Proteomics in multiomic strategies

Groups such as those headed by Mathias Uhlén at Stockholm’s Royal Institute of Technology (KTH) and Leroy Hood at the Institute for Systems Biology (ISB) in Seattle have been spearheading the use of large-scale, multiomic studies of healthy populations as a tool to enable knowledge-driven advances in precision medicine. One landmark study from the KTH group demonstrated that individuals generally maintain a remarkably stable molecular profile over time, but that these profiles vary significantly among different people.

By following these “wellness” cohorts longitudinally as diseases develop later in some individuals, invaluable predictive biomarkers can be uncovered. A wellness study from the ISB group, for example, identified a predictive marker for metastatic cancer that could be measured up to 2 years prior to clinical diagnosis.

The combination of genetic and protein expression data (“proteogenomics”) has long been recognized as a highly relevant and powerful combination of molecular analysis. In its simplest form, this approach can be used to calculate the percentage of variation in protein expression among individuals that is due to genetics. This is important information when analyzing the results of population cohorts etc. A good example of this was reported by Ulf Gyllensten’s group from Uppsala University, who described strong effects of genetic and lifestyle factors on protein biomarker variation in a population cohort from northern Sweden.

When cis-pQTL data is combined with phenotypic data in a Mendelian Randomization (MR) analysis, it provides extremely strong evidence that the protein plays a causal role in the disease or biological process being studied.

More recently, proteogenomics is increasingly used to identify specific genetic loci that affect the levels of individual proteins. These associations are known as protein Quantative Trait Loci (pQTLs), and when located very close to the genetic locus encoding the affected protein (“cis-pQTL”), these are powerful tools for gaining actionable insights into disease biology and to drive new drug development. When cis-pQTL data is combined with phenotypic data in a Mendelian Randomization (MR) analysis, it provides extremely strong evidence that the protein plays a causal role in the disease or biological process being studied. This is particularly important for drug target identification, where causality is a prerequisite to take the protein into a drug discovery and development program but cannot be reliably determined from either genomics or proteomics alone.

This powerful approach is being rapidly adopted by the scientific community and has also resulted in a major international consortium dedicated to large-scale collaboration and pQTL data sharing. SCALLOP is an independent collaborative framework for the discovery and follow-up of genetic associations with proteins for researchers generating data using the Olink platform. The aim of the SCALLOP consortium is to identify novel molecular connections and protein biomarkers that are causal in diseases, and to date, 35 PIs from 28 research institutions have joined the effort, which now comprises summary level data for almost 70,000 patients and controls from 45 cohort studies.

The transcription of genes as mRNA plays an essential intermediate role in converting genetic information to phenotype. Quantitative RNA measurements in the form of transcriptomics have long been used as a relevant proxy for biological output from the genome. Powerful techniques such as RNAseq have enabled large-scale studies of gene expression that have contributed to many systems biology studies. Since many factors other than the rate of gene transcription are involved in protein abundance and function, however, proteomics solution can now provide the key “missing link” between genetic information and real-time biological activity.

An increasing number of studies are combining transcriptomics and Olink proteomics to gain a more complete picture, leveraging on the complementary potential of these approaches, for example to study localised changes at the tissue/cellular level by RNA analysis in relation to systemic changes at the proteomics level.

Case study

A study in collaboration with Massachusetts General Hospital Cancer Center, performed deep proteomic profiling together with single-cell RNAseq in a metastatic melanoma cohort during treatment with checkpoint inhibitor immunotherapy. Around 1500 proteins were measured, resulting in plasma protein profiles associated with both predicted treatment response and overall survival. Single-cell RNAseq data from individual immune cells within the tumors of melanoma patients showed that differentially expressed genes between responders and non-responders were primarily derived from myeloid cells, macrophages and dendritic cells, with an excellent correlation to the plasma proteomics data.

Biological systems typically involve an intricate interplay between different cell types, signaling molecules and other proteins. Approaches that combine cellular and molecular analyses can provide unique insights into complex biological questions, and nowhere is this more important than in studies involving the immune system. Advanced, high throughput technologies such as mass cytometry (e.g. CyTOF) enable detailed analysis of immune cell populations and how they change during health and disease. More recently, mass cytometry has been used in combination with Olink analysis to gain unique insights into immunological questions through this systems-level approach.

Olink and mass cytometry studies

In a landmark study from Dr. Petter Brodin’s group at Karolinska Institute, Stockholm, they used Olink and CyTOF to show for the first time that the immune systems of newborn children evolve in a stereotypic manner that is similar in diverse children, not predictable from cord blood measurements and driven by environmental factors such as the colonizing microbiome.

Such studies have provided vital insights into severe versus mild disease, recovery from severe disease, and in better understanding variants of the disease

This combination of omics also found great utility during the huge research effort into the COVID-19 pandemic, where the complex immune system response to viral infection is a critical factor in the development and severity of the disease. Such studies have provided vital insights into severe versus mild disease, recovery from severe disease, and in better understanding variants of the disease such as Multisystem Inflammatory Syndrome in Children.

🔗 Tebani et al. 2020, Nature Communications

🔗 Magis et al. 2020, Scientific Reports

🔗 Enroth et al. 2014, Nature Communications

🔗 A rare but severe complication in children with COVID-19

🔗 Systems-level immunomonitoring of severe COVID-19