receptors, including the acquisition of invasive properties as well as the promotion of neoangiogenesis ( ).

CX3CR1 As a Molecular Determinant of Organotropic Diffusion of Lung Cancer

An increasing scientific interest has been devoted to the study of the chemokine receptor CX3CR1, a transmembrane protein that uniquely interacts with CX3CL1 or fractalkine, a large, 373 amino acid-long chemokine that exists as a membrane bound protein or in a soluble form ( ). The soluble pool of fractalkine is maintained by cleavage of the membrane-bound reservoir, which is operated constitutively by the protease ADAM10 and can be enhanced following tissue injury by ADAM17 ( ). Fractalkine expression is mainly distributed on endothelial cells, lymphocytes, neurons, microglial cells, and osteoblasts. Besides its established role in chemotaxis, fractalkine promotes neuroprotective effects during brain injury, regulates endothelial cells and tissue homeostasis by exerting anti-apoptotic effects ( ).

In solid malignancies, the expression of fractalkine and its receptor CX3CR1 modulates the biologic behavior of the disease in a histotype-dependent manner. With the exception of colorectal and hepatocellular carcinoma, where elevated expression of fractalkine ( ) and CX3CR1 ( ) exerts a protective effect on patients’ prognosis, possibly through a direct enhancement of antitumor cytotoxic immune response, the effects of fractalkine/CX3CR1 signaling seem to mediate predominantly pro-tumoral effects in most other tumor types. The production of fractalkine by mature osteoblasts and bone marrow endothelial cells has been identified as a major determinant of skeletal diffusion of CX3CR1-bearing prostate adenocarcinoma cells ( ).

A similar molecular interplay guides the neurotropic potential of human pancreatic ductal adenocarcinoma, where malignant cells but not their normal exocrine epithelial counterparts largely express CX3CR1. The interaction between CX3CR1-positive cell clones and fractalkine enhances adhesion and migration through nerve sheaths via conformational changes in β-integrins and activation of focal adhesion kinase. In clinical samples, CX3CR1 positivity was associated with worse disease-free survival and increased rates of perineural invasion, a recognized adverse prognostic factor in early-stage pancreatic carcinoma ( ).

In epithelial ovarian carcinoma, the CX3CR1/fractalkine axis guides transcelomic diffusion by favoring migration of cancer cells in ascitic fluid and modulating their adhesion to mesothelial surfaces ( ), whereas in breast cancer, CX3CR1 overexpression in the primary tumor confers a 10-fold risk increase for the development of brain metastases in a group of 142 patients with node-positive breast carcinoma undergoing radical treatment ( ).

Based on the increasing body of evidence supporting the role of CX3CR1 and its ligand as molecular determinants of metastatic diffusion in a wide range of solid tumors, our interest in CX3CR1 was focused on understanding its role in the malignant progression of lung cancer and investigating whether CX3CR1 may represent a molecular factor involved in the metastatic spread of both SCLC and NSCLC.

To test this hypothesis we assembled a large isogeneic collection of primary lung cancer specimens and matching metastatic deposits by reviewing post-mortem examination records of 499 patients with a definitive histological diagnosis of lung cancer and absence of active anticancer treatment prior to the patient’s death. Original paraffin embedded samples were reviewed and only 98 cases with satisfactory preservation of the tissue were sampled to construct a tissue microarray block. CX3CR1 expression was assayed using immunohistochemistry.

Our results show that the CX3CR1 expression in the primary lung tumors was significantly higher in NSCLC (63%) compared to SCLC (7%). Histological subclassification revealed large cell neuroendocrine carcinoma (100%) and squamous cell carcinomas (93%) as being most predominant in CX3CR1 expression when compared to adenocarcinomas (50%).

The disproportion of CX3CR1 immunopositivity seen in the primary tumors was preserved in metastatic deposits, where NSCLC yielded a significantly higher proportion of CX3CR1 immunopositivity (65%) compared to SCLC (29%). Interestingly, the distribution of CX3CR1 expression across the sampled metastatic sites of NSCLC showed that atypical sites of metastatic spread including thyroid, heart, small bowel and skin were entirely positive to CX3CR1, followed by more than 75% of the kidney, liver, bone and brain deposits. All the sampled splenic metastases from NSCLC primaries were negative.

In SCLC cases, 83% of the CNS metastatic sites were positive to CX3CR1, whereas only a minority of the other sites, including locoregional lymph nodes (14%) and liver (28%), expressed the receptor. Adrenal gland metastases arising from SCLC were all negative.

While metastatic disease was predominantly CX3CR1 positive, the subset of primary tumors lacking CX3CR1 expression displayed a significantly higher proportion of liver and brain metastatic deposits, with a predominance of lung adenocarcinoma, a histotype with peculiar neurotropic and hepatotropic potential.

Because fractalkine is expressed in the normal CNS including activated brain endothelial cells, we hypothesize that its constitutive expression may provide a chemoattracting gradient promoting trans-endothelial migration of metastasizing tumor cells ( ). Similarly, the abundant expression of fractalkine by hepatic sinusoidal cells may explain the significant association between CX3CR1 status in the primary tumor and the concurrent presence of liver metastatic diffusion ( ).

The high expression rates observed in liver and the central nervous system (CNS) metastatic sites support the hypothesis that the acquisition of CX3CR1 positivity is not a random event, but may represent a site-selective molecular event conferring growth advantage to the metastasized cell clone.

In view of the recognized role of fractalkine in the promotion of angiogenesis, it is also reasonable to presume that the increased proportion of CX3CR1 positivity observed within the brain deposits may reflect a locally enhanced pro-angiogenic environment. As previously shown, brain metastases do not share the same angiogenic phenotype of the corresponding primary NSCLC. In particular, the high production of vascular endothelial growth factor (VEGF) observed in the brain secondary tumors may be mirrored by the consistent and potentially angiogenesis-driven upregulation of CX3CR1 ( ). Expression of CX3CR1 receptor in metastatic sites is also relevant to ensure sustained survival of disseminated cancer cells through the activation of discrete intracellular pathways such as AKT and ERK1/2 ( ).

Evidence emerging from the study of lung cancer specimens suggests that CX3CR1 is involved in the clinical progression of NSCLC, being expressed in the majority of NSCLC metastatic deposits as well as in a predominant proportion of SCLC brain metastases. Expression of CX3CR1 is differentially distributed across the diverse histological subgroups of primary lung cancer, with almost three-quarters of brain metastatic counterparts spread from both SCLC and NSCLC primaries expressing CX3CR1.

The vast proportion of CX3CR1 positivity observed in metastatic disease provides preliminary, although convincing evidence supporting a potential role for specific inhibitors of the fractalkine/CX3CR1 axis in the setting of advanced lung carcinoma and possibly, as an adjuvant strategy to treat micrometastatic disease. Moreover, pending confirmation in independent, recruited patient cohorts, routine assessment of CX3CR1 status by immunohistochemistry could prove useful in stratifying patients who are more at risk of CNS or hepatic relapse following radical treatment.