Partners Adaptive Biotechnologies Corp., which is based in Seattle, and Redmond, Wash.-based Microsoft Corp. have started sharing an open database that details the immune response in COVID-19 patients with researchers and public health officials. The project is analyzing thousands of de-identified patient blood samples submitted from institutions around the world and is dubbed ImmuneCODE.
Adaptive is also conducting a prospective study known as ImmuneRACE (Immune Response Action to COVID-19 Events) that is enrolling 1,000 U.S. participants who had COVID-19. The aim is to analyze a set of T cells that recognize the SARS-CoV-2 virus, including data on the cellular response to viral antigens and patient outcomes, in order to better understand immunity. Adaptive is working to develop a COVID-19 diagnostic based on this T cell response. Microsoft is also an investor in Adaptive – and has roughly 40 full-time employees devoted to this relationship.
How it works
“Our work with Microsoft started a few years before and we basically had realized that in any given person there are many millions of different T cell receptors,” Adaptive co-founder and CSO Harlan Robins told BioWorld. “The idea is that the diversity of these T cell receptors is designed to protect you against any possible disease.”
“The basic philosophy is the immune system is just better at diagnosing disease than we are – and that the information about how it's diagnosing disease is contained in these T cell receptors because they are the ones that bind to the disease and specifically expand,” he continued. “Your immune system responds to just about every disease. So, if we could just identify which ones were expanding at a given time, we could know all the diseases you have just from a simple blood test.”
Immune cell receptors from all the T cells in a given patient blood sample are sequenced using Illumina equipment and mapped onto SARS-CoV-2-specific antigens that are identified by Adaptive to induce an immune response.
Adaptive has already collected roughly 1,000 retrospective blood samples from hospitals around the world, including the Seattle-based health system Providence, as well as several from Italy and Spain. These were all locations that were hard hit by COVID-19, and also include various strains of the virus. Enrollment for the U.S.-based, prospective ImmuneRACE study is almost at the halfway mark.
As for Microsoft, it was already powering Adaptive’s data-intensive research efforts and it has continued with the COVID-19 project to provide its Azure hyperscale cloud and machine learning offerings to analyze and cross-check this data in real-time as samples are sequenced.
“We had created a set of technologies that connects those receptors to the disease state. We had it at a high scale, but it's not anywhere near high enough scale to cover the huge diversity across the entire population,” explained Robins on the origins of the Adaptive relationship with Microsoft. “So, we needed to take the set of information we created and learn from it, to extrapolate to a much bigger set of data.”
“Basically, we needed machine learning to get from where we were and what we're able to generate to the much bigger result of being able to do this thoroughly,” he continued. “So, we started looking for a partner that was a machine learning expert and able to help us amplify that signal and Microsoft, for many different reasons, ended up being our partner of choice and has been a fantastic partner in this regard to me.”
T cells are part of adaptive immunity, cells that respond to antigen-specific threats, as are B cells. T cells respond quickly, after the initial response from innate immune cells, to fight the disease, while B cells are activated to produce antibodies typically weeks later.
Adaptive expects its focus on the T cell receptor will also help elucidate what parts of SARS-CoV-2, beyond the well-publicized spike protein, interact with the immune system. Its research is also expected to outline the huge variety of individual immune responses in this disease.
In both the retrospective and prospective studies, as well as with the development of its T cell receptor COVID-19 diagnostic, Adaptive is detecting the number of T cells with specific receptors for SARS-CoV-2. The number of T cells expands rapidly during active infection and then slowly dwindles. The expectation is that measuring the T cell response in combination with antibodies could provide some assessment of immunity.
Robins is critical of how meaningful SARS-CoV-2 serology antibody tests are on their own, a perspective that is widely shared. But he thinks the pairing of the two kinds of tests together could prove more useful in establishing a measure of immunity.
“If you just measure the antibody response, you're assuming the T cell response comes for free, so to speak. But that doesn't necessarily have to be true for the coronavirus,” said Robins. “About 25% or 30% of colds are caused by four different coronaviruses. You can get reinfected with those viruses on a relatively continual basis right, you don't get long-term protection against those viruses. There are plenty of studies where patients who even had significant antibody response titers against those viruses, still got reinfected. So, we fundamentally don't understand what it is to say that you're immune.”
“We don't have a great correlate of protection or assessment of what truly creates immunity. It's probably going to have to be both,” he continued. ‘If you asked me personally, what would be my best bet for someone who has increased immunity to the virus, I would say someone who has robust antibodies and T cells in their memory specific for the virus. That would be my best guess. If you're missing one or the other, you're probably not immune. To really get at immunity you probably have to have both.”
Adaptive plans to issue early data from its research in the near term within months as a preprint study, a practice that has become commonplace amidst the pandemic, that would be followed by a paper in a peer-reviewed journal. On its COVID-19 diagnostic, it is in early discussions with the FDA on what will be required to demonstrate precisely the sensitivity and specificity that will be required by the agency.
The company was in conversations with the FDA around its first diagnostic effort in Lyme disease prior to the encroachment of the pandemic. So, the agency already is familiar with the platform. Adaptive plans to submit to go ahead with a clinical trial sufficient for FDA approval this year in Lyme disease.