NEW YORK (GenomeWeb) – Akoya Biosciences has launched an early-access program for its CODEX tissue imaging system.
The Menlo Park, California-based firm, which specializes in multiplexed protein analysis, has placed CODEX systems in ten sites around the country and is planning a broader launch of the platform later this year, said Brian McKelligon, Ayoka's CEO.
CODEX was developed in the lab of Stanford University researcher Garry Nolan, who is a co-founder of Akoya. A significant portion of the development work was done by Julia Kennedy-Darling, then a postdoc in Nolan's lab and now director of applications development at Akoya. The company was founded in late 2015 to commercialize the technology and closed a $9 million equity funding round last summer.
The CODEX platform uses oligonucleotide-linked antibodies to detect proteins of interest in tissue samples. Capable of measuring dozens of different protein targets in a piece of tissue, the system is designed to interface with conventional fluorescence microscopes, enabling them to image samples with high levels of multiplexing and at single-cell resolution.
The key to the system is the use of antibodies labeled with oligo barcodes. Researchers can stain their sample of interest with as many antibodies as they like (typically to between 25 and 50 targets, according to the company). They then add reagents targeting not the antibodies but the bound oligos, causing those oligos to fluoresce, making them detectable via microscope. By analyzing the sample iteratively — exciting a set of three oligos, denaturing them and stripping them from the sample, then repeating the processes with the next set of three — researchers can collect data on many more proteins than is possible using conventional fluorescence microscopy.
The system fits into a larger trend in life sciences research toward measuring large numbers of parameters per sample. Nolan is a leading researcher in this area, having played a role in the development of other systems for highly-multiplexed single-cell analysis like Fluidigm's CyTOF and multiparameter ion beam imager technology.
"If you look at any of the cell measurement technologies that are out there, there's a push towards measuring larger numbers of parameters either on a per-cell basis or a per-tissue image basis," he said. "We know that three elements of a cell or three elements of some expression profile in tissue is insufficient to really describe what's in there, the so-called heterogeneity."
"The idea is, basically, you want to measure as many things as you reasonably can and in a short enough time frame and on instruments that are cheap enough and with reagents that aren't so expensive that they break the bank every time you run an experiment," he said.
Conventional fluorescence microscopy can only look at around three to four parameters in a sample. The MultiOmyx system from clinical testing firm NeoGenomics is able to image dozens of protein or nucleic acid markers, but that system uses harsh treatments of the sample between iterative series of stains to collect this data — an approach that damages the tissue, limiting the number of markers that can be multiplexed and the ability to use analyzed samples in future assays, Nolan said.
The oligo denaturing and washing process used by the CODEX is much gentler, Nolan said. "The tissue ends up still being good at the end, and so you can do beautiful imaging."
It also means researchers can go back and re-interrogate samples with additional markers if new questions arise, he said. "You just go back in and re-register the tissue and let the computer find the alignment and you're ready to go."
Nolan said his lab had multiplexed as many as 120 markers in a tissue sample using the system but noted that 50 is a more typical number for experiments in his lab. In addition to 120-marker experiments being fairly expensive in terms of reagent costs, current informatics tools struggle to manipulate and analyze the data from experiments of that size, he said.
Another current limitation of the technique is speed, Nolan said. "Even though we're using the best of the microscopes that are out there, they still require a certain amount of time to do the cross-tissue [scan]."
This requires researchers to balance how many markers they want to measure and, along with it, how many iterations are needed, with how much of the sample they want to image and how long they want the process to take.
"It's always, how deep do you want to go on the parameters per cell; how high a resolution do you want; and how much of the tissue do you feel you need to capture to know that you've captured some biological or clinical meaning," he said. He added that one approach his lab is exploring is adapting the technique to light-sheet microscopy, which could speed the process by roughly 10-fold.
Nolan said his lab is using the CODEX in large part for immunophenotyping of cancers — looking at what kinds of immune cells are present in the tumor microenvironment for purposes like predicting clinical outcomes or the effectiveness of treatments like immunotherapy.
McKelligon said research into the tumor microenvironment is one of the main areas Akoya is targeting commercially as well, though he added that the company expects it will prove useful for a variety research questions. Among the areas of focus for the company's early-access customers are neurobiology, hematopoiesis, autoimmune disease, cell therapy, and developmental biology.
McKelligon said Akoya plans to offer oligo-linked antibodies to individual targets, allowing customers to build their own panels from reagents the company has validated for use in human and mouse fresh, frozen, and formalin-fixed paraffin-embedded tissue. It will also offer oligo conjugation kits so customers can generate their own antibodies in cases where Akoya doesn't offer ones they are interested in.
On the software side, the company is tackling three distinct challenges, McKelligon said. The first, which it is mainly working on internally, is developing tools to control and integrate the CODEX system and the microscope it is coupled to. The second, which it is partnering on with an outside lab, is developing data analysis and visualization tools. The third challenge is continuing to evolve and develop these analysis and visualization tools. For this, McKelligon said the company will "need to be proactive in partnering to build up a community and ecosystem of the best data analysis methods, many of which are not going to come from internally."
As Nolan noted, highly-multiplexed, high-resolution tissue imaging is a rapidly growing research area with a number of technologies and companies entering the space. For instance, last year, Fluidigm launched its Hyperion Imaging Mass Cytometry platform, which allows researchers to add sample spatial and structural data to the multiplexed molecular analyses enabled by the company's traditional mass cytometry systems. (The imaging system was developed in part by Bernd Bodenmiller, assistant professor for quantitative biology at the University of Zurich and another of Nolan's postdocs.)
On the mass spec front, Bruker in particular continues to move forward with MALDI imaging applications and has begun working with commercial and academic pathology labs in Europe with the aim of developing MALDI-based assays that could prove clinically useful.
Gordon Mills, formerly at MD Anderson Cancer Center and now head of precision oncology at Oregon Health & Science University's Knight Cancer Institute, is working to adapt NanoString's nCounter platform to measure large multiplexes of proteins in tissue while retaining spatial information.
In 2016 NanoString announced the release of its Digital Spatial Profiling technology, which allows multiplexed measurements of up to 800 proteins and RNA while retaining spatial information. That system uses antibodies attached to photocleavable nucleic acid barcodes to stain target molecules in samples of interest. The tissue is then treated with focused UV beams to cleave the barcodes, which are then collected and analyzed, providing a count of the tagged proteins or nucleic acids in the tissue region treated by the UV beam. According to the company, it is able to focus the beams to the single-cell level, meaning the technique can, in theory, achieve single-cell resolution.
Nolan said he saw the CODEX system, which lists for $65,000, fitting into academic or smaller pharma and pathology labs without the money for high-end mass spec systems or the technical expertise to run them.
McKelligon suggested the system would prove complementary to other imaging and single-cell analysis tools.
"Whether you're talking about single-cell sequencing, about mass spec approaches, about much higher-throughput but lower-plexing approaches like PerkinElmer has with the Polaris [pathology imaging system], I think it's all highly complementary, because we're filling an important and underserved [market]," he said.