Working with challenging samples for additional layers of biological insights
Working with biological samples limited in volume and/or accessibility presents substantial challenges in achieving reliable data and actionable biological insights in protein biomarker research. This obstacle is present in a wide range of research domains and includes volume-limited samples, such as tumour biopsies, ocular matrices, or pediatric specimen, as well as samples with challenging workflows, such as extracellular vesicles, cerebrospinal and interstitial fluid.
However, these sample types often carry important biological insights into human health and disease by closely reflecting physiological changes and sourcing protein biomarkers associated to disease initiation, progression, and response to treatment. The ability to perform high-throughput protein analysis from minute volumes of such precious sample types can therefore make a critical difference in:
- understanding pathological mechanisms,
- improving patient stratification,
- identifying therapeutic targets,
- accelerating drug development.
Understanding human disease through studying alternative matrices
A clear example showcasing the value of using challenging samples for deeper clinical insights is represented by a study comparing plasma protein levels with aqueous (AH) and vitreous humor (VH) in patients with different retinal pathologies. This study revealed that the correlation between the ocular matrices and serum was weak and limited to a few proteins. However, a significant portion of the proteins analyzed (∼41% of detectable proteins) was shown to be correlated between AH and VH, including proteins relevant to retinal pathology, such as angiogenesis regulators, immune-related proteins, and matrix metalloproteinases (Figure 1). Overall, these findings show that AH, which is associated with significantly less vision threatening complications upon sampling compared to VH, could serve as a representative source of biomarker measurements for retinal disease research and monitoring (1).
Cytokine profiling of AH also enabled the identification of three prognostic tumor clusters in uveal melanoma, which differ in patient age and disease stage, pointing to the role of these cytokines in the lack of effective antitumor immune responses (2).
Figure 1. Scatterplots of the protein quantities (NPX) in AH and VH relevant to ocular biology. Blue lines show linear fit. Spearman correlation R2 values are indicated.
While cerebrospinal fluid (CSF) is a sample type with a challenging and invasive collection method, it offers crucial insights into neurological disease development and progression. In a study performing proteomic profiling of CSF samples sourced from patients with Parkinson’s disease (PD), corticobasal syndrome, progressive supranuclear palsy, multiple system atrophy and controls aimed to identify novel diagnostic biomarkers for PD using discovery and validation cohorts. Six proteins were significantly different in PD compared to controls in both cohorts (including midkine, DDC, SMAD5, CCL17, TFPI-2 and TF), serving as potential novel diagnostic biomarkers for PD (Figure 2) (3).
Figure 2. Association of CSF protein expression with PD clinical parameters. Spearman’s rank, adjusted for age and sex, between CSF levels of significantly differentially expressed proteins in PD Stockholm cohort patients and disease duration, scores of Hoehn&Yahr, Unifed Parkinson’s Disease Rating Scale (UPDRS) part 3, Montreal Cognitive Assessment (MoCA), Hospital Anxiety and Depression Scale (HADS), Non-Motor Symptoms Questionnaire (NMSQ) and Pittsburgh Sleep Quality Index (PSQI), and Levodopa equivalent daily dose (LEDD) (BH adjusted P-values; *P<0.05; **P<0.01).
Gaining more from limited blood samples
The ability to extract protein insights from limited sample volumes is especially notable in studies involving pediatric and neonatal blood. The use of human specimen involves stringent regulations when working with pediatric cohorts, often limiting the acceptable sample volume, collection method and analysis. However, these samples have proven essential for identifying novel protein biomarkers for disease prediction and diagnosis in pediatric research.
For instance, a study by Dong et al. examined inflammatory proteins involved in necrotizing enterocolitis (NEC), a serious gastrointestinal disease of the newborn causing high mortality in premature infants (4). Proteomics analysis of plasma samples from newborn infants with NEC or sepsis, as well as healthy controls revealed 11 inflammatory proteins (IL-8, IL-24, TSLP, LIF, OPG, TRAIL, TNFSF14, CCL20, MCP-4, CXCL1, MMP-10) that could distinguish NEC from controls (Figure 3). Furthermore, the combinations of these markers showed much better diagnostic value than any of the individual proteins and commonly used infection marker CRP. These results therefore offer a new strategy for early detection NEC.
Figure 3. Principal component analysis based on protein levels observed in NEC, sepsis, and HC groups. Wayne diagram highlighting the common and unique proteins in NEC and sepsis groups.
Limited blood samples pose an issue beyond pediatrics research. An example of this is intracranial serum, available in minute volumes upon collection, however, found to be an important source of insights into ischemic stroke diagnostics and prognostics. A study by Maglinger et al. revealed systemic and intracranial VCAM1 associations to stroke comorbidities, stroke severity, functional outcomes, as well as the role of VCAM1 in molecular signaling pathways (5). These findings may help predict stroke severity, improve stroke diagnostics and prognostics, while also providing practical therapeutic targets for drug development and drug repurposing.
The Proximity Extension Assay as a solution to limited sample volume
The examples of challenging and precious sample usage outlined above were enabled by Olink’s proximity extension assay (PEA), a robust platform for multiplex protein analysis requiring just 1 μL of sample. Due to its in-built quality control system that monitors each step of the PEA protocol, only single measurements are required to ensure reliable, high-quality data. By circumventing the need for replicates, PEA ensures accurate protein measurements while saving precious sample volume. Furthermore, PEA has been widely referenced in scientific literature to work with a wide array of sample matrices across major therapeutic areas.
Read more about how PEA is being used to reveal previously inaccessible protein insights in the White Paper Discovering more with less, by utilizing samples such as:
- Extracellular vesicles
- Interstitial fluid
- Cerebrospinal fluid
- Tears, aqueous and vitreous fluid
- Mouse serum, plasma and tissue lysate, and more
References
- Wilson S, Siebourg-Polster J, Titz B, et al. Correlation of aqueous, vitreous, and serum protein levels in patients with retinal diseases. (2023) Transl Vis Sci Technol. DOI: 10.1167/tvst.12.11.9
- Wierenga A, Cao J, Mouthaan H, et al. Aqueous Humor Biomarkers Identify Three Prognostic Groups in Uveal Melanoma. (2019) Invest Ophthalmol Vis Sci. DOI: 10.1167/iovs.19-28309
- Paslawski, W., Khosousi, S., Hertz, E. et al. Large-scale proximity extension assay reveals CSF midkine and DOPA decarboxylase as supportive diagnostic biomarkers for Parkinson’s disease. (2023) Transl Neurodegener. DOI: 10.1186/s40035-023-00374-w
- Dong, H., Zhang, L., Li, B. et al. Screening inflammatory protein biomarkers on premature infants with necrotizing enterocolitis. (2023) Inflamm Res. DOI: 10.1007/s00011-023-01702-6
- Maglinger B, Sands M, Frank JA, et al. Intracranial VCAM1 at time of mechanical thrombectomy predicts ischemic stroke severity. (2021) J Neuroinflammation. DOI: 10.1186/s12974-021-02157-4