Putting COVID-19 on the ‘Map’

Adaptive Biotechnologies and Microsoft are turning their nearly two-year-old partnership to map and decode the human immune system into a potential weapon against COVID-19.

The partners have agreed to study the global pandemic in an expansion of their collaboration, which focuses on mapping population-wide adaptive immune responses to diseases at scale, and to share the data they gather on COVID-19 publicly.

“We expect, probably by June, we’ll have sequencing of samples, and we’ll start to have done enough computational biology that we can start to make this knowledge more broadly known,” Adaptive Bio chief medical officer Lance Baldo, MD, told GEN Edge. “Our top objective is to make this information known to researchers, healthcare professionals, and public health officials throughout the world. And of course, we do hope that there are going to be some great products that can come out of this.”

One such product, Baldo added, could be a blood test for detecting SARS-CoV-2.

“This platform can absolutely develop a test for COVID-19. And throughout the process of this research, we’re absolutely going to evaluate how and if that makes sense,” Baldo added.

“Universal” goal

Development of a more expansive “universal” blood test for multiple diseases was among the purposes articulated by Adaptive and Microsoft when they launched their TCR-Antigen Map collaboration in 2018, with the aim of using immunosequencing and machine learning to map T-cell receptor (TCR) sequences to diseases and disease-associated antigens.

That goal remains, and Adaptive Bio is actively engaged with the FDA in talks to establish a regulatory path that would advance the universal test into diagnostic and clinical use, Baldo said.

Last year, the partners confirmed clinical signals in two diseases, and established their first proof of concept in Lyme Disease.

“We’re launching that clinical program toward the end of this year to validate the platform and then ultimately put that platform and put those data in front of the FDA,” Baldo said.

He added that Adaptive Bio is looking at celiac disease as a potential second area of high unmet need—one focused outside of infectious disease, on autoimmunity and anti-gliadin antibodies, and has early research focused on earlier detection of cancer.

The companies carry out immunosequencing through Adaptive Bio’s proprietary immune medicine platform, developed by co-founder and chief scientific officer Harlan Robins, PhD, while at Fred Hutchinson Cancer Research Center. Robins, brother of co-founder and CEO Chad Robins, headed the computational biology program before stepping down from the Fred Hutch faculty last year.

The platform takes the key cells of the adaptive immune system that enable the body to mount responses against antigens—T cells and B cells—and extracts their genomic DNA, sequences them, then identifies the receptor sequences that are contained within those cells. T cells and B cells randomly assemble different gene segments–called variable (V), diversity (D) and joining (J) genes–in a process called VDJ recombination, designed to generate antigen receptors that can collectively recognize molecules posing exogenous or endogenous threats to the body.

Looking for Clonal Expansion

In the case of COVID-19, Baldo said, Adaptive Bio will extract genomic DNA from the T cells, then look for signs of clonal expansion, in which T cell receptor sequences appear to be expanded in people who are exposed to or have active SARS-CoV-2 infection. Adaptive will carry out immunosequencing, apply computational biology and machine learning from Microsoft—as well as do its own in vitro studies to map the most likely T cell receptor sequences that will bind to, and potentially neutralize, the virus.

“The relevant receptor sequence might be very, very potent, but very, very infrequent. So, we’re going to find them all,” Baldo said. “We’re going to find the common sequences. We’re going to fund the uncommon sequences. And then ultimately with some of the in vitro work that we’re doing, hopefully we’ll be able to characterize the ones that seem to be most effective in terms of antigen binding.”

To generate immune response data, Adaptive plans to begin this month collecting de-identified blood samples through a LabCorp mobile phlebotomy service from patients diagnosed with or recovered from COVID-19, at their place of choice, in a virtual clinical trial managed by LabCorp’s drug development business Covance. The partners have an initial clinical collaborator—Providence, whose 51 hospitals include one near Seattle that treated the first U.S. COVID-19 patient—and are seeking additional institutions and research groups worldwide to contribute blood samples.

Immune cell receptors from the blood samples will be sequenced using Illumina platform technology—the specific platform had yet to be determined at deadline—and mapped to SARS-CoV-2-specific antigens confirmed by Adaptive’s proprietary immune medicine platform to induce an immune response. The immune response signature found from the initial discovery work and the initial set of samples will be uploaded to the open data access portal.

Using Microsoft’s hyperscale machine learning capabilities and Azure^® cloud computing platform, the companies said, they plan to improve and update the accuracy of the immune response signature online in real time as more trial samples are sequenced from the study.

$45M Investment

When the companies signed their collaboration agreement, Microsoft invested approximately $45 million in Adaptive Bio convertible preferred stock, while Adaptive Bio agreed to spend a minimum of $12 million over the agreement’s seven-year term on Microsoft Azure cloud services, Adaptive Bio disclosed in its S-1 registration statement when it launched its initial public offering in June 2019, raising approximately $321 million in net proceeds.

The companies are proverbial neighbors: Adaptive is based in Seattle, while Microsoft is headquartered 15 miles northeast, in the Seattle suburb of Redmond, WA.

“Making critical information about the immune response accessible to the broader research community will help advance ongoing and new efforts to solve this global public health crisis, and we can accomplish this goal through our proven TCR-Antigen mapping partnership with Adaptive,” Peter Lee, Microsoft corporate vice president, AI and Research, said in a statement.

The core of the TCR-Antigen Map is the biological data and models derived from Adaptive’s high-throughput, patented immunoSEQ sequencing and MIRA (Multiplexed Identification of T cell Receptor Antigen Specificity) antigen-mapping technologies to determine all possible T-cell receptors that bind to clinically relevant antigens across diseases.

immunoSEQ is designed to sequence single chains of “Y-shaped” T cell receptors or B cell receptors using next-generation sequencing (NGS), enabling users to understand the quantity and diversity of T and B cells in a biological sample.

MIRA maps millions of TCRs to thousands of clinically relevant antigens. When combined with immunoSEQ, the technologies elucidate what potential diseases a patient’s immune system has seen or is actively fighting.

“If you could eventually sample millions of T cell and B cell repertoires, you could start to identify which receptor sequences are specific to which antigens. And then use that to be able to either diagnose or to treat disease,” Baldo said. “Being able to create this tremendous library of T cell receptor sequences that then pair with particular antigens of interest, in some ways it’s really just allowing the immune system to tell us what threat it either is facing, or what it has faced in the past.”