Genentech Builds Single Cell Proteomics Platform for Real Time Mass Spec Analysis
By Allison Proffitt
March 3, 2022 | Our technical capabilities are becoming increasingly granular, and proteomics is no exception. At the recent Molecular & Precision Med TRI-CON, Christopher Rose reported on how his group within the Microchemistry, Proteomics & Lipidomics department at Genentech where he currently leads the Discovery Proteomics group is working to build a single-cell proteomics platform.
The Discovery Proteomics group at Genentech uses tools—generally mass spectrometry—to interrogate potential targets, aid in compound selection, do some toxicology, and identify target protein or peptide sequences and biomarker selection. Rose said. Mass spec has been the historical tool of choice, though he highlighted changes as sample-sizes continue to decrease.
When his team surveyed the field of pharma proteomics researchers (DOI: 10.1080/14789450.2021.1962300), Rose said, single cell proteomics was a dominant area of interest. It’s particular attractive, Rose said, because as opposed to antibody-based methods, single-molecule sequencing and mass spec allow for untargeted analysis. “These are attractive because we don’t have to define the target beforehand. We don’t have to have an antibody reagent. But we can just go into a cellular system and then profile the proteome in an unrestricted manner and be able to find something new.”
It’s a promising area of work, Rose said, with many recent papers outlining opportunities including identification of membrane proteins, untargeted spatial proteome analysis of tissues, proteome analysis of cells microdissected from tissues, and multiomics, perhaps combining with single cell RNAseq to learn where proteomics can give the most unique and impactful information and how to use the technologies together.
And proteomics can deliver a wealth of data, especially with large sample sizes. “But as you go down toward a single cell, it becomes much more challenging to identify a large number of proteins. Part of that is due to the sensitivity of our measurements; part of it is due to how the samples are prepared and the losses you might incur along the way,” Rose said. Over the past five years, he highlighted, proteomics sample sizes have shrunk to smaller and smaller numbers, now hundreds of thousands of cells at a time. “The biology is pushing us toward more specific cell populations, and trying to understand how particular cell populations respond to a particular therapeutic,” he said, highlighting that single cell proteomics offers the opportunity to understand cell heterogeneity and study rare cell populations. But there are still significant challenges to single cell proteomics, with a limited depth available (1,000 to 1,500 he surmised) and analysis of cellular signaling events still difficult.
A New Analysis Pipeline
Since about 2018, the prevalent method of single cell proteomics was SCoPE-MS, which mixed up to eight barcoded single cells at a time with carrier cells. The carrier proteome gives enough signal to detect the single-cell proteomes of interest, which can be fragmented to get peptide identification and the quantitative information in the barcodes can be revealed. But the levels of individual protein within the cell are hard to measure against the very strong signal from the carrier cells, and variation in measurement increases as you add more carrier proteome. A workaround, Rose said, has been to sample more peptides. The instrument takes longer per peptide, but it does give more accurate information.
Rose and his colleagues studied the carrier proteome limit for single cell proteomics and examined the accuracy of the SCoPE-MS approach in a paper published at the end of 2020 (DOI: 10.1038/s41592-020-01002-5). They showed that early single cell proteomics data was very noisy, and while many proteins were identified, that data quality was still quite poor.
The team came up with a method to separate the carrier proteome from the single cell proteomes of interest within the mass space. This allows researchers to sequence the proteomes of interest separately and focus all of the instrument’s attention on the single cell proteomes, not the carrier proteome.
That data analysis originally required collecting the raw data from the mass spec and sending it into the data analysis pipeline. Matching of the spectra outputs to peptide sequences happened offline servers and database searches. Users could view the quantitative data via web interface, and analyzed and shared the data via Spotfire dashboards.
This was bulky and time consuming, so the team looked for a way to take a piece of spectra data off the mass spec and match it to a peptide in real time. “The beauty of this is once we’ve identified a peptide sequence, we could understand or tell the instrument to do something new. We have the ability to send instructions back to the instrument to tell it to perform a different type of analysis,” Rose explained.
His lab’s version of this workflow—inSeqAPI—mimics the workflow users would see on the instrument and it allows researchers to use fast scans to match spectra to peptides, and save slow, quantification scans for peptides of interest. “By doing that, we can dedicate more of the instrument’s time to these slower, quantification scans,” he said. “This has been a real help in single-cell proteomics, especially when the sample is somewhat limited.”
The inSeqAPI pipeline is very fast, doing real time search matching peptide sequences to the spectra happens in about 5 milliseconds, and yields an increase in signal that is “pretty dramatic,” Rose reported. The strong signal remains even as carrier protein concentrations increase, making it a robust approach as well.
Rose also mentioned improved sample prep with CellenONE and the proteoCHIP system, and outlined next steps for his team’s work. He hopes to apply these techniques to understand single cell response to therapeutics and use those learnings to inform therapeutic development. He also hopes to work to improve proteomics depth further for single cell approaches.