Session 1

Abstract

1.1

Identifying and Targeting Tumor Neoantigens

Catherine Wu

Department of Medical Oncology, Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA, USA

With the recent availability of novel immunologic agents, priority has shifted to understanding the mechanisms of and predicting responses to each treatment. At the heart of cancer and host immune cell interactions is the tumor antigen and host antigen-specific T cell interaction, with the cytotoxic T cell-cognate antigen interaction forming the mechanistic basis for immune-mediated recognition and the killing of malignant cells. While the search for immunogenic tumor antigens has been the subject of decades-long studies, multiple lines of evidence have convincingly demonstrated tumor neoantigens as an important class of immunogenic tumor antigens. Neoantigens arise from amino acid changes encoded by somatic mutations in the tumor cell and have the potential to bind to and be presented by personal HLA molecules. Using next-generation sequencing approaches, we can now systematically identify mutations leading to amino acid changes that can be potentially recognized immunologically through the implementation of neoantigen discovery pipelines. In recent studies, we have demonstrated that neoantigen load is associated with clinical outcome to immune-based therapies, and neoantigens can be safely and feasibly targeted to generate customized cancer vaccines. Within this therapeutic landscape, the identification of HLA-bound peptides by liquid chromatography-tandem mass spectrometry (LC-MS/MS) is poised to transform our understanding of rules underlying antigen presentation, through the generation of data to reveal subdominant binding motifs and to understand the factors critical to epitope presentation, such as protein cleavage and gene expression. A MS-based approach to directly identify antigen targets from tumor cells also promises to impact strategies to develop and implement personal cancer vaccines and other immunotherapies.

1.2

Mapping the Glycoproteome with Activated Ion Electron Transfer Dissociation

Nicholas M. Riley ^1,2, Alexander S. Hebert ¹, Michael S. Westphall ¹, Joshua J. Coon ^1,2,3,4

¹Genome Center of Wisconsin, Departments of ²Chemistry and ³Biomolecular Chemistry, University of Wisconsin-Madison, Madison, WI, USA; ⁴Morgridge Institute for Research, Madison, WI, USA

Protein glycosylation is a prevalent, chemically complex, and biologically diverse post-translational modification (PTM) involved in a wide array of intra- and inter-cellular functions. Approximately half of all expressed proteins undergo glycosylation, and this heterogeneous modification accounts for the greatest proteome diversity over any other PTM. Changes in protein glycosylation are associated with cellular proliferation, inter-cellular communication, and metabolic processes, making the characterization of the cellular landscape of protein glycosylation integral to advancing our understanding of cell biology. Glycan microheterogeneity, i.e., different glycans modifying the same glycosite, makes glycan identity at a given site crucial to the biological context of the modification. This unique feature of glycosylation makes analysis of intact glycopeptides imperative for glycoproteome characterization, but current analytical tools are ill-suited for this task. Tandem mass spectrometry (MS) is an ideal platform to advance glycoproteomic technology, but current dissociation methods are often suitable only for characterization of either peptide or glycan moieties. This mandates multiple analyses of the same precursor ions that limit throughput and challenge data interpretation. We have developed a tandem MS dissociation method called activated ion-electron transfer dissociation (AI-ETD) that addresses several of challenges of intact glycopeptide analysis. Through the use of concurrent ion-ion reactions and infrared photo-activation, AI-ETD can access glycan and peptide information from intact glycopeptides in a single MS/MS scan, and with this AI-ETD provides (1) improved product ion generation for peptide backbone sequencing, (2) higher MS/MS success rates to sequence more glycopeptides per experiment, and (3) valuable fragmentation for glycan composition determination. Here we showcase AI-ETD for large-scale intact glycopeptide characterization on the newest generation of Orbitrap instruments. We show that >2000 localized N-glycosites can be confidently identified from approximately 30,000 localized N-linked glycopeptide spectral matches (>7,500 unique) in mouse brain tissue. This represents a more than 3-fold increase over recent studies in the number of glycopeptides and glycosites that can be identified via intact glycopeptide analyses, and it rivals, if not outmatches, glycoproteomic studies that analyze only deglycosylated peptides.

1.3

Highly multiplexed and targeted imaging of tissues by mass cytometry

Bernd Bodenmiller

Institute of Molecular Life Science, University of Zurich, Switzerland

Cancer is a tissue disease. Heterogeneous cancer cells and normal stromal and immune cells form a dynamic ecosystem that evolves to support tumor expansion and ultimately tumor spread. The complexity of this dynamic system is the main obstacle in our attempts to treat and heal the disease. The study of the tumor ecosystem and its cell-to-cell communications is thus essential to enable an understanding of tumor biology, to define new biomarkers to improve patient care, and ultimately to identify new therapeutic routs and targets. To study and understand the workings of the tumor ecosystem, highly multiplexed image information of tumor tissues is essential. Such multiplexed images will reveal which cell types are present in a tumor, their functional state, and which cell-cell interactions are present. To enable multiplexed tissue imaging, we developed imaging mass cytometry (IMC). IMC is a novel imaging modality that uses metal isotopes of defined mass as reporters and currently allows to visualize over 50 antibodies simultaneously on tissues with subcellular resolution. In the near future we expect that over 100 markers can be visualized. We applied IMC for the analysis of breast cancer samples in a quantitative manner. To extract biological meaningful data and potential biomarkers from this dataset, we developed a novel computational pipeline geared for the interactive and automated analysis of large scale, highly multiplexed tissues image datasets. Our analysis reveals a surprising level of inter and intra-tumor heterogeneity and identify new diversity within known human breast cancer subtypes as well as a variety of stromal cell types that interact with them. In summary, our results show that IMC provides targeted, high-dimensional analysis of cell type, cell state and cell-to-cell interactions within the TME at subcellular resolution. Spatial relationships of complex cell states of cellular assemblies can be inferred and potentially used as biomarkers. We envision that IMC will enable a systems biology approach to understand and diagnose disease and to guide treatment.

1.4

Pulsed SILAC and TMT labeling reveals proteoform resolved protein dynamics

Bernhard Küster, Jana Zecha, Chen Meng, Mathias Wilhelm

Technische Universität München, Freising, Germany

The coordination of protein synthesis and degradation for regulating protein abundance is a fundamental requirement for cellular homeostasis. We investigated the feasibility of a workflow combining pulsed SILAC and TMT-labelling for parallel and independent determination of protein degradation and synthesis rates with peptide level resolution. HeLa cells were pulse labelled in four cell culture replicates using SILAC medium supplemented with 13C and 15N containing lysine and arginine. Peptides derived from 10 different pulse time points were modified using TMT10plex reagents and measured on an Orbitrap Fusion Lumos in SPS-MS3 mode. Following data base search using MaxQuant and subsequent data normalization, synthesis and degradation rates were estimated by fitting exponential equations to TMT reporter ion intensities. Synthesis and/or degradation curves were obtained for more than 56,000 peptides mapping to more than 7,300 protein groups. Following further data processing, protein synthesis and degradation rate constants obtained from different replicates showed good reproducibility. Importantly, half-lives determined in this way were comparable to data obtained by a standard pulsed SILAC approach. In contrast, the SILAC-TMT approach overcomes the issue of missing values in the pulse time series. This not only led to higher proteome coverage but also enabled turnover determination at peptide level. The latter feature led to the discovery of differences in turnover of protein splice variants and post-translationally processed proteins revealed by significant rate constant differences of peptides nominally belonging to the same gene product. Our data shows that the commonly used pulsed SILAC concept can be extended to TMT labeling of different pulse time points. Moreover, this multiplexing strategy should also be applicable to disturbed, non-steady-state systems because the separate determination of changes in synthesis and degradation rates can now be independently measured. As a consequence, studying proteome dynamics under differential biological conditions is strongly facilitated, eventually improving our understanding of cellular proteostasis in health and disease.

1.5

Evolution, dynamics and genetics of protein post-translational control

Pedro Beltrao

European Bioinformatics Institute (EMBL-EBI), Cambridge, England, UK

Cells need to constantly adapt to changes in conditions and use post-translational regulation as a fast way to transfer information from sensors to effectors of cellular responses. Advances in mass-spectrometry now allow us to identify post-translational modification (PTMs) sites in large scale and to quantify their changes across different conditions. However, little is known about how the thousands of recently discovered PTMs modulate protein function, coordinate cellular responses and change across species or in disease. We have been working to study the evolution, dynamics and genetic determinants of protein phosphorylation and degradation. We have observed that only a small fraction of phosphosites are ancient in origin and we have suggested that a fraction of these may be neutral. To address this, we are working on methods to rank phosphosites according to their functional relevance by integrating diverse set of evolutionary, structural and regulatory features. In addition, we have been analysing how cell signalling pathways change in cancer by combining genetic and proteomic information from publicly available patient samples. Using these data we have observed that around 25% of genes have copy number changes that are not propagated to changes in protein levels. This gene dosage compensation is prevalent for protein complex subunits and may be due to changes in degradation of free versus bound complex members. We think these studies will give us a better understanding of the degree of functionally relevant PTMs and their dysregulation in cancer.

2.1

Proteomic analysis of cell cycle progression in asynchronous cultures, including mitotic subphases, using PRIMMUS

Tony Ly ¹, Pat Wadsworth ², Angus Lamond ¹

¹University of Dundee, Scotland, UK; ²University of Massachussetts-Amherst, MA USA

The temporal regulation of protein abundance and post-translational modifications is a key feature of cell division. Recently, we analysed gene expression and protein abundance changes during interphase under minimally perturbed conditions (Ly et al. 2014; Ly et al. 2015). Here we show that by using specific intracellular immunolabeling protocols, FACS separation of interphase and mitotic cells, including mitotic subphases, can be combined with proteomic analysis by mass spectrometry. Using this PRIMMUS (PRoteomic analysis of Intracellular iMMUnolabeled cell Subsets) approach, we now compare protein abundance and phosphorylation changes in interphase and mitotic fractions from asynchronously growing human cells. We identify a set of 115 phosphorylation sites increased during G2, which we term 'early risers'. This set includes phosphorylation of S738 on TPX2, which we show is important for TPX2 function and mitotic progression. Further, we use PRIMMUS to provide a proteome-wide analysis of protein abundance remodeling between prophase, prometaphase and anaphase.

2.2

Proteome dynamics

Matthias Selbach

Max-Delbrück-Center for Molecular Medicine (MDC), Berlin, Germany

Understanding how the information in the genome is interpreted to produce a specific phenotype is probably the most important question in the post-genomic era. Proteins are the central link between the genome and the phenotype. Therefore, studying proteome dynamics can help to understand (i) how genomic information is processed to yield a specific proteome and (ii) how the proteins expressed at a certain condition affect the phenotype. I will give two examples how mass spectrometry can be used to study proteome dynamics. First, I will present global pulse-chase experiments to quantify the kinetics of cellular protein degradation. These data show that many proteins are degraded non-exponentially. Non-exponentially degraded proteins are often members of multiprotein complexes which are produced in super-stoichiometric amounts. This has important consequences for complex assembly and aneuploidy. Second, I will show how quantitative proteomics can help to understand the molecular mechanisms how disease-associated mutations cause disease. To this end, I will show the results of a proteomic screen designed to investigate how mutations in intrinsically disordered regions of proteins affect protein-protein interactions.

2.3

Quantitative profiling of highly polymorphic multigene products

Robert J. Beynon

Institute of Integrative Biology, University of Liverpool, England, UK

All mass spectrometry-based strategies for quantification of individual proteins within a proteome rely on at least one peptide (SRM based methods) and often more peptides (label-free quantification) that are used as surrogates for protein abundance. Both approaches can be effective, particularly in the context of relative quantification. However, both methods become more challenging when resolving polymorphic protein families, in which similarities between individual members of the family reduce the opportunities for unique quantitative solutions. We are particularly interested in the families of lipocalins known as MUPs (major urinary proteins) in rodents, small (18–19kDa) beta-barrel structures that play multiple roles in chemical communication between individuals. The house mouse (Mus musculus domesticus) possesses at least 21 genes encoding these proteins, and the sequence differences between individual mature gene products can be sufficiently small to render quantification a major challenge. Further, the allelic and polymorphic heterogeneity in MUPs is pronounced in wild house mice, adding the additional (impossible?) challenge of quantification when the protein sequence might not be known. A final complication is common to all urinalysis; the need to correct for urine dilution. In my presentation, I will introduce this MUP protein family, discuss the remarkable range of behavioural effects that they elicit and set the scene for the requirement for accurate, absolute quantification of each unique protein. This cannot be achieved by a 'one size fits all' strategy, and we approach this by a combination of analyses of MUPs as intact proteins, as immunoreactive molecules and as a source of peptides for targeted quantification.