Investigating Disease Mechanisms with Single-Cell Sequencing

Single-cell sequencing enables researchers to investigate complex biological systems at the level of individual cells. This approach provides a detailed view of cellular function, communication, and changes in these features across conditions. In a joint webinar hosted by Biocompare and SEQanswers, three experts shared their approaches for using single-cell techniques to study cellular composition and variation in type 2 diabetes and central nervous system tumors.

β-cell dysfunction in type 2 diabetes

Researchers from the Jackson Laboratory used single-cell RNA sequencing to create a detailed map of cell-specific gene expression changes in human pancreatic islets, aiming to better understand type 2 diabetes (T2D). These pancreatic islet cells, which secrete insulin and glucagon to regulate glucose levels, are closely linked to the development and progression of T2D. In this study, presented by Khushdeep Bandesh, the team examined how gene expression and islet cell proportions vary with diabetes status and identified potential contributors to β-cell dysfunction.¹

The study involved a detailed analysis of 245,878 cells from 48 donors, including 17 non-diabetic (ND), 14 pre-diabetic (PD), and 17 with diagnosed T2D. Bandesh explained that unsupervised clustering of the donor transcriptomes identified 14 islet cell types along with signature genes enriched in hormone-related pathways. Not surprisingly, insulin-producing β cells were noticeably reduced in donors with T2D. This corresponded to a 25–30% drop in functional β-cell mass, due to both fewer total β cells and a shift toward senescent subpopulations. However, α-cell proportions increased, likely reflecting compensatory responses. Detailed subclustering of the β cells showed that eight distinct subtypes were present across all donors. In T2D samples, one β-cell subtype tied to insulin secretion declined, while another marked by senescence increased. Together, these changes accounted for a major portion of the β-cell reduction.

After assessing cell proportions, Bandesh performed differential expression analysis and identified 511 genes altered in T2D β cells with 316 upregulated and 195 downregulated. About two-thirds had not been previously linked to T2D. A subsequent pathway analysis revealed higher expression of genes involved in neuronal signaling and lower expression of those tied to hormone regulation and lipid metabolism. Interestingly, β cells in T2D donors showed increased expression of neuroreceptor genes, suggesting a disruption in neuroendocrine control. Among the downregulated genes were those involved in vitamin A metabolism, which has been linked to β-cell survival. Functional studies in EndoC-βH3 cells supported roles for several of these genes in insulin secretion and cell viability.

To strengthen their findings, the team cross-referenced genetic data from large-scale genomic studies, linking known T2D risk variants to specific gene expression changes within β cells. From the 19,900 T2D-linked variants, 461 were associated with expression changes in 41 β-cell DEGs. Additionally, another set of 27 genes had matching effects from both the genetic variant and disease, highlighting them as potential targets for further investigation.

The list was further refined through proteomic data and mouse knockout phenotypes. Of the 106 DEGs with available protein data, several also showed consistent changes in T2D donors. Mouse knockout data revealed that 13 β-cell DEGs led to glycemic defects, further supporting a potential role in glucose regulation. In total, 92 β-cell DEGs stood out as high-priority targets based on their signals across transcriptomic, proteomic, genetic, and functional datasets. Bandesh noted that the results are accessible through multiple open platforms, including an R Shiny app (TAPIC), with raw data available through GEO, Zenodo, and other repositories.

In summary, the study identified a 25–30% reduction in functional β-cell mass in T2D, driven by fewer total β cells and a shift toward senescent subpopulations. Around 66.5% of the differentially expressed genes were previously unreported and linked to altered neurotransmission. Lastly, Bandesh concluded that using multiomics, genetic, and experimental data, the team nominated 92 β cell differentially expressed genes as potential targets for further investigation.

Immune profiling of CSF reveals disease-specific signatures

In the next two presentations, Paula Nieto and Juan Nieto, from the Single-Cell Genomics Team at the National Center for Genomic Analysis, presented different aspects of their high-resolution analysis of cerebrospinal fluid (CSF) from patients with central nervous system (CNS) tumors and leptomeningeal disease (LMD). Using a multiomic approach, they were able to characterize the immune dynamics in CNS lymphoma, glioblastoma (GB), and brain metastases (BrM).²

The study began by collecting CSF, blood samples, and brain biopsies. CSF proved especially useful as it surrounds the brain and spine and can serve as a diagnostic tool for tracking CNS tumors. After collection, the research team performed single-cell RNA and TCR sequencing, spatial transcriptomics, and nanopore sequencing of cell-free DNA.

The initial analysis showed that CSF samples carried disease-specific immune patterns depending on the type of tumor or inflammation. CNS lymphoma samples were enriched in cytotoxic, pre-exhausted, and exhausted T cells, suggesting an anti-tumor immune response. Conversely, BrM and GB samples largely comprised myeloid populations, including blood-derived monocytes and CNS-resident macrophages with anti-inflammatory or metabolically shifted phenotypes. Along with these results, patients with neuroinflammatory diseases showed a more dendritic-cell-rich environment, while healthy controls had predominantly central memory CD4+ T cells.

Next, the team tracked clonal T cell expansions across blood and CSF using TCR sequencing data. Most of the expanded clones found in the CSF were also detected in peripheral blood, which indicates peripheral activation and migration into the CNS. However, the higher expansion and reduced TCR diversity in the CSF suggested local proliferation in response to CNS antigens.

The research team then used longitudinal sampling to examine how CSF composition shifts during treatment and disease progression. In one patient with primary CNS lymphoma, single-cell analysis detected expanding exhausted CD8+ clones, even as the patient received high-dose steroids. Another case involving breast cancer with BrM and LMD showed clonal evolution of tumor cells in the CSF, including a drug-resistant clone that emerged during therapy and displayed angiogenic and differentiated gene signatures.

Further single-cell profiling of myeloid populations revealed distinct patterns tied to tumor type. The CNS lymphoma and inflammatory cases showed higher infiltration of blood-derived cells, while GB and BrM samples contained more resident-like macrophages. A gene module analysis suggested metabolic rewiring of myeloid cells. Most notably, lactate-associated modules correlated with CSF lactate levels measured in clinical diagnostics, indicating a possible role in immune suppression or tumor-promoting activity.

The integration of spatial transcriptomics to brain tissue confirmed that CSF immune profiles closely mirror the cellular environments of corresponding CNS lesions. This overlap was most evident in T-cell populations, which implies active movement between tissue and CSF, while myeloid cells showed more localized functional shifts.

Altogether, these findings suggest that CSF immune signatures may reflect the tumor environment in less accessible areas of the CNS. This work also offers a framework for integrating CSF liquid biopsy with spatial and single-cell technologies to monitor immune responses and tumor progression. While the clinical utility of this approach requires further validation, the dataset provides a foundation for future investigations into CNS disease biology.

References

1. Bandesh K, Motakis E, Nargund S, et al. Single-cell decoding of human islet cell type-specific alterations in type 2 diabetes reveals converging genetic- and state-driven β-cell gene expression defects. bioRxiv. Published online January 22, 2025:2025.01.17.633590. 

2. Nieto P, Klinsing S, Caratù G, et al. Decoding the immune response in leptomeningeal disease through single-cell sequencing of cerebrospinal fluid. bioRxiv. Published online January 27, 2025:2025.01.27.634744.