Discovering circulating protein biomarkers through in-depth plasma proteomics

Abstract

Plasma, i.e., the liquid component of blood, is one of the most clinically used samples for biomarker measurement. Despite that plasma proteins and metabolites are the most frequently analysed biomarkers in practice, identifying and implementing new circulating protein biomarkers for diagnosis, treatment prediction, prognosis, and disease monitoring has been limited.

This PhD thesis compiles the discovery of systemic alterations in the blood plasma proteome and potential biomarkers related to disease status, prognosis, or treatment through plasma proteomics. We analysed plasma and serum samples with global proteomics by high-resolution isoelectric focusing (HiRIEF) and liquid chromatography coupled with mass-spectrometry (LC-MS/MS), and targeted proteomics by antibody-based proximity extension assays (PEA) in three diseases that would benefit from blood biomarkers: stage IV metastatic cutaneous melanoma (mCM), glioblastoma (GBM), and coronavirus disease 2019 (COVID-19). Specifically:

a.) New treatment options for mCM substantially prolong overall survival (OS), but multiple patients do not respond to treatment or develop treatment resistance, thus having shorter progression free survival (PFS). Corroborated by the presence of multiple metastases, which makes biomarker sampling difficult, circulating proteins derived from the tumour and in response to treatment could serve as predictive and prognostic biomarkers in mCM.

b.) GBM is the most malignant primary brain tumour with limited treatment options and notoriously short OS. Sampling biomarkers for GBM requires an invasive surgical intervention on the skull, which makes GBM a good candidate for circulating protein biomarkers for prognosis and monitoring.

c.) COVID-19 is an inflammation-driven infectious disease that affects multiple organs and systems, thus making the plasma proteome a good source to explore systemic biological processes occurring in COVID-19.

In papers I and II, using HiRIEF LC-MS/MS and PEA, we explored the treatment-driven plasma proteome alterations in mCM patients treated with anti-PD-1 immune checkpoint inhibitors (ICI) and MAPK-inhibitors (MAPKi), respectively, and identified potential treatment predictive and monitoring biomarkers. mCM patients treated with anti-PD-1 ICI had a strong increase in soluble PD-1 levels during treatment, and upregulation of proteins involved in T-cell response. BRAF[V600]-mutated mCM patients treated with MAPKi had deregulation in proteins involved in immune response and proteolysis. CPB1 had the highest increase in patients treated with BRAF- and MEK-inhibitors and was associated with longer PFS. Higher levels of several proteins involved in inflammation before treatment were associated with shorter PFS regardless of ICI or MAPKi treatment.

In paper III, using HiRIEF LC-MS/MS and PEA, we longitudinally analysed the plasma proteome dynamics of GBM patients, collecting plasma samples before surgery and at three timepoints after surgery. Through consensus clustering, based on treatment-naïve plasma protein levels, we identified two patient clusters that differed in median OS. The association between the cluster membership and OS remained consistent after adjustment for age, sex, and treatment. Through machine learning, we identified protein panels that separated the patient clusters and may serve as prognostic biomarkers. The largest alterations in the plasma proteome of GBM patients occurred within two months after surgery, whereas the plasma protein levels at later timepoints had no difference compared to pre- surgery levels. We observed a decrease in glioma-elevated proteins in the blood after surgery, identifying potential monitoring biomarkers.

In paper IV, using HiRIEF LC-MS/MS, we analysed serum proteome alterations in hospitalised COVID-19 patients in comparison to healthy controls, and identified a strong upregulation in inflammatory, interferon-induced, and proteasomal proteins. Several protein groups showed association with clinical parameters of COVID-19 severity, including proteasomal proteins. Serum proteome alterations were traceable to proteome alterations induced in a lung adenocarcinoma cell line (Calu-3) by infection with SARS-CoV-2. Finally, we performed the first meta-analysis of global proteomics studies of the soluble blood proteome in COVID-19, providing estimates of standardised mean differences and summary receiver operating characteristics curves. We demonstrate the high accuracy and precision of HiRIEF LC-MS/MS when compared to the meta-analysis estimates and pinpoint proteins that may serve as biomarkers of COVID-19.

In summary, this thesis postulates that new circulating protein biomarkers would be clinically useful. By combining mass-spectrometry- and antibody-based-proteomics, we demonstrate the potential of in-depth analyses of the plasma proteome in capturing systemic alterations related to treatment, survival, and disease status, pinpointing potentially novel biomarkers that require validation in larger cohorts.