Fredrik Edfors

A pathology atlas of the human cancer transcriptome.

Modeling the cancer transcriptome Recent initiatives such as The Cancer Genome Atlas have mapped the genome-wide effect of individual genes on tumor growth. By unraveling genomic alterations in tumors, molecular subtypes of cancers have been identified, which is improving patient diagnostics and treatment. Uhlen et al. developed a computer-based modeling approach to examine different cancer types in nearly 8000 patients. They provide an open-access resource for exploring how the expression of specific genes influences patient survival in 17 different types of cancer. More than 900,000 patient survival profiles are available, including for tumors of colon, prostate, lung, and breast origin. This interactive data set can also be used to generate personalized patient models to predict how metabolic changes can influence tumor growth. Science, this issue p. eaan2507 A systems biology approach should allow genome-wide exploration of the effect of individual proteins on cancer clinical outcomes. INTRODUCTION Cancer is a leading cause of death worldwide, and there is great need to define the molecular mechanisms driving the development and progression of individual tumors. The Hallmarks of Cancer has provided a framework for a deeper molecular understanding of cancer, and the focus so far has been on the genetic alterations in individual cancers, including genome rearrangements, gene amplifications, and specific cancer-driving mutations. Using systems-level approaches, it is now also possible to define downstream effects of individual genetic alterations in a genome-wide manner. RATIONALE In our study, we used a systems-level approach to analyze the transcriptome of 17 major cancer types with respect to clinical outcome, based on a genome-wide transcriptomics analysis of ~8000 individual patients with clinical metadata. The study was made possible through the availability of large open-access knowledge-based efforts such as the Cancer Genome Atlas and the Human Protein Atlas. Here, we used the data to perform a systems-level analysis of 17 major human cancer types, describing both interindividual and intertumor variation patterns. RESULTS The analysis identified candidate prognostic genes associated with clinical outcome for each tumor type; the results show that a large fraction of cancer protein-coding genes are differentially expressed and, in many cases, have an impact on overall patient survival. Systems biology analyses revealed that gene expression of individual tumors within a particular cancer varied considerably and could exceed the variation observed between distinct cancer types. No general prognostic gene necessary for clinical outcome was applicable to all cancers. Shorter patient survival was generally associated with up-regulation of genes involved in mitosis and cell growth and down-regulation of genes involved in cellular differentiation. The data allowed us to generate personalized genome-scale metabolic models for cancer patients to identify key genes involved in tumor growth. In addition, we explored tissue-specific genes associated with the dedifferentiation of tumor cells and the role of specific cancer testis antigens on a genome-wide scale. For lung and colorectal cancer, a selection of prognostic genes identified by the systems biology effort were analyzed in independent, prospective cancer cohorts using immunohistochemistry to validate the gene expression patterns at the protein level. CONCLUSION A Human Pathology Atlas has been created as part of the Human Protein Atlas program to explore the prognostic role of each protein-coding gene in 17 different cancers. Our atlas uses transcriptomics and antibody-based profiling to provide a standalone resource for cancer precision medicine. The results demonstrate the power of large systems biology efforts that make use of publicly available resources. Using genome-scale metabolic models, cancer patients are shown to have widespread metabolic heterogeneity, highlighting the need for precise and personalized medicine for cancer treatment. With more than 900,000 Kaplan-Meier plots, this resource allows exploration of the specific genes influencing clinical outcome for major cancers, paving the way for further in-depth studies incorporating systems-level analyses of cancer. All data presented are available in an interactive open-access database (www.proteinatlas.org/pathology) to allow for genome-wide exploration of the impact of individual proteins on clinical outcome in major human cancers. Schematic overview of the Human Pathology Atlas. A systems-level approach enables analysis of the protein-coding genes of 17 different cancer types from ~8000 patients. Results are available in an interactive open-access database. Cancer is one of the leading causes of death, and there is great interest in understanding the underlying molecular mechanisms involved in the pathogenesis and progression of individual tumors. We used systems-level approaches to analyze the genome-wide transcriptome of the protein-coding genes of 17 major cancer types with respect to clinical outcome. A general pattern emerged: Shorter patient survival was associated with up-regulation of genes involved in cell growth and with down-regulation of genes involved in cellular differentiation. Using genome-scale metabolic models, we show that cancer patients have widespread metabolic heterogeneity, highlighting the need for precise and personalized medicine for cancer treatment. All data are presented in an interactive open-access database (www.proteinatlas.org/pathology) to allow genome-wide exploration of the impact of individual proteins on clinical outcomes.

Gene-specific correlation of RNA and protein levels in human cells and tissues

An important issue for molecular biology is to establish whether transcript levels of a given gene can be used as proxies for the corresponding protein levels. Here, we have developed a targeted proteomics approach for a set of human non‐secreted proteins based on parallel reaction monitoring to measure, at steady‐state conditions, absolute protein copy numbers across human tissues and cell lines and compared these levels with the corresponding mRNA levels using transcriptomics. The study shows that the transcript and protein levels do not correlate well unless a gene‐specific RNA‐to‐protein (RTP) conversion factor independent of the tissue type is introduced, thus significantly enhancing the predictability of protein copy numbers from RNA levels. The results show that the RTP ratio varies significantly with a few hundred copies per mRNA molecule for some genes to several hundred thousands of protein copies per mRNA molecule for others. In conclusion, our data suggest that transcriptome analysis can be used as a tool to predict the protein copy numbers per cell, thus forming an attractive link between the field of genomics and proteomics.

Immunoproteomics Using Polyclonal Antibodies and Stable Isotope–labeled Affinity-purified Recombinant Proteins

The combination of immuno-based methods and mass spectrometry detection has great potential in the field of quantitative proteomics. Here, we describe a new method (immuno-SILAC) for the absolute quantification of proteins in complex samples based on polyclonal antibodies and stable isotope–labeled recombinant protein fragments to allow affinity enrichment prior to mass spectrometry analysis and accurate quantification. We took advantage of the antibody resources publicly available from the Human Protein Atlas project covering more than 80% of all human protein-coding genes. Epitope mapping revealed that a majority of the polyclonal antibodies recognized multiple linear epitopes, and based on these results, a semi-automated method was developed for peptide enrichment using polyclonal antibodies immobilized on protein A–coated magnetic beads. A protocol based on the simultaneous multiplex capture of more than 40 protein targets showed that approximately half of the antibodies enriched at least one functional peptide detected in the subsequent mass spectrometry analysis. The approach was further developed to also generate quantitative data via the addition of heavy isotope–labeled recombinant protein fragment standards prior to trypsin digestion. Here, we show that we were able to use small amounts of antibodies (50 ng per target) in this manner for efficient multiplex analysis of quantitative levels of proteins in a human HeLa cell lysate. The results suggest that polyclonal antibodies generated via immunization of recombinant protein fragments could be used for the enrichment of target peptides to allow for rapid mass spectrometry analysis taking advantage of a substantial reduction in sample complexity. The possibility of building up a proteome-wide resource for immuno-SILAC assays based on publicly available antibody resources is discussed.

Immunocapture strategies in translational proteomics

Aiming at clinical studies of human diseases, antibody-assisted assays have been applied to biomarker discovery and toward a streamlined translation from patient profiling to assays supporting personalized treatments. In recent years, integrated strategies to couple and combine antibodies with mass spectrometry-based proteomic efforts have emerged, allowing for novel possibilities in basic and clinical research. Described in this review are some of the field’s current and emerging immunocapture approaches from an affinity proteomics perspective. Discussed are some of their advantages, pitfalls and opportunities for the next phase in clinical and translational proteomics.

Solid-phase cloning for high-throughput assembly of single and multiple DNA parts

We describe solid-phase cloning (SPC) for high-throughput assembly of expression plasmids. Our method allows PCR products to be put directly into a liquid handler for capture and purification using paramagnetic streptavidin beads and conversion into constructs by subsequent cloning reactions. We present a robust automated protocol for restriction enzyme based SPC and its performance for the cloning of >60 000 unique human gene fragments into expression vectors. In addition, we report on SPC-based single-strand assembly for applications where exact control of the sequence between fragments is needed or where multiple inserts are to be assembled. In this approach, the solid support allows for head-to-tail assembly of DNA fragments based on hybridization and polymerase fill-in. The usefulness of head-to-tail SPC was demonstrated by assembly of >150 constructs with up to four DNA parts at an average success rate above 80%. We report on several applications for SPC and we suggest it to be particularly suitable for high-throughput efforts using laboratory workstations.

Targeting CDK2 overcomes melanoma resistance against BRAF and Hsp90 inhibitors

Novel therapies are undergoing clinical trials, for example, the Hsp90 inhibitor, XL888, in combination with BRAF inhibitors for the treatment of therapy‐resistant melanomas. Unfortunately, our data show that this combination elicits a heterogeneous response in a panel of melanoma cell lines including PDX‐derived models. We sought to understand the mechanisms underlying the differential responses and suggest a patient stratification strategy. Thermal proteome profiling (TPP) identified the protein targets of XL888 in a pair of sensitive and unresponsive cell lines. Unbiased proteomics and phosphoproteomics analyses identified CDK2 as a driver of resistance to both BRAF and Hsp90 inhibitors and its expression is regulated by the transcription factor MITF upon XL888 treatment. The CDK2 inhibitor, dinaciclib, attenuated resistance to both classes of inhibitors and combinations thereof. Notably, we found that MITF expression correlates with CDK2 upregulation in patients; thus, dinaciclib would warrant consideration for treatment of patients unresponsive to BRAF‐MEK and/or Hsp90 inhibitors and/or harboring MITF amplification/overexpression.

A Protein Standard That Emulates Homology for the Characterization of Protein Inference Algorithms

A natural way to benchmark the performance of an analytical experimental setup is to use samples of known composition and see to what degree one can correctly infer the content of such a sample from the data. For shotgun proteomics, one of the inherent problems of interpreting data is that the measured analytes are peptides and not the actual proteins themselves. As some proteins share proteolytic peptides, there might be more than one possible causative set of proteins resulting in a given set of peptides and there is a need for mechanisms that infer proteins from lists of detected peptides. A weakness of commercially available samples of known content is that they consist of proteins that are deliberately selected for producing tryptic peptides that are unique to a single protein. Unfortunately, such samples do not expose any complications in protein inference. Hence, for a realistic benchmark of protein inference procedures, there is a need for samples of known content where the present proteins share peptides with known absent proteins. Here, we present such a standard, that is based on E. coli expressed human protein fragments. To illustrate the application of this standard, we benchmark a set of different protein inference procedures on the data. We observe that inference procedures excluding shared peptides provide more accurate estimates of errors compared to methods that include information from shared peptides, while still giving a reasonable performance in terms of the number of identified proteins. We also demonstrate that using a sample of known protein content without proteins with shared tryptic peptides can give a false sense of accuracy for many protein inference methods.

High Cell Density Perfusion Culture has a Maintained Exoproteome and Metabolome

The optimization of bioprocesses for biopharmaceutical manufacturing by Chinese hamster ovary (CHO) cells can be a challenging endeavor and, today, heavily relies on empirical methods treating the bioreactor process and the cells as black boxes. Multi-omics approaches have the potential to reveal otherwise unknown characteristics of these systems and identify culture parameters to more rationally optimize the cultivation process. Here, the authors have applied both metabolomic and proteomic profiling to a perfusion process, using CHO cells for antibody production, to explore how cell biology and reactor environment change as the cell density reaches ≥200 × 106  cells mL-1 . The extracellular metabolic composition obtained in perfusion mode shows a markedly more stable profile in comparison to fed-batch, despite a far larger range of viable cell densities in perfusion. This stable profile is confirmed in the extracellular proteosome. Furthermore, the proteomics data shows an increase of structural proteins as cell density increases, which could be due to a higher shear stress and explain the decrease in cell diameter at very high cell densities. Both proteomic and metabolic results shows signs of oxidative stress and changes in glutathione metabolism at very high cell densities. The authors suggest the methodology presented herein to be a powerful tool for optimizing processes of recombinant protein production.

Enhanced validation of antibodies for research applications

There is a need for standardized validation methods for antibody specificity and selectivity. Recently, five alternative validation pillars were proposed to explore the specificity of research antibodies using methods with no need for prior knowledge about the protein target. Here, we show that these principles can be used in a streamlined manner for enhanced validation of research antibodies in Western blot applications. More than 6,000 antibodies were validated with at least one of these strategies involving orthogonal methods, genetic knockdown, recombinant expression, independent antibodies, and capture mass spectrometry analysis. The results show a path forward for efforts to validate antibodies in an application-specific manner suitable for both providers and users.Five validation pillars have been proposed to verify the specificity of research antibodies. Here the authors screen 6,000 antibodies from the Human Protein Atlas with these methods to provide an antibody validation resource for providers and users.