High-throughput Single-cell Proteomics Enabled by Integrating nPOP-workflow with Quantitative Hyperplexing

This study presents a high-throughput single-cell proteomics platform that integrates a modified nPOP workflow with IBT16-TMT16 quantitative hyperplexing to achieve deep proteome coverage (up to 3,000 proteins per cell) and ultrahigh throughput (~2,000 cells/day), successfully revealing distinct molecular features in cholangiocarcinoma cells.

The Gist

Imagine a bustling city. For years, scientists studying this city (our bodies) have only been able to take a "smoothie" of the entire population. They blend millions of cells together, crush them up, and analyze the mixture. While this tells them what the city on average looks like, it completely hides the unique stories of the individual citizens. A baker, a doctor, and a firefighter might all be in the mix, but their specific jobs and struggles get lost in the noise.

Single-cell proteomics is the attempt to stop blending the smoothie and instead look at every single citizen individually to see exactly what they are doing.

This paper introduces a new, super-powered toolkit that makes looking at these individual cells faster, clearer, and more detailed than ever before. Here is the breakdown of their breakthrough:

1. The Problem: The "Blurry Camera" and the "Slow Assembly Line"

Previously, looking at a single cell was like trying to take a high-definition photo of a tiny ant with a blurry camera. You could see something, but you missed the details. Also, the process was incredibly slow. You could only photograph a few ants a day, which isn't enough to study a whole city.

2. The Solution: Two New Super-Tools

The researchers built two different "workflows" (methods) to solve this. Think of them as two different ways to organize a massive library.

Tool A: The "Super-Sensitive Microscope" (Label-Free)

• How it works: They didn't use any chemical tags or labels. Instead, they optimized their equipment (the "camera" and the "lens") to be incredibly sensitive.
• The Analogy: Imagine they upgraded their microscope so it can see the tiny dust motes dancing in a sunbeam without needing to paint the dust.
• The Result: They could look at a single cell and identify over 3,000 different proteins (the workers doing jobs in the cell).
• Real-World Test: They used this on liver cancer cells (Cholangiocarcinoma). They found that cancer cells and healthy cells next to them were doing totally different things. The cancer cells were revving up their "energy engines" and stress-response systems, while the healthy ones were calm. This helps doctors understand exactly how the cancer is behaving.

Tool B: The "High-Speed Barcode Scanner" (Hyperplexing)

• How it works: This is the game-changer for speed. In the old days, you could only scan a few cells at a time. The researchers combined two different types of "barcodes" (chemical tags called IBT and TMT) to create a 32-channel barcode system.
• The Analogy: Imagine a post office. Before, you could only sort 16 letters at a time. Now, they invented a machine that can sort 32 letters simultaneously, and they did it so efficiently that they can process thousands of letters in a single day.
• The "nPOP" Workflow: They used a tiny droplet system (like a digital rain machine) where each cell gets its own microscopic drop of liquid to be processed. This prevents the cells from getting lost or damaged.
• The Result: They could process 2,000 single cells in a single day. They successfully identified 1,400 to 2,000 proteins per cell with 95% accuracy.
• Why it matters: They tested this on four different types of human cells (like a skin cell, a lung cell, etc.). The machine didn't just count them; it understood their unique "jobs." It knew the lung cells were focused on breathing and stress, while the liver cells were focused on moving materials around.

3. Why This Changes Everything

• Speed: They went from taking a photo of one cell an hour to taking photos of 2,000 cells a day.
• Clarity: They can now see the "individual stories" of cells. In cancer, this is huge because not all cancer cells are the same. Some are aggressive, some are hiding. This tool can spot the "bad actors" that other methods miss.
• Scalability: Because it's so fast and accurate, this isn't just for lab experiments anymore. It's ready to be used on thousands of patient samples to help design better, personalized cancer treatments.

The Bottom Line

The authors built a high-speed, high-definition camera system for the microscopic world. They figured out how to take a picture of a single cell, identify its entire "to-do list" (proteins), and do it fast enough to study entire populations of cells. This allows scientists to finally understand the unique personalities of individual cells, which is the key to unlocking better cures for complex diseases like cancer.

Technical Summary

1. Problem Statement

Single-cell proteomics (scProteomics) is essential for understanding cellular heterogeneity and dynamic molecular mechanisms that bulk proteomics obscures. However, the field faces two primary bottlenecks:

• The Throughput vs. Depth Trade-off: Label-free methods offer high sensitivity and deep proteome coverage (identifying >3,000 proteins/cell) but suffer from low throughput and poor quantitative consistency for large cohorts. Conversely, isobaric labeling methods (e.g., TMT) improve throughput but are limited by the finite number of channels (e.g., TMT 16-plex or 18-plex) and often require extensive offline fractionation, which reduces sensitivity for single cells.
• Sample Loss and Cost: Handling picoliter-to-nanoliter volumes in single-cell workflows often leads to significant sample transfer losses, high reagent costs, and batch effects.
• Lack of Integrated Hyperplexing: While "hyperplexing" (combining different tagging chemistries to expand channels) has been demonstrated in bulk proteomics, it has not yet been successfully integrated with high-efficiency single-cell sample preparation workflows like nPOP (nano-proteomic sample preparation).

2. Methodology

The authors developed a dual-strategy workflow to address sensitivity and throughput simultaneously:

A. Optimization of Label-Free scProteomics (High Sensitivity)

To establish a baseline for maximum sensitivity, the authors optimized the liquid chromatography-mass spectrometry (LC-MS) configuration for single-cell inputs:

• Instrumentation: Compared timsTOF SCP vs. timsTOF Pro. The SCP showed superior sensitivity (3.35x more protein groups at 0.2 ng input).
• LC Configuration: Evaluated trap-column vs. trap-free setups. A trap-free, single-column configuration significantly improved proteome identification (up to 81% increase) and peak resolution for low-input samples.
• Column & System: Compared nanoElute vs. Evosep One. The nanoElute system paired with a homemade analytical column (20 cm × 50 μm i.d., 1.9 μm C18 resin) and an integrated spraytip yielded the highest ion intensity and proteome coverage.
• Acquisition Mode: DIA-PASEF (Data-Independent Acquisition) consistently outperformed DDA-PASEF across all sample amounts.
• Sample Prep: Used the cellenONE system to sort single cells into 3 μL master mixes containing DDM, HEPES, and trypsin for direct lysis and digestion.

B. Integration of nPOP with Quantitative Hyperplexing (High Throughput)

To scale up to large cohorts, the authors integrated a modified nPOP workflow with a Quantitative Hyperplexing strategy:

• Sample Preparation: Automated dispensing of single cells into nanoliter-scale droplets on a fluorocarbon-coated glass slide. Each droplet contained 13 nL lysis buffer.
• Hyperplexing Strategy: Combined IBT16 (Isobaric Tags) and TMTpro16 reagents to create a 32-plex workflow.
  • To avoid channel overlap on the timsTOF SCP, the 32 channels were split into two groups (Group A and Group B) per run.
  • Included specific reference, carrier, and blank channels for normalization and background correction.
• Workflow: Cells were lysed and digested in droplets, labeled with IBT or TMT, quenched, and then pooled by group for LC-MS analysis.
• Data Analysis: Used PEAKS software for identification and quantification, employing KNN imputation for missing values and normalization against designated reference channels.

3. Key Results

Label-Free Performance (Sensitivity)

• Proteome Depth: On average, >3,000 protein groups were identified from individual 293T and HeLa cells (using timsTOF SCP + nanoElute).
• Reproducibility: High intra-group correlation (coefficients close to 1) and clear separation in PCA between cell types (293T, HeLa, and bulk controls).
• Clinical Application (CCA): Applied to cholangiocarcinoma (CCA) and matched paracancerous cells. Identified ~2,000 proteins/cell. Revealed distinct metabolic and translational regulation patterns, including upregulation of stress response (HSPD1) and energy metabolism (PDHA1) in tumor cells, and downregulation of drug metabolism enzymes (CYP4F3, CES1).

Hyperplexing Performance (Throughput & Quantification)

• Multiplexing Capacity: Successfully expanded multiplexing to 32 channels (IBT16 + TMTpro16) per run.
• Efficiency: Achieved >95% labeling efficiency across four human cell lines (293T, HeLa, A549, LM3).
• Proteome Depth: Consistently quantified 1,400 to 2,000 protein groups per single cell.
• Dynamic Range: Captured proteins across a five-order-of-magnitude dynamic range.
• Biological Discrimination: PCA and hierarchical clustering clearly separated the four cell lines based on biological identity rather than labeling artifacts. GO enrichment analysis revealed cell-type-specific signatures (e.g., RNA metabolism in 293T, DNA replication in HeLa, metastasis pathways in LM3).

Throughput Estimation

• The workflow supports the preparation of ~1,440 single cells in a single batch using cellenONE.
• When coupled with next-generation instruments (e.g., Orbitrap Astral Zoom or timsUltra AIP), the authors estimate a detection throughput of ~2,000 single cells per day.

4. Key Contributions

1. Workflow Integration: First demonstration of integrating nPOP (nanoliter droplet processing) with Quantitative Hyperplexing (IBT16 + TMTpro16) for single-cell proteomics.
2. Hardware Optimization: Systematic optimization of LC-MS parameters (trap-free, homemade columns, DIA-PASEF) that pushed label-free single-cell sensitivity to >3,000 proteins/cell.
3. Scalability: Developed a strategy that overcomes the channel limit of standard TMT, enabling 32-plex analysis with high quantitative accuracy and minimal sample loss.
4. Clinical Relevance: Validated the workflow on fresh-frozen human cholangiocarcinoma tissue, successfully resolving tumor heterogeneity and metabolic reprogramming at the single-cell level.

5. Significance

This study provides a scalable, efficient, and robust platform for large-scale single-cell proteomics. By solving the trade-off between depth and throughput, the authors enable:

• Large Cohort Studies: The ability to profile thousands of cells (e.g., ~2,000/day) makes clinical studies with significant statistical power feasible.
• Deep Biological Insight: The combination of high sensitivity and high multiplexing allows for the detection of rare cell populations and subtle proteomic shifts that are masked in bulk analyses.
• Future Expansion: The modular nature of the workflow suggests compatibility with even higher multiplexing levels (e.g., TMTpro 32-plex + IBT32-plex) and other labeling chemistries, paving the way for comprehensive proteomic atlases of human tissues and diseases.