Trials and Tribulations of Cancer Immunotherapy

This is the authors' abstract. We don't add key points for this paper.

Since the publication 7 years ago of the landmark paper by Stupp et al survival of glioblastoma multiforme (GBM) has reached a plateau, with only limited non-significant increases in survival.1 Despite the most aggressive therapies this devastating malignancy has a near perfect recurrence rate, accounting for nearly 15% of all brain tumors in adults between the ages of 45 and 70.2 Over the last few decades however, great strides have been made in cancer immunotherapy, most notably with the creation of the human papilloma virus (HPV) vaccine for cervical cancer, and T-cell therapies for advanced melanoma.3 A shortcoming to these therapies is that they require the cultivation of substantial amounts of lymphocytes using uncommon technical expertise and facilities.3 Less labor-intensive immunotherapy modalities like vaccines may prove more efficacious. Dendritic cells (DCs) theoretically are a prime candidate for cancer vaccines. Specifically, they are known to be the most powerful antigen-presenting cells (APCs) and are largely responsible for the initial sensitization of naïve T-cells to antigens. Plus they are present in nearly all sites of pathogen entry.4 Thus it seemed reasonable to harness the power of these cells to achieve the strongest immunity. Furthermore, DCs can be derived ex vivo from blood monocytes exposed to granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4).5 DC immunotherapy and vaccination has been studied in a wide variety of cancers, Phuphanich et al have recently reported the results of a phase I clinical trial in which patient-derived DCs were pulsed with multiple tumor-associated antigens (TAAs) that are expressed on gliomas and known to be overexpressed in the cancer stem cell (CSC) population.6 Patients with newly diagnosed or recurrent GBM who underwent a gross total resection and were HLA-A1 and/or A-2 positive were enrolled. A well understood mechanism for tumors to avoid immune detection and eradication is through minimizing the presentation of antigens to class I major histocompatability complexes (MHCs). This failure has been partly traced to mutations and deletions affecting the expression of human leukocyte antigen (HLA) molecules.7-10 The autologous DC vaccine was prepared from monocytes obtained through leukapheresis and cultured ex vivo. To target the DCs to the CSC population HLA-1 restricted TAA's MAGE1 and AIM-2, and HLA-2 restricted TRP-2, gp100, HER2, and IL13Ra2 were pulsed for 16 to 20 hours. Subsequently, sterility was tested and patients were then given 3 injections of the vaccine, and monitored for adverse events (AE). Upon analysis of the expression of the TAA's targeted by the vaccine, it was shown that expression of the HLA-1 restricted TAA's significantly correlated with progression-free survival (PFS) and overall survival (OS). Additionally, the expression of HER2 and gp100 appeared to exhibit a trend towards longer PFS and OS. In response to vaccine administration, A2 epitopes showed significant downregulation (P = .023; Fisher's exact test) relative to upregulation in 3 patients where responses were positive. Comparatively, the downregulation A1 epitope AIM2 was not significant yielded a P value of .21 (Fisher's exact test). The characterization of GBM CSCs has lead to the defining surface marker of this putative cell population: CD133. Cells expressing this marker (CD133+) have been shown to be more resistant to radiation11 and chemotherapy,12 and be correlated to a poor clinical outcome.6 Moreover, CD133 positivity has been reported to increase in recurrence, possibly owing to the severity and rapid clinical degradation. Intriguingly, the vaccine illustrated a preferential targeting of CD133 positive cells. Previous studies by this group and others have consistently shown increased CD133 in recurrent tumors, especially in recurrent tumors even after TAA loaded DC vaccination.6 In newly diagnosed and recurrent patients, the CD133 expression was shown to markedly decrease with 1 recurrent and 1 newly diagnosed patient becoming CD133 negative. Analysis of immune responses 7 days before vaccine administration compared to samples obtained 56 days after the first vaccine administration showed a 33% positive vaccine response of >.5-fold increase. PFS and OS and long-term survivors exhibiting increasing trends towards better survival, however they were not significant. At a median follow-up of 40.1 months, 6 of the 16 newly diagnosed patients showed no evidence of tumor recurrence. Median PFS and OS in the newly diagnosed patients were 16.9 months and 38.4 months respectively. Despite these median survival times being comparable to current survival, the ICT-107 vaccine illustrated an impressive percent of responders, exhibiting an increased immune response, 33%, higher than recent immunotherapy vaccines trials. A central problem in developing cancer vaccines is that all immunotherapies, vaccines included, are dependent on their ability to generate immune responses to specific and defined targets that must be presented by the tumor cells. Historically this has been the greatest challenge in creating novel vaccines. The immunoediting hypothesis13 may partly help to explain the limited success of recent cancer vaccines, and the difficulty in targeting tumor cells. Furthermore, the full potential of dendritic cells may not have been fully exploited or the hypothesis that DCs are not an effective pathway to pursue may hold true. Careful re-evaluation of current vaccine design is crucial; the natural course of DCs and their role in the immune system all must be taken into account. Equally important, soluble factors such as IL-12 must be considered particularly for their immuno-modulating capabilities.3 Finally, clinical trials need to incorporate validated prognostic methods (eg, delayed-type hypersensitivity responses, T-cell cytotoxicity assay) to monitor anticancer immune activity in vivo as an intermediary measure of efficacy.4 The notion of inducing the body to fight the human body, the central notion of the immune system, works for the copious other maladies that affect the human condition. Advances in our understanding of neuroimmunology will greatly increase our efficacy with cancer vaccines and reverberate in rational vaccine development, clinical trial designs, and analyses.Figure 1: Schema of treatment schedule and events (A) The correlation of quantitative antigen expression on primary tumor from new diagnosed patients (n = 13) with progression-free survival (PFS; left set of 6 graphs) and overall survival (OS; right set of 6 graphs). Logarithmic plots of antigen expression were determined by qRT-PCR. A significant correlation between HLA-1 antigens and PFS and OS was observed while gp100 and HER2 antigens trended towards a positive correlation in both PFS and OS. (B) Kaplan-Meier estimates of PFS (left panel) and OS (right panel) for newly diagnosed patients (n = 16) (Dotted line illustrates 95% CI; hash marks are for censored patients.) (C) CD133 expression by RT-PCR in primary tumor and subsequent tumor samples from newly diagnosed and recurrent patients. Expression is calculated relative to GADPH. CD133 expression was negative or decreased in subsequently after vaccine administration (D). With kind permission from Springer Science+Business Media: Cancer Immunology Immunotherapy, Phase I trial of a multi-epitope-pulsed dendritic cell vaccine for patients with newly diagnosed glioblastoma, 2012 Jul 31. [Epub ahead of print].