Cytokine patterns in patients with cancer: a systematic review

Summary

Active, but dysfunctional, immune responses in patients with cancer have been studied in several tumour types, but owing to the heterogeneity of cancer theories of common reaction mechanisms seem to be obsolete. In this Review of published clinical studies of patients with cancer, expression and interplay of the following cytokines are examined: interleukin 2, interleukin 6, interleukin 8, interleukin 10, interleukin 12, interleukin 18, tumour necrosis factor α (TNFα), transforming growth factor β (TGFβ), interferon-γ, HLA-DR, macrophage migration inhibitory factor (MIF), and C-X-C motif chemokine receptor 4 (CXCR4). Clinical data were analysed in a non-quantitative descriptive manner and interpreted with regard to experimentally established physiological cytokine interactions. The clinical cytokine pattern that emerged suggests that simultaneous immunostimulation and immunosuppression occur in patients with cancer, with increased concentrations of the cytokines MIF, TNFα, interleukin 6, interleukin 8, interleukin 10, interleukin 18, and TGFβ. This specific cytokine pattern seems to have a prognostic effect, since high interleukin 6 or interleukin 10 serum concentrations are associated with negative prognoses in independent cancer types. Although immunostimulatory cytokines are involved in local cancer-associated inflammation, cancer cells seem to be protected from immunological eradication by cytokine-mediated local immunosuppression and a resulting defect of the interleukin 12–interferon-γ–HLA-DR axis. Cytokines produced by tumours might have a pivotal role in this defect. A working hypothesis is that the cancer-specific and histology-independent uniform cytokine cascade is one of the manifestations of the underlying paraneoplastic systemic disease, and this hypothesis links the stage of cancer with both the functional status of the immune system and the patient's prognosis. Neutralisation of this cytokine pattern could offer novel and so far unexploited treatment approaches for cancer.

Introduction

The immune system can recognise transformed cells, and both innate and adaptive immune reactions to cancer have long been described. The tumour microenvironment contains macrophages, neutrophils, mast cells, myeloid-derived suppressor cells, dendritic cells (DCs), natural killer (NK) cells, and T and B lymphocytes.¹ Tumour-associated antigens and T lymphocytes that are able to recognise tumour-specific antigens have been described.2, 3, 4, 5 Tumour-associated macrophages and tumour-infiltrating leucocytes accumulate within neoplastic tissue.⁶ Activated immune effector cells, such as NK cells, cytotoxic T cells, and macrophages, are present both at the tumour site and in the circulation of patients with cancer, but immune responses against cancer seem to be dysfunctional and tumours progress despite existing immunological activity.5, 7

The possibility of local tumour immune escape or even tumour-induced immune suppression has been studied and discussed in detail.7, 8, 9 The term immunoediting has been used to describe the immunological selection of resistant tumour variants by elimination of immunosensitive malignant cells.10, 11 The resulting avoidance of immune destruction has been defined as a hallmark of cancer.¹² An inflammatory microenvironment seems to be a consistent component of malignant tumours, suggesting the presence of a cancer-related immune reaction.1, 6 As summarised by Mantovani and colleagues,⁶ there is increasing evidence that inflammation contributes to the development of cancer and also that cancer seems to directly promote the generation of an inflammatory microenvironment. Trinchieri¹³ stated in a recent review of the links between inflammation and cancer that “the class of inflammation and immunity that is responsible for tumour initiation and early progression is likely also to be the same class that makes the immune system unable to destroy the tumours successfully”. Whether inflammatory conditions increase the local cancer risk or if genetic alterations such as oncogenes cause inflammation and neoplasia is so far unclear,⁶ but that cancer cells actively interfere with the patient's immune system has been established. Several recent reviews summarised the cellular interaction between cancer and the immune system in the tumour microenvironment.7, 8, 9 However, with respect to the heterogeneity of cancer, a theory of a common or uniform immune reaction pattern seems so far to be obsolete.

Cytokines are intricately involved in all immune reactions. Since cytokines interfere directly or indirectly with each other's expression, the isolated effect of one cytokine can seem less relevant unless assessed in the context of its position in the hierarchy of a defined cascade. Cytokine interactions comprise sophisticated interdependent positive and negative feedback mechanisms, thus providing homoeostasis and immune control. Any attempt to outline these feedbacks in a model is necessarily simplistic because a more accurate representation of a large feedback system, ideally through a mathematical model, would create a non-linear and practically unpredictable relation. Isolated cytokine reactions have been described in individual tumour types, but so far no systematic model of the cytokine-related immune reaction in patients with cancer has been developed.

In summary, there is increasing evidence from experimental studies that malignant tumours utilise local mechanisms within their microenvironment to prevent activation of immunological effector functions, thereby protecting the tumour from a potential immune attack.8, 9

I undertook the present Review of cytokine interactions in patients with cancer with an emphasis on the immunological mechanisms that are associated with clinical disease progression. The approach is intentionally descriptive. Because of this narrow focus, several aspects of cytokine activation and signalling mechanisms were not considered. The format of the present Review is an initial step to detect common patterns in the immune response in patients with cancer, which so far might not be fully represented in experimental systems. The aim of the present Review was to point out reproducible and consistent features among reported clinical data in many and unrelated cancer types. By this translational approach, I attempt to use the context of recent experimental data to interpret the clinical findings. The working hypothesis here is that the emerging cytokine pattern potentially is a manifestation of the underlying systemic paraneoplastic immune response that seems to be a consistent pattern in patients with cancer. Further specification of this system could offer novel therapeutic approaches.⁷