Current Cancer Therapy Reviews - Volume 4, Issue 2, 2008

Earlier this year our dear friend and colleague Professor Tony Dodi died suddenly. He was a Scientist of immense integrity and worked with many of the authors who have contributed to the current series of articles focussing on Cancer Biomarkers. His laboratory at Nottingham Trent University (NTU) authored the article entitled “NK Cells, monocytes and Non T-cell biomarkers as prognostic indicators in cancer immunotherapy” in issue 1 and he contributed enormously to the progress and wellbeing of the ENACT Research programme funded by the EU under FP6. During his career, Tony worked with Robert Lachler and Richard Bachelor at The Royal Postgraduate Medical School, Hammersmith London, prior to joining Professor Alejandro Madrigal at the Anthony Nolan Research Institute. In 2005, Tony was appointed as Professor of Haemato-oncology and Head of Leukaemia Research at NTU. He also continued to head the Immunotherapy Group at the Anthony Nolan Research Institute in London. His legacy exists in the research he undertook in the field of Cancer Immunotherapy and Transplantation Immunology and he was recently involved in the development of the Anthony Nolan Cord Blood Programme, which was developed as a result of his enthusiasm and energy. From the outset, I admired his research ability and determination and consider it to be a distinct privilege as well as pleasure to have known Tony as a colleague and close friend. He is sorely missed by all who knew him and we wish to dedicate this series of papers to his memory. We will not forget Tony's contribution to our research endeavours.

Cancer is a genetic disorder resulting in a change in cell growth pattern, cell phenotype and functional traits. Tissue integrity is challenged as a result of malignancy, placing the host-cell equilibrium under threat. In particular the immune system may respond to cancer in several ways: (1) to the benefit of the host by mediating the destruction of the malignant cells, (2) provide the tumour with an additional growth advantage by changing the tissue micro-environment to enhance tumour cell survival and growth potential or (3) become tolerant to the presence of the tumour through interaction with tumour or host cell derived products, thereby allowing the tumour to develop “unchecked”. Although single biomarkers have been identified for several cancers, the dynamic biological relationship between the host and the cancer will be reflected through extensive and differing genetic, protein and cell response “signatures”, derived from both host and tumour cells. Biomarkers may predict the presence of cancer and associate with diagnosis, prognosis and outcome of therapy. The development of advanced bio-molecular techniques for the interrogation of complex genetic, protein, biological data and clinical parameters has led to new opportunities for defining cancer biomarkers. There is a pressing need for more reliable markers that associate with disease status and importantly with response to therapy. Many cancers are detected late in the disease process when more aggressive therapy is required and although treatment may succeed in eradicating the disease it is often difficult to predict patient outcome following treatment. It is well reported that first line treatment may fail to stem the disease course; thus, providing a more accurate assessment of the likely outcome of therapy would allow appropriate treatment options to be administered at an earlier stage. For these reasons it is important to focus research on the identification of new cancer biomarkers with predictive capability. This publication provides an insight into important research areas of current interest, describing “state of the art” methods that are being used to identify new, disease -associated genes / gene products that have utility as cancer biomarkers. Many of the studies focus on markers detected in response to immunotherapy, for example, where a specific immune response can be defined and monitored, which in some cases may be indicative of patient outcome. This involves the assessment of adaptive Tcell mediated immunity and innate immune mechanisms using phenotypic and functional characteristics or the measurement of soluble serum factors produced either by leukocyte populations or as a consequence of host cell - tumour cell interaction (Pawelec et al., Dodi et al., Schadendorf et al.). Some of the many characteristics of cancer cells, for example the loss of MHC which represents a tumour escape mechanism, have been recognised as indicators of patient prognosis (Garrido et al.). Another important area of research focuses on profiling autoantibody responses against tumour -associated proteins, many of which have been discovered through screening cDNA libraries and antigen arrays. Protein micro-arrays in particular may prove to be beneficial for the rapid screening of patient sera for antibody responses to onco-proteins and normal cellular proteins (Miles et al.). Gene arrays and interrogation for the serum and cancer cell proteome using currently available separation and analytical methods coupled with bioinformatic analysis of complex data sets, allows an integrated approach to be adopted for biomarker discovery. The combined assessment of proteomic, genetic, immunological and clinical data may prove to be of greater value in defining a set of markers with greater accuracy of prediction (Laversin et al.). This issue further continues with reviews on cancer diagnosis, prognosis and predictive responses that are likely to influence therapeutic intervention in the future. The host-cancer relationship is complex and the sequel of pathological mechanisms occurring during progression of the disease will dictate the genetic and protein expression profile that can be detected. Biomarkers may be used to predict the presence of cancer, although there is a continuing need for more reliable indicators of disease and it is unlikely that single biomarkers will prove to be accurate and reliable. The paper by Labarriere et al. describes the current status of research using adoptive cellular therapy, as developed by the group of Francine Jotereau in Nantes. They describe the clinical and biological consequences of adoptive T-cell transfer and the clonal response and T-cell phenotype / genotype associating with long term patient survival. Mihaylova et al. further discusses the role of NK associated ligands and receptors as potential markers associating with susceptibility to cancer. The paper includes important concepts of KIR - HLA interactions and NKG2D -Mic proteins, that impact on genetic diversity.

The most common cancer for women in the United Kingdom is breast cancer and it is well known that genomic modifications are responsible for breast carcinogenesis. Major developments in genomics and proteomics have uncovered biomarkers which have a crucial role in this process. To move towards individualized and improved breast cancer care, there is a need for tumor markers that are associated with disease diagnosis, prognosis, treatment and monitoring. General opinion is that biomarkers can bring more accurate and sensitive tools to the clinical settings than conventional approved histopathological parameters. However, clinical applications of breast tumor markers remain quite low in numbers because the data regarding these novel markers for breast cancer is limited and do not allow well-grounded conclusions. This review presents a non exhaustive selection of biomarkers published, some validated, and their uses in breast cancer management.

Prostate cancer remains one of the most common cancers in men and treatment options for patients with advanced or metastatic disease are limited. Immunotherapeutic approaches, such as passive or active immunotherapy, have been shown to be successful in treating some cancer patients. These rely on the identification of cancer specific antigens that can be targeted by the immune system or used to diagnose as early as possible and/or to monitor the disease and its progression. Several prostate cancer antigens have already been identified; these include Prostate-Specific Membrane Antigen (PSMA), Prostate Specific Antigen (PSA), Prostatic Acid Phosphatase (PAP), PAGE-4, PSP 94, differential display 3, prostate androgen-regulated transcript 1 and six-transmembrane epithelial antigen prostate-restricted antigens and these represent potential candidate for immunotherapy against prostate cancer. The following review describes the importance of these antigens with a particular emphasis to the PAP antigen and T21, a recently identified prostate cancer associated antigen, as very promising targets for prostate therapy.

Biomarkers become more and more important tools in clinical diagnosis of cutaneous malignancies. Furthermore, after vigorous validation, biomarkers may have influence on future prognostic classification systems. Clinical and histopathological parameters such as anatomic site, type of the primary tumor, tumor size, invasion depth, ulceration, vascular invasion, lymph node involvement and others are well established for their prognostic value. Additionally, an increasing variety of molecular markers have now been identified and described, providing a potential powerful platform for a more precise diagnosis and a more accurate prognostic sub-grouping of tumor entities. This might lead to future changes in existing classification systems, and might have consequences for the treatment and outcome of skin cancer patients. Recently published gene expression or proteomic profiling data relate to new marker molecules involved in skin cancer pathogenesis, which may, after validation by suitable studies, represent future prognostic or predictive biomarkers in cutaneous malignancies. The aim of this report is to summarize the currently known serologic and newer immunohistochemical biomarker molecules in the most common cutaneous malignancies malignant melanoma, squamous cell carcinoma, and cutaneous lymphoma, particularly emphasizing their prognostic and predictive significance.

The progress in genomics and proteomics resulted in increasing number of tumor-associated antigens (TAA) being discovered as cancer biomarkers and targets for immunotherapy. The key role played by HLA class I antigens in immune reactivity against malignant and virally infected cells via binding to the peptides of TAA and subsequent presentation to cytotoxic T-lymphocytes stimulates interest in the characterization of their expression in tumor cells. Various types of HLA class I alterations with different underlying molecular mechanisms are found in different malignancies. Loss or downregulation of tumor HLA class I antigen expression represents one of the main mechanisms used by cancer cells to evade immunosurveillance since it limits the ability of cytotoxic T-cell to eliminate these cells and reduces the clinical efficacy of T-cell-based cancer therapy. As a result of the immune selection, HLA class I negative variants escape and lead to tumor growth and metastatic colonization. Altered HLA class I expression on malignant cells frequently correlates with poor survival, disease progression and limited response to T-cell-based therapy. Early cancer detection and treatment require more effective cancer biomarkers, or molecular signatures, for diagnosis, prognosis, and therapeutic efficacy. Analysis of the tumor expression of HLA class I antigens as biomarkers of cancer development might help to choose an appropriate treatment protocol and monitor clinical response to cancer immunotherapy.

The immune system plays an important role in eliminating tumors. Accordingly, many immunotherapy protocols have been developed in attempts to boost anti-tumor immune responses in cancer patients. These trials most commonly aim at augmenting T-cell-, especially CD8+ T-cell, responses. In clinical trials performed to date, however, little correlation has been observed between parameters measured to assess cellular immune responses and clinical outcome. Thus, identification of reliable surrogate predictors for evaluation of cancer vaccine efficacy still remains to be accomplished and would be crucial for successfully pursuing clinically significant results. In this review, we discuss the role of different cellular components of the immune system in orchestrating and regulating an anti-tumor immune response, and how to reliably monitor these using polychromatic flow cytometry (FC). The increasing use of multi-parameter FC is expected to be as revolutionary for immunomonitoring as was the introduction of first generation technology. The application of this novel technology and more standardised monitoring assays should help identify biomarkers reliably distinguishing patients who respond to treatment.

Identification and characterisation of tumour associated antigens is a pre-requisite for an effective and targeted immunotherapy. Tumour antigens are used as markers to improve diagnosis and to predict the outcome of a cancer. This review emphasises on a few important cellular target molecules such as indoleamine 2,3-dioxygenase (IDO), interleukin- 15 (IL15), mucin 1 (MUC1) and Major histocompatibility complex class I related chain A and B (MICA/B) which are involved in the downstream pathways of cancer progression. Unique aberrant and over-expression of these molecular targets modulates the physiological and immunological micro-environments in the host tissue. IDO, IL15 and MICA/B provoke NK cell and dendritic cell mediated anti-tumour responses due to their strong binding affinity to activator receptors on the immune cells. The application of these random markers derived from immunological mechanisms may help to improve and delineate some of the predictive indicators obtained with random serum profiles.

During the last few years, adoptive cellular therapy (ACT) — the isolation of antigen-specific lymphocytes, their ex vivo expansion and activation, and subsequent autologous administration — has been tested for treatment of melanoma tumours. Initial ACT used melanoma -infiltrating lymphocytes (TIL) that often contain tumour reactive lymphocytes, of diverse, mostly unknown, specificities [1, 2]. Recently, the identification of melanoma antigens [3] and the development of techniques for selection and expansion of epitope specific T cells has opened the way to the use of tumour antigen specific T cells, and importantly to immune follow up of ACT [4-9]. Despite these advances, several issues remain to address to achieve the major aims of ACT, as the rapid production of clinical grade T cells capable of eliciting a significant destruction of tumour tissue.

Progress in the field of NK cell receptors (NKRs) and their ligands has revolutionized our concept of how NK cells selectively recognize and lyse tumor while sparing normal cells. Major families of NKRs that inhibit and activate NK cells to lyse target cells have been characterized, including killer cell immunoglobulin-like receptors (KIRs), C-type lectins, and natural cytotoxicity receptors. Identification of NK receptor ligands such as HLA, MHC class I chain-related genes A and B - MICA/B, etc. further clarifies the role of NKR/ligand interactions in NK cell function. Given the extensive genomic diversity of KIRs, HLA and MIC, it is reasonable to imagine that genotypes encoding imbalanced inhibitory and activating interactions may contribute to susceptibility or resistance to human diseases, including cancers. This review will discuss two cell-surface receptor/ligand systems involved in NK cell recognition of tumor targets - KIR/HLA and NKG2D/MIC and the potential relevance of their genetic polymorphism in malignancies.

To date, no effective treatment is available for advanced cancers, which remain a major cause of morbidity and mortality. Clearly, there is urgent need to unravel novel biomarkers for early detection. Proteomic approaches for the identification of novel biomarkers for cancer diagnosis and staging have traditionally relied on the identification of differentially expressed proteins between tumour cells and their normal counterparts based on the patterns of protein expression observed by two-dimensional gel electrophoresis (2DE-PAGE). Recent advances in mass spectrometry and bioinformatics and statistical algorithms necessary to interpret mass spectrometric data have revolutionized the approach to defining new tumour markers. Proteomics studies have generated numerous datasets of potential prognostic, diagnostic and therapeutic significance in human cancer. In this review we will discuss the available proteomic technologies their limitations and highlight the key areas of research required for understanding the aetiology of cancer and how they have been used to discover cancer biomarkers.

A lot of experimental data has been produced which suggests that there is a crucial role for glycans during the different stages of melanoma progression. Malignant transformation is associated with various alterations in the glycosylation pattern of the protein glycans. The most common changes are associated with the presence of more branched and hypersialylated N-linked oligosaccharides, re-expression of fetal-type antigens and premature terminated glycans. Some of these changes may provide the tumor cells with an advantage in influencing their social behavior and facilitating metastasis formation so, glycan-target therapy in combination with existing protocols may have an important impact in the treatment of melanoma.

The production of autoantibodies against tumour-derived proteins has been observed in the most, if not all cancer patients therefore they seem to be very attractive biomarkers for the diagnosis or early detection of cancer. Moreover, even if they by themselves have a minor role in the anti-tumour immune response, the autoantibody profile likely reflects the repertoire of activated CD4+ T cells, presumably, including both, the helper cells and the regulatory T cells. Hence, they may turn out to be valuable biomarkers for monitoring patient's response to immunotherapy. So far, their clinical utility has been hampered by the low frequency against each antigen, the inability to differentiate between different types of cancer, large interindividual variations in autoantibody repertoires and ambiguous biological and clinical significance. At least partially these limitations could be overcome by applying high-throughput proteomic techniques to cancer serology that allows definition of a comprehensive set of antigens in each type of cancer and analysing the whole autoantibody profiles in patients' sera. In this review we aim to give an insight into the recent advances in applying various proteomic approaches, including phage display-based SEREX, SERPA, MAPPing and protein microarrays for the identification of tumour-associated antigens and autoantibody profiling, and discuss the relevance of autoantibodies for the early detection of cancer and monitoring patients' response to therapy.

The advent of post genomic technologies and their application to biomedical problems resulted in a massive increase the complexity of data generated. This has resulted in the need for more refined statistical and computational methods for analysis of this type of data Through comprehensive analysis and modelling approaches, structures or classes in the data can be defined, denovo predictive biomarkers identified and clinical decision support systems developed. This review seeks to provide an overview of some of the computational methodologies that have been used for data-mining complex high throughput data for the identification of new biomarkers or improvement of existing biomarkers. Given the literature is vast within the area of computational algorithms we seek to present commonly used methods.

Biomarkers can be used to aid the diagnosis of a disease (particularly early detection of occult tumours), to aid in the staging of disease, to predict outcome of disease, to act as a surrogate for clinical progression and to monitor responses to treatment. It is therefore important when referring to biomarkers to state exactly which of these indications is being sought. There is a real clinical need for biomarkers in all these stages, especially where patient stratification prior to treatment can considerably improve the outcome of clinical trials. Cancer, being a heterogeneous disease, is unlikely to respond to any one specific treatment and biomarkers that could tell us if the patient and their cancer was resistant to certain drugs would be very useful and, indeed, these are being sought. In an analogous way these are even more necessary with regards to vaccines, which would benefit greatly from patient stratification and early prediction of disease response [1].