To gain new insights about the long-term impact of SARS-CoV-2 infection, researchers in Sweden used a Capitainer self-sampling device to gather dried blood spot samples from random, consenting adults in 2020 and 2021. Samples were collected by individuals in their own homes and analysed using commercial proteomics assays from Luminex and Olink. The study detected clinically-known markers for SARS-CoV-2 infection that align with described infection stages. It also highlights the power of a combined Capitainer and Olink workflow to analyse hundreds of proteins simultaneously and gain new insights into human health.
Four years after the onset of the COVID-19 pandemic, many questions remain about the impact of SARS-CoV-2 infection, particularly in individuals who experienced mild or no symptoms and did not seek medical attention. COVID-19 was initially believed to be a respiratory disease but it has since been established that other major organ systems may be impacted and that hyperactivation of the immune response (which leads to cytokine storm) is a key step in its pathogenesis.
Population studies to understand the heterogeneous phenotypes reported in infected individuals as well as environmental and genetic risk factors for serious disease and mortality are hampered by inclusion bias since certain cohorts, e.g., undiagnosed or non-hospitalised individuals, are easily left out when sampling occurs via the healthcare system. To address that challenge, researchers at KTH and SciLifeLab in Sweden explored the potential of Capitainer quantitative dried blood spot (qDBS) self-sampling to analyse the proteomes of a general population that included undiagnosed and non-hospitalised individuals. Their findings were recently published in Communications Medicine (1).
Study design: sampling a broad patient population with the aid of self-reporting questionnaires
To initiate the study, Capitainer’s self-sampling qDBS cards (Capitainer®B) and questionnaires for self-reporting of COVID-19 symptoms were mailed to 4,000 random, consenting adults in metropolitan Stockholm and Gothenburg during 2020 and 2021.
qDBS samples were self-collected by finger-pricking and posted back to the clinical laboratory for analysis. Proteins were then eluted from pre-cut filter discs within Capitainer®B, and COVID-19 serostatus was determined by assaying for antibodies against the SARS-CoV-2 nucleocapsid and spike proteins using MAGPLEX multi-analyte serology technology from Luminex Corp.
By measuring the titres of SARS-CoV-2-specific IgM (marker of early/acute infection) and IgG (marker of late/past infection) antibodies, the samples were stratified as representing early- or late phase SARS-CoV-2 infection. On a subset of 228 samples, 276 proteins involved in cardiovascular disease and metabolism in qDBS samples from three study groups:
Study Group 1: Individuals from random households in Sweden who performed DBS self-sampling during the first pandemic wave in 2020. These samples were either deemed seropositive and represented the peak of the immune response to SARS-CoV-2 (n = 44, IgM+/IgG+) or seronegative (n = 37, IgM-/IgG-). The seropositive subgroup was only exposed to the wild-type SARS-CoV-2 variant and there were no significant differences in self-reported symptoms within the entire group.
Study Group 2: Individuals who had donated DBS samples in 2020, which were stratified as being in the early (n = 26) or late (n = 40) phase of confirmed SARS-CoV-2 infection. Stratification was based on the titres of anti-spike antibodies where acute infection was defined as having an IgM+/IgG+ serostatus while late-stage infection was marked by an IgM-/IgG+ serostatus. This entire group was only exposed to the wild-type SARS-CoV-2 variant and there were no significant differences in self-reported symptoms within the group.
Study Group 3: Unvaccinated seropositive (n = 37, IgG+) and seronegative (n = 40, IgG-) individuals who donated DBS samples more than one year into the pandemic (i.e., during the third wave in 2021). This group was stratified based on the presence of IgG antibodies against the nucleocapsid and spike proteins of SARS-CoV-2. Self-reported symptoms varied in this group, with approx. 30 % of asymptomatic seropositive donors. During the time of sampling in this group, the Alpha variant (B.1.1.7) was dominant and the prevalence of the Delta variant (B.1.617.2) was increasing. This group was therefore likely exposed to numerous SARS-CoV-2 variants.
A combined Capitainer and Olink proteomics workflow to detect hundreds of target proteins
The samples in all study groups were assayed for 276 target proteins using Olink’s proximity extension assay (PEA) technology. This study specifically used Olink’s Cardiovascular III, Metabolism, and Cardiometabolic antibody panels. Each panel includes 92 pairs of oligonucleotide-labelled antibodies, and when incubated with sample eluates, every pair of antibodies binds a target protein in a pair-wise manner. When a pair of antibodies binds the same target protein, the unique oligonucleotides on those antibodies are brought into proximity, where they hybridise to form a DNA reporter template that can be amplified, identified and quantified using quantitative real-time PCR. This provides semi-quantitative data about the abundance of target proteins in all samples.
The team confirmed the compatibility of Capitainer® samples for deep multiplex proteomics by carrying out side-by-side comparisons of 92 proteins related to cardiovascular diseases in venous blood samples and DBS samples collected at the same visit from the same 12 donors (= 24 samples). Despite differences in the protein expression levels between venous and qDBS samples as discussed in the article (1), 91 of the 92 assayed proteins were detected in more than 90 % of the 24 samples and these proteins could be measured in venous and DBS samples. These findings highlight the suitability of a combined Capitainer and Olink workflow for multiplex protein analysis.
Results – high target protein detection rate in Capitainer® and venous samples
An analysis of all samples revealed that 260 of the 276 target proteins could be detected in more than 90 % of the samples from all study groups. A replicate analysis was then carried for 264 of the 276 proteins in five unique qDBS eluates. These 264 proteins were included because they were quantified to be above the detection limit in at least half of the samples assayed, and the replicate analysis revealed high reproducibility, with more than 90 % of the proteins reporting a coefficient of variation of less than 10 %.
After statistical analysis to remove potential outliers, proteomics data from 228 of the samples were further studied. Overall, the findings revealed proteins that are relevant to COVID-19 pathogenesis and the documented immune response against SARS-CoV-2.
Detection of known markers of mortality and prolonged COVID-19 disease
The data obtained from Study Group 1 revealed some physiological aspects of COVID-19 pathology. For example, several proteins in the seropositive subgroup were previously associated with intensive care unit mortality (e.g., MBL2) or prolonged illness (e.g., IL2RA) in severe COVID-19 patients. Overall, the protein expression signatures represented cell-mediated immune responses and tissue damage, which are known pathogenesis mechanisms in severe COVID-19 disease.
According to self-reporting in the supplied questionnaires, individuals became infected on average five months (with a range of 0-12 months) before sampling took place for this study. This variation in interval between infection and sampling, and other variable factors including the use of medication to relieve symptoms, may have contributed to the heterogeneity of donors within each group. In addition, because older individuals were the first to be vaccinated (when Group 3 was sampled), the individuals included in Group 3 were generally younger than those in Groups 1 and 2.
To investigate changes in protein levels between time of infection and time of sampling, and to test the advantage of repetitive blood testing that is made possible with qDBS sampling, a longitudinal pilot study was carried out using five qDBS samples collected during the recovery period by an individual confirmed to be COVID-19-positive via a standard PCR test. Indeed, this experiment revealed changes in protein levels between the time of infection and time of sampling, mirroring the results obtained in the qDBS-based population surveys.
Capitainer®B and Olink are compatible for large-scale proteomics workflows
While longitudinal differences may have impacted the results obtained from the various patient groups in this study, this is the first report of Capitainer® self-sampling for proteomics in a general population study. Importantly, the study demonstrates that the combined Capitainer and Olink workflow is a suitable for home sampling for future large-scale population health studies.
References
- Fredolini C, Dodig-Crnković T, Bendes A, et al. Proteome profiling of home-sampled dried blood spots reveals proteins of SARS-CoV-2 infections. Commun Med (Lond). 2024 Apr 2;4(1):55.