Core Framework & Diagram How Did Single-Cell Technology Transform Immune Research?
7 月 24, 20261 Min Read Major Controversies in Immunology History
7 月 24, 2026Before, we were looking at a forest. Now we can see each tree clearly
—— How single-cell sequencing technology remapped the immune system.
I. mBefore single-cell sequencing: feeling the elephant in the dark
Before single-cell technology, immunology's research paradigm relied on 'population averages' — mixing millions of similar cells for analysis, obtaining the entire population's average gene expression, average protein levels, and average functional characteristics. This approach was effective in many situations, but has a fundamental blind spot: heterogeneity within cell populations was smoothed over. Imagine: you want to study whether 'activated T cells' express a specific gene. You mix a million activated T cells and measure the gene's expression, getting a number — say '100.' But this '100' might come from: a million cells each expressing 1 unit, or ten thousand cells each expressing 100 units while the other 990,000 cells express nothing at all. These two situations have completely different biological meanings, but under traditional methods you get the same number.
This blind spot led to large amounts of incorrect or incomplete understanding. For example, regulatory T cells (Treg) were long viewed as a functionally uniform cell type, but single-cell sequencing shows Treg is actually a highly heterogeneous cell population containing multiple subgroups with completely different functions — some mainly active in lymph nodes, some specifically existing in the tumor microenvironment, some playing pro-inflammatory roles in chronic inflammation. These subgroup discoveries completely changed immune regulatory strategies targeting Treg.
2. scRNA-seq technology: taking a 'full-body photo' of each cell
The core scRNA-seq technical process can be divided into three steps: first, cell separation — separating individual cells from tissue or blood samples one by one into single suspended cells; second, single-cell capture and library construction — usually using microfluidic droplet technology (like 10x Genomics' Chromium system), encapsulating each cell in an independent oil droplet with uniquely barcoded beads, reverse transcribing the cell's internal mRNA into cDNA with the cell's unique identifier; third, large-scale parallel sequencing — simultaneously determining cDNA sequences of thousands to tens of thousands of cells, then using barcodes to assign each read back to the corresponding individual cell. The final output is a massive 'cell × gene' matrix: each row is a cell, each column is a gene, and the number in each box represents this cell's expression intensity for that gene. This matrix typically contains 5,000 to 50,000 cells and over 20,000 gene positions — enormously informative, requiring specialized computational biology methods (dimensionality reduction, clustering, trajectory analysis, etc.) to extract meaningful structure. In 2019, CITE-seq appeared as an important upgrade: while measuring each cell's RNA expression profile, simultaneously measuring surface protein expression using DNA-barcoded antibodies — obtaining both transcriptome and proteome information for each cell in the same experiment, greatly improving cell identity definition precision.
3. Single-cell technology's revolution in tumor immunology
The tumor microenvironment (TME) is one of single-cell technology's most important application areas. Before single-cell technology, researchers knew tumors contained 'tumor-infiltrating lymphocytes (TIL)' and TIL density correlated with patient prognosis — higher density usually meant better prognosis. But TIL's internal heterogeneity, and different subpopulations' relationships with tumors and treatment response, was almost completely a black box. Single-cell sequencing opened this black box. In 2018, Zhang Zemin's team published the first systematic single-cell analysis of hepatocellular carcinoma tumor microenvironment, mapping the complete distribution of T cells, NK cells, macrophages, and dendritic cells within tumors, discovering multiple previously unknown cell subpopulations — including one called 'progenitor exhausted T cells,' with limited self-renewal capacity, serving as the key to maintaining T cell pools in tumors, and the primary effector cells of PD-1 immune checkpoint treatment responses.
This discovery has direct clinical significance: the proportion of 'progenitor exhausted T cells' in tumors can predict PD-1 treatment response likelihood — patients with higher proportions, after releasing the immune brake, have these progenitor cells that can proliferate and differentiate into effector T cells, producing durable anti-tumor responses; patients with lower proportions, even after releasing the brake, don't have enough T cells to execute the killing task. Spatial transcriptomics further extended this map to the tissue dimension: not only knowing which cells are in the tumor, but also knowing their specific locations in tissue and signal transmission relationships between neighboring cells. Researchers discovered immune cells in different tumor areas (core vs. margin) have completely different activation states, and this spatial heterogeneity closely relates to patient prognosis and treatment response.
4. Human Cell Atlas: immunology's 'periodic table'
Single-cell technology's grandest application is the ongoing 'Human Cell Atlas' (HCA) project — a large-scale international collaboration involving hundreds of laboratories globally, aiming to use single-cell sequencing technology to map complete reference profiles of all cell types in the human body. In the immune system domain, HCA has already produced numerous breakthrough discoveries. In 2022, HCA's immune section published a panoramic description of human fetal immune system development, revealing immune cells' complete differentiation trajectory from early hematopoietic stem cells to various mature immune cell subtypes, as well as fetal-specific immune cell types (that disappear after birth but may reappear under certain developmental conditions). This 'periodic table's' establishment is not only academically meaningful. It provides a reference standard: when researchers observe a certain immune cell in a diseased state (autoimmune disease, infection, cancer), they can compare it with the normal reference map, quickly judging which cells are abnormally activated, which are absent, and which are in abnormal differentiation states.
5. Limitations and what comes next
Despite single-cell technology's power, it has clear limitations. First, the 'snapshot' problem: scRNA-seq provides a cell snapshot at a specific time point, while the immune system is highly dynamic. Is a cell labeled 'exhausted' in early or terminal exhaustion? Will it further exhaust or be reactivated? Static sequencing data can't directly answer these questions. Second, information loss: during cell separation, positional information in tissue, direct contact information between cells, and extracellular signals are all lost. Spatial transcriptomics can partially compensate for position information loss, but current spatial resolution and gene detection sensitivity are still lower than traditional scRNA-seq. Third, cost and data analysis barriers: a typical scRNA-seq experiment's data analysis usually requires specialized bioinformatics teams, analysis periods ranging from weeks to months. Next-generation technologies are attempting to break through these limitations: multi-omics single-cell technology (simultaneously measuring genome, transcriptome, proteome, epigenome), spatial multi-omics (simultaneously measuring multi-level information in tissue), and single-cell temporal tracking (lineage tracing) — linking static snapshots into dynamic stories.
Frequently Asked Questions