New Technique for Single-Cell Sequencing Enhances Drug Screening Research

A new study by scientists at the Herbert Irving Comprehensive Cancer Center (HICCC) introduces a more cost-effective, scalable single-cell sequencing technique for co-sequencing RNA and DNA. The tool, called DEFND-seq (DNA and Expression Following Nucleosome Depletion sequencing), uses an existing microfluidics platform that can process cells in parallel, allowing cancer researchers to better explore questions related to disease progression, tumor evolution, and drug resistance. 

Analyzing both RNA and DNA from several thousands of single cells within a tumor sample could help researchers understand how the genome and transcriptome interact to drive cancer growth. Current methods for simultaneous sequencing are limited by throughput, only processing a few thousand cells at once. They rely on manually manipulating individual cells in multi-well plates, flat plates embedded with many small test tubes.  

“The use of multi-well plates requires more material, in terms of enzymes and other expensive ingredients, and puts a cap on the number of cells you can practically analyze from a single sample,” says Peter Sims, PhD, associate professor of systems biology at Vagelos College of Physicians & Surgeons (VP&S) and co-leader of the Precision Oncology and Systems Biology program at the HICCC. “With this new technique, we're using a droplet microfluidic system that makes it really easy to process tens of thousands of cells in parallel.” 

An Accidental Discovery Sparks a New Technique 

DEFND-seq leverages an existing commercial droplet microfluidics system from 10x Genomics, a platform that is widely and routinely used in single-cell studies around the world. Reactions take place in tiny droplets, rather than wells, minimizing the amount of sample material and reagents needed for each reaction.  

DEFND-seq originally came about from a mistake in the lab. As scientific director of the Sulzberger Columbia Genome Center, Sims observed many experiments where an existing test, called ATAC-seq, would fail. Normally, the proteins act like spools of thread that wind up regions of DNA so they can't be accessed, and ATAC-seq analyzes the regions that are left open. Sims realized that disrupting the chromatin packaging would enable sequencing to occur on the entire DNA strand.  

“Some accidents in the lab led us to stumble into this possibility of basically doing ATAC-seq wrong on purpose,” he says. “Essentially, we're messing up an ATAC-seq protocol so that all the protein falls off of the DNA, and that allows us to use very similar chemistry to profile the genome more uniformly.”  

Enhancing Drug Screening in Brain Tumors 

Tim Olsen, a postdoctoral researcher in the lab, performed a series of benchmarking experiments where they applied DEFND-seq to various glioblastoma samples, including archived surgical specimens. Glioblastoma is an aggressive primary brain tumor with a median survival of 15 months. Sims, whose lab focuses on glioblastoma and other neurological disorders, plans to use DEFND-seq to help analyze cell type-specific drug responses.  

“We can take a surgical resection from a human tumor and perform a drug screen directly on the intact tumor tissue, including with drug combinations,” says Sims. “Historically, we've used single-cell RNA sequencing as our primary readout for these screens and that allows us to identify the cell types that are impacted by each drug, including non-tumor cell types.” 

With the addition of high-resolution DNA data provided by DEFND-seq, Sims and his colleagues can easily map the drug responses for subclones, defined as cells with distinct genetic lineages, within heterogeneous tumors. In particular, the technology can identify specific subclones that drive mechanisms of drug resistance.  

“We've got some really exciting work going on in our cancer center, including efforts to develop local delivery of drugs across the blood-brain barrier that we think are going to be really impactful,” he says. “We're trying to make this drug screen platform as sophisticated as possible so that we can look on a cell state-by-cell state basis at what drugs we can combine together to maximally ablate the tumor.”