Targeted Proteomics with Single-cell Resolution
Amy E. Herr, Department of Bioengineering, University of California, Berkeley
Innovation in measurement science is needed to transform pathology tools from subjective, qualitative visual inspection of biopsied tissues for just a handful of protein markers to objective, quantitative measurement of surface and intracellular signaling proteins, including (importantly) protein isoforms and sub-cellular location. In this talk, I will focus on my laboratory’s recent efforts to measure multiple protein biomarkers to enhance tumor classification. Our overarching goal is to overcome limitations of workhorse immunohistochemistry (IHC) tools which rely on high-specificity antibody probes which are often not available. This presents an important analytical gap, as single-cell protein assays form the basis for cancer sub-type classification and treatment decisions.
Protein tools optimized for minute tissues and even single-cells are needed for three reasons. First, comparative analyses of intra-tumor regions (e.g., leading vs. interior regions <500 microns in diameter) elucidates resistance progression, which develops differently in different tumor regions. Conventional western blotting (WB) is not sensitive enough. Second, cell-to-cell variation is a hallmark of cancer; measurement of oncoprotein isoform heterogeneity is, in many cases, currently impossible. Third, as with all biobanks, limited tumor specimens are available. This is particularly true with HER2+ BCa, which is often biopsied early when tumors are small. Through microfluidic design strategies, we address both gaps by: single-cell resolution WB detection limits are 0.1% of conventional WB and well-matched to tissue region analysis. To achieve the lower limits of detection, we use an innovative strategy that differs from most microfluidic electrophoresis separations. We do not use a “cross-t” injector that injects <1% of the sample into the separation channel. Our distinct approach borrows from classic electrophoresis formats that pre-concentrate and then inject all material from the sample well into the separation lane. Minimizing dilution allows the scWB to detect protein from ~1 cell. Essentially we employ the same physics as flow cytometry which routinely and successfully measures single-cell protein levels. As with flow cytometry, the scWB limits dilution of cell contents; the protein concentrations in single mammalian cells can be quite high (i.e., nM to uM in ~1pL cell). Like flow cytometry, the scWB has a limit of detection of ~104 molecules.
Our single-cell WB platform is designed for both simplicity and performance. The assay performs ~103 concurrent single-cell Western blots in ~4 hours. The scWB device is a microscope slide coated with a thin layer (30 μm) of photoactive polyacrylamide gel. Using this simple device, the scWB assay steps include: (1) settling of single cells into microwells stippled in the gel layer (30 um diameter), (2) in-situ cell lysis, (3) polyacrylamide gel electrophoresis (PAGE), (4) photo-initiated blotting to immobilize proteins, and (5) antibody probing with scWB fluorescence readout using a microarray fluorescence scanner. PAGE is performed in ultra-short 500 um to 2 mm separation distances compared to 10’s of centimeters for slab-gels.
As is directly relevant to next-generation pathology tools, we are applying the single cell targeted proteomic tool to assess oncoprotein isoforms. In particular, with our collaborators at the Stanford University School of Medicine, we measure truncated oncoprotein isoforms are not detectable via genomic or transcriptomic tools. The truncated proteins are created by proteolytic cleavage (ADAM10) and alternative translation, thus requiring direct protein measurement. While much attention has centered on nucleic acid assays in single cells and small tissue sections, paltry attention has focused on direct measurement of protein targets, a major innovation of our work. Immunoassays for truncated oncoprotein isoforms are impossible, as isoform-specific antibodies do not exist. While mass spectrometry is ideal for protein isoform detection, mass spectrometry is not sensitive enough for single-cell measurements. Consequently, immunoassays are the de facto standard for protein measurement in tissues down to single-cells. Yet, immunoassays fail if target-specific antibodies do not exist, as is the case for oncoproteins including HER2; available antibody reagents cannot distinguish the various truncated HER2 isoforms. Our isoform measurement innovation draws on microfluidic integration of multi-stage assays. We will measure proteins directly by employing separations first, and only after isoforms are resolved by size do we use antibody probes. Consequently, even moderate-to-poor performance antibodies are suitable, as separations mitigate poor antibody selectivity. While developed for HER2, tools for protein isoforms will be broadly relevant, as several driver oncoproteins are isoforms (e.g., RON33, MDM234, MYC35, BAG-136, PPM1D37, FLIP38).
Further, high background signal and off-target interactions limit the number of proteins detectable in the same, individual cell to ~5 (workhorse ICC & flow cytometry). Multiplexing to >10 proteins has been demonstrated, but employed uncommon tools that can be difficult to apply. To multiplex a Protein Panel, we use a dual-pronged approach: (1) create micro-gradient gels (similar to gradient gels) but created via gray-scale photolithography to achieve high quality microscale separations and (2) employ a novel photo-active polyacrylamide gel that supports protein electrophoresis (molecular sieving) then switches to a protein immobilization scaffold for subsequent immunoprobing. During brief UV illumination, benzophenone moieties in the polymer abstract hydrogen from C—H bonds on the protein, resulting in stable covalent protein immobilization. Covalent immobilization allows chemical stripping for tens of re-probing rounds thus allowing multiplexing up to 10’s of protein targets
Classification of tumors suffers from tremendous pathologist-to-pathologist variability in interpreting IHC. Gene screens yield quantitative readouts, oncoprotein isoforms cannot be measured, and multiplexed signaling protein assays are not feasible. I will discuss these advances in microfluidic assay integration for high protein specificity even in small tissue samples and single cells as is directly relevant to precision medicine.
C-C Kang, J-M Lin, Z Xu, S Kumar & AE Herr. "Single-cell western blotting after whole-cell imaging to assess cancer chemotherapeutic response." Analytical Chemistry, 2014, 86(20):10429-36.
AJ Hughes*, DM Spelke*, Z Xu, C-C Kang, DV Schaffer, & AE Herr. "Single-cell western blotting." Nature Methods, 2014, 11(7):749-55.*Equal contributors
See more of this Group/Topical: 2015 Annual Meeting of the AES Electrophoresis Society