SAN FRANCISCO (GenomeWeb) – Correlating genomic information with spatial location was a big theme at last week's Advances in Genome Biology and Technology meeting in Marco Island, Florida.
In particular, 10x Genomics announced that it will commercialize a spatial gene expression technology in the second half of the year and NanoString presented customer data from its Digital Spatial Profiler platform, GeoMx, which it plans to launch this month at the American Association for Cancer Research meeting.
Researchers attributed the interest in spatial genomics to a number of factors. First, with the increase in single-cell sequencing studies and Cell Atlas projects that seek to characterize every cell type, there's an increasing interest in not only sequencing and subtyping specific cells, but also understanding where those cells are located. In addition, for applications like cancer, spatial information is crucial for understanding tumor heterogeneity.
10x's Visium Spatial Gene Expression Solution will be based on technology originally commercialized by Science for Life Laboratory spinout Spatial Transcriptomics, which 10x acquired last year.
The Spatial Transcriptomics technology relies on overlaying a pathology tissue slice onto a microarray containing discreet barcodes in order to generate spatially barcoded cDNA molecules, after which standard library prep and sequencing can be performed.
"The power of the approach is that you are getting two orthogonal data types — a classic histopathology high-resolution image and also a gene expression profile from the tissue," Tarjei Mikkelsen, 10x's vice president of biology, said during a company-sponsored workshop at AGBT.
The Spatial Transcriptomics technology uses printed microarrays with 100-micron diameter features that are spaced along the array with 200 microns between the center of each. The features contain barcodes and capture probes. A histopathology slide is placed on top of the microarray and cDNA synthesis occurs directly on the array, afterwhich the barcoded molecules are cleaved off for further analysis.
In a Nature Communications study last year, researchers from the Karolinska Institute and the SciLife Lab in Sweden described the technology and how they used it to analyze the transcriptomes of around 6,000 tissue regions from 12 tissue biopsies of a prostate cancer tumor.
10x continues to sell that product, but is also developing what will be the Visium product. Mikkelsen said that the firm has optimized each step of the workflow, which will reduce the turnaround time from three days to nine hours and improve the sensitivity.
In addition, the company is upgrading the array to make it denser and to make the 100-micron features smaller, Mikkelsen said. The current 100-micron features capture between three and 30 cells each, depending on the size of the cell. 10x has not yet disclosed the feature size it will commercialize on Visium, but Mikkelsen said it would be smaller than 100 microns.
Aside from changing the array itself, Mikkelsen said the company is also integrating its own reagents into the process, which will lead to an increase in the number of unique molecular identifiers and thus the number of genes that can be analyzed within any given feature. Visium will be a standalone product, and not something that is run on 10x's Chromium instrument.
At AGBT, Sanja Vickovic, a Wallenberg Fellow at the Broad Institute who previously worked on the Spatial Transcriptomics technology at SciLife Lab, described in a presentation how her group at the Broad was using the technology to evaluate host microbiome interactions in the gut. To enable microbiome analysis, Vickovic said that the team modified the protocol to also ligate 16S primers to the barcoded beads, so that the host RNA could be sequenced along with the bacteria, and it would be possible to identify where in the colon certain bacteria species were more common.
Vickovic said that the goal is to use the method to analyze colorectal cancer samples, since some studies have shown that the certain bacteria may play a role in colorectal cancer growth.
In addition, she said the Broad team is continuing to work with the Spatial Transcriptomics technology and is further developing it independently of 10x Genomics. For instance, in a paper published on the BioRxiv preprint server last week, the team demonstrated their efforts to increase resolution by using attaching barcoded oligonucleotides to beads that it randomly deposits into wells that are two microns in diameter.
"We're continuing to further develop the technology and make it more robust," she said. "And I don't think we'll stop at 2 micrometers."
NanoString, meanwhile, first highlighted customer data from its GeoMx spatial profiling instrument at last year's AGBT. GeoMx uses slides stained with oligo barcodes to tag various protein and RNA molecules on a slide. The user looks at the slide under a microscope and selects regions of interest for further analysis. When UV light is shone on the region of interest, the oligo tags are cleaved and removed for further analysis. In addition, the process does not destroy the slide, Joe Beechem, NanoString's senior vice president of R&D, said at AGBT. So, after an initial analysis, researchers can store the slide and analyze a different region of it later.
Originally, NanoString designed GeoMx to be compatible with its nCounter instrument, but it also plans to make the system compatible with next-generation sequencing.
In addition, two groups from MD Anderson and the Netherlands Cancer Institute, who were part of NanoString's technology access program, have already published study data on the system in Nature Medicine. Both groups used it to evaluate biomarkers associated with immunotherapy response.
During a presentation at AGBT, Katherine McNamara, an MD/PhD student at Stanford University, described how her lab has used GeoMx to profile proteins and RNAs in the tumor microenvironment of HER2 amplified breast cancer patients to better understand treatment response.
In her presentation, she discussed how analyzing proteins and RNA of HER2-positive patients at multiple time points throughout treatment enabled the group to determine how the tumor microenvironment and HER2 signaling pathway changed throughout treatment.
McNamara's group used a 108-gene RNA panel and a 43-protein panel to evaluate patients. Importantly, she noted, decreases in the HER2 protein occurred after just one cycle of targeted therapy. When the group analyzed patients who ultimately achieved a complete response to those who did not, the decrease in HER2 signaling was "more dramatic" in those who achieved a complete response versus those who did not.
McNamara also noted that being able to pick regions of interest to profile helped to "better tease out what's going on." For instance, she said, the spatial information allowed the team to identify both decreases in the AKT protein and the AKT1 mRNA, a change that was not identified by analyzing bulk sequencing data. In that case, she said, "you're not profiling the tumor in isolation, but also with the surrounding stroma, and tumor microenvironment."
She added that in the future, the team plans to use GeoMx to analyze how targeted therapy impacts immune changes and to correlate protein and RNA markers during treatment within individual cases.
Also at AGBT, two customers discussed their early experiences with GeoMx in a NanoString-sponsored workshop. John McPherson, deputy director of the University of California, Davis Comprehensive Cancer Center, discussed how his lab is using GeoMx to better understand the molecular signatures associated with melanoma, with a goal of better diagnosing the disease early and identifying prognostic biomarkers.
"Early diagnosis of melanoma is critical to improving outcomes," but it can be tricky to distinguish between a regular mole, precursor to cancer, and cancer itself, McPherson said,. While molecular diagnostics are being developed to help, "low tumor cellularity and purity complicate the molecular analysis," he said.
As such, he said, he wanted to see whether being able to integrate spatial information — so that researchers could see which section of the tumor they were analyzing with molecular methods — could help improve diagnosis.
McPherson said that his team first ran GeoMx to validate its performance and then analyzed 12 patient samples using an RNA panel of 1,412 genes. The first thing the team noticed was how the data clustered. Data from an immune-rich region, for instance, clustered together, as expected. But, notably, the data also clustered based on the sample's biological classification — the benign mole samples clustered together and the melanomas clustered together.
"So it does seem that there are distinct signatures," McPherson said.
When the researchers drilled down into the data, they identified some known genes, such as PRAME, which were enriched in the melanoma samples. That gene is already used as a diagnostic biomarker using RT-qPCR in conjunction with another gene, McPherson said. "We're now looking to find more genes that could potentially increase the diagnostic value of that assay," he said. McPherson said that one next step would be to look at specific tumor regions using GeoMx. The team also plans to analyze samples from patients with known outcomes to look for prognostic biomarkers.
Another early-access user, Peter Nelson, a member of the divisions of human biology and clinical research at the Fred Hutchinson Cancer Research Center, discussed at AGBT how he is using GeoMx to look at the tumor microenvironment of prostate cancer.
Nelson said that his group initially analyzed serial FPPE samples from metastatic tumors, testing an RNA panel of 109 genes, a protein panel, and an RNA panel of 1,412 genes. "The reproducibility was remarkable," he said of the system.
Nelson said that his team is using the system to try to better understand inter- and intra-tumor heterogeneity, for instance, how expression of the androgen receptor gene varies. Assessing tumor heterogeneity, particularly for the AR gene, is important, he said, because therapy often targets the AR pathway and mutations in the gene can cause resistance. "Therapeutic pressure appears to be creating new subtypes," he said, and assessing those subtypes can be "important for determining the likelihood of a therapeutic to eradicate the tumor."