Optimization of dendritic cell based immunotherapies

Abstract

Dendritic cells (DC) are professional antigen presenting cells that continuously sample the environment; capture and process antigens; and transmit gathered information to T cells (1–3). In presence of danger signals, DC become activated and trigger an inflammatory response against processed antigens, otherwise they remain in an immature state that lead to immune tolerance (4). Given their pivotal role in shaping the immune system, DC are considered among the most promising cell-based immunotherapeutic approach in cancer setting aiming at activating an immune response against tumor associated antigens (TAA) (1). Among different approaches, the most explored one is vaccination with ex vivo generated monocyte-derived DC (5). Over the last twenty years, in fact, many strategies have been developed to differentiate monocytes into immature DC and mature DC (6). However, despite of the large number of studies and over 300 clinical trials on ex vivo-generated DC vaccines, clinical results so far obtained were disappointing. Even though these vaccines proved to be safe and capable to induce immune and clinical response in patient with melanoma, prostate carcinoma, glioma, and renal cell carcinoma, the overall response rate was usually below 15% (7). Many reasons have been hypothesized for such low response rates, among which the generation of DC with suboptimal potency is considered the most relevant. It’s not known yet how to generate the most potent DC; furthermore, differences in clinical setting, study design, sources of antigens, and route of administration make it almost impossible to compare results from previously conducted trials in order to clearly delineate the shared determinants of in vivo efficacy of DC-based vaccines. Differently from previous attempts to optimize DC by modifying differentiation and/or maturation procedures, this project explores the possibility to identify factors affecting DC potency/efficacy in vivo in order to gain knowledge of molecular determinants essential for specific DC and that can thus be used for quality assessment of manufactured DC. Therefore, this project aimed at identifying factors affecting DC consistency and candidate molecular biomarkers of consistency, potency and efficacy of GMP manufactured DC. In the first part of this project, we therefore analyzed factors affecting DC consistency as well as genes and proteins mostly affected by these factors. We analyzed a specific type of DC that is being tested clinically that are DC differentiated by GM-CSF and IL-4 and matured with LPS and IFN-gamma (LIg-DC). We showed that even when highly standardized procedures are used to generate LIg-DC, manufacturing, intra-donor and inter-donor related factors may affect DC phenotype with the last one being the most relevant (8). Interestingly, these three factors mainly affected expression level of different genes and, while intra-donor variability diminished during differentiation (probably because of strong differentiation signals), inter-donor variability increased upon differentiation/maturation. Additionally, we observed that, while most of the well-known and usually tested DC markers (e.g., CD80, CD86, CD83, HLA-DR) did not show any differences in expression among LIg-DC generated at different times from different donors, the expression of several genes and the levels of several key secreted cytokines and chemokines showed significant variability among LIg-DC products. In particular, among top variable genes, many are likely to be functional important for LIg-DC and their expression correlated with the levels of inflammatory IL-12 as well as other key chemokines, such as MDC, MIG and CXCL10. Then, in order to analyze whether such variability can be responsible of functional differences in vivo, we characterized from a molecular as well as immunophenotypic point of view LIg-DC vaccines administered to stage D0 prostate cancer patients. We observed a strong correlation between DC phenotype and development of clinical and immunological response in patient after vaccination. In particular, we identified a 303-gene signature made up of several well-known tolerogenic DC factors, such as CD14 and IL-10, that was capable to discriminate DC of patients that later showed clinical/immunological response from the ones of non-responders. The differential expression of CD14 and IL-10 was confirmed at the proteomic level and we also observed that MCP-1 and MDC protein levels correlated with the expression of the tolerogenic gene signature. Even though IL-10 secretion levels were able to predict strong immunological responses, it was only by combining CD14, IL-10, MCP-1 and MDC protein measures that it was possible to obtain an index able to replace the tolerogenic gene expression signature in its ability to discriminate both clinical and strong immunological responses. In the final part of the project, we explored whether monocyte-derived DC differentiated in presence of GM-CSF and interferon-alpha (IFNa-DC) show patterns of variability similar to LIg-DC and whether biomarkers of efficacy are shared with LIg-DC. The DC used in this study were also manufactured to sustain a phase I clinical study aiming at activating an immune response in advanced melanoma patients. Even IFNa-DC were showing pretty invariable expression levels of major histocompatibility complex class I and class II molecules, as well as co-stimulatory receptors CD80 and CD11c marker. However, we did observe high variation in the level of expression of CD86, CD40, CD83 and CD1a among IFNa-DC made from different patients. At gene expression level, instead, we did not observe the existence of the tolerogenic signature we detected in LIg-DC, but even in these cells immune response genes showed high level of variability, therefore pointing to functional differences in DC vaccines. Also proteomic analysis suggested that lot-to-lot variability shown by IFNa-DC affects different cytokines and chemokines compared to LIg-DC. Altogether, this project developed a methodological framework for the identification of biologically-relevant quality control markers of DC by combining genomic and proteomic analysis. When applied to clinical DC, such approach was able to identify genes and proteins that correlated with clinical and immunological response and that can therefore be used as efficacy biomarkers of LIg-DC. However, as highlighted from the analysis of IFNa-DC, such newer markers are specific for DC used. On a broader range, these results strongly support the need for in-depth analysis of DC for the identification of newer quality assessment markers and factors essential for DC activity in vivo. Once identified, these markers can be used for the advancement of DC immunotherapies and to foster their implementation in clinic.