Scientists from the Fred Hutchinson Cancer Research Center and recent spinout company Adaptive Biotechnologies have published research demonstrating that the use of a synthetic analogue of a somatically rearranged immune cell receptor locus can be used to quantify and correct multiplex PCR amplification bias.
Published last week in Nature Communications, the study is the first to explicitly show how the method can be implemented in a clinical capacity upstream of high-throughput sequencing to provide accurate, quantitative measurement of clonal rearranged immune cell receptor loci to monitor minimal residual disease in T-cell acute lymphoblastic leukemia patients, Hutch researcher and corresponding author Harlan Robins told PCR Insider.
Adaptive Biotechnologies has been using the technique in its CLIA-certified laboratory for several months as part of a service-based immune cell sequencing offering called ImmunoSeq and a clinical assay called ClonoSeq to detect MRD in leukemia and lymphoma patients. Last year, PCR Insider sister publication Clinical Sequencing News had noted that the company was aiming for CLIA certification by the end of that year.
Having achieved this milestone, the company is now also developing a kit so other clinical labs can implement the method on their own for a variety of applications, Robins said.
"This would start out as a [research use-only] kit for whatever purposes someone wanted to use it for," Robins said. "If they wanted to look at MRD, they could, but they could also look at the immune response in, [for example], rheumatoid arthritis. I have projects ongoing right now looking at changes in the adaptive immune system of the elderly. You could look at changes in immune response to help with vaccine development or to find neutralizing antibodies to HIV. The potential set of applications is enormous."
Immune diversity and the body's ability to mount a response against nearly any pathogen is possible in part because the genomes of B and T cells undergo somatic rearrangement of cell surface receptor gene segments, the researchers explained in their paper.
Most of the diversity in B- and T-cell receptors is contained in a specific region called CDR3, and in T-cell receptors these regions are formed by rearrangements of so-called variable and joining, or V-J, gene segments, among others. Furthermore, once a specific rearrangement occurs, the immune cells harboring these unique receptors can clonally expand to mount an immune response — in an adult human, there are millions of different T-cell receptor rearrangements carried by several billion circulating T cells, the researchers explained.
Accurately measuring changes in the abundance of each clone is vital for understanding the adaptive immune response; or, in the case of leukemia and lymphoma, it's crucial for accurately quantifying cancerous immune cells with unique receptors for the purposes of MRD monitoring.
Researchers have previously attempted to use multiplexed PCR to simultaneously amplify multiple loci in an effort to monitor the adaptive immune repertoire, but it has proven challenging because of massive amplification bias caused by non-specific amplification, primer-dimer formation, and other problems. This bias is magnified when the multiplex PCR is being used upstream of highly quantitative sequencing.
"This has been a long-standing question — how do you do a multiplex PCR without having massive bias?" Robins said. "You're putting a whole set of PCR reagents into the same tube, and in principal they're amplifying all different molecules, but how do you know that you're not getting varying levels of amplification depending on what those molecules you're amplifying are?"
Robins noted that researchers have developed "four or five tricks" to try and balance PCR primers and help eliminate some of this bias. However, "I think it is abundantly clear that this just doesn’t work that well," he said. "We and everybody else were doing the best we could, and to be honest it just wasn't very good."
When attempting to sequence rearranged adaptive immune cell receptors — essentially immune receptors that have highly variable V-J combinations — Robins' group was "still getting hundreds of fold difference … even with our best PCR design," he said. "For instance, you might have five copies each of one molecule and another molecule, and you could get a 200-fold difference in the amount of amplified products between the two. It just wasn't quantitative."
To solve this the researchers developed a multiplexed PCR and sequencing approach that uses a synthetic analogue of a somatically rearranged T-cell receptor locus, called TCRG, to quantify and correct multiplex PCR amplification bias. The actual in vivo TCRG repertoire is impossible to know a priori, they explained in the paper; therefore, they generated a synthetic repertoire that includes a template for every possible V-J combination.
Using those synthetic templates, Robins and colleagues first identified and corrected the amplification bias present in their initial assay. Then, they measured the effect of primer concentration on amplification rates, and used that information to titer the relative concentration of each primer in the multiplex reaction such that all V-J combinations amplified with similar efficiencies. Finally, they used computational methods to remove residual differences in amplification efficiency.
Next, the researchers demonstrated the clinical applicability of their assay by using it to quantitatively measure clonal TCRG sequences for MRD monitoring in T-cell acute lymphoblastic leukemia patients. Specifically, they applied their assay to samples collected from 36 T-ALL patients, and found that, for patients with a clonal TCRG rearrangement, their assay results were concordant with multi-parameter flow cytometry, with no false negatives. The PCR-based assay was also able to detect MRD in 10 additional patients with a greater sensitivity than flow cytometry.
"Because these cells have virtually unique rearrangements in their genomic DNA, we use that rearrangement as a molecular tag of a cancer," Robins said. Others have taken this approach using allele-specific PCR methods, but that technique requires the generation of new assays for every patient.
"The reason we're not doing a different [assay] for each person is the whole multiplex is in the same well," he said. "That concept only works if you're quantitatively getting the same answer no matter what the different cancer is."
Clinical implementation of the method involves Adaptive Biotechnologies receiving a bone marrow sample at the time of patient diagnosis. Using acute lymphoblastic leukemia as an example, Robins said, a patient with the disease would have a single B- or T-cell receptor clone that would likely constitute an enormous percentage of all such cells in the bone marrow. Whereas, if a patient did not have leukemia, "the largest B- or T-cell receptor clone in the sample … would be a very, very small percent — way less than one percent," Robins said.
Then, a patient receives treatment with the goal being to eliminate all the cancer cells. "A few months after treatment, you send us a sample of your bone marrow or blood, and we look and see if you still have those cancerous cells in there," Robins said. "Now we have a molecular tag of that B-cell or T-cell rearrangement — it is the molecular tag of the cancer. We're able to read that at one cell in a million. This works, and it's way more sensitive than flow cytometry, which works at a level of about one in 10,000 cells."
The researchers believe that their method can be generally applied to other adaptive immune receptor loci, and in fact should enable the development of any multiplex PCR system where quantitative results are of interest — for example in downstream qPCR as opposed to high-throughput sequencing. As such, the company is working on moving its assay into kit form so that "we can truly get the same answer between different labs that would use this technology," Robins said.
Robins, who besides his position at Hutch is a consultant to Adaptive, said that the company is also in the early stages of working with other research groups at the Hutch to develop additional cancer diagnostic applications based on the method.
"Take, for example, sequencing some number of exons as a cancer diagnostic," he said. Currently a popular approach is to enrich target molecules using a hybridization method. "This would be a completely different strategy [that] would have some significant advantages over using hybridization technology," he said. "We are applying it in that case for a couple of different … cancer diagnostics looking at differences in the genome." Robins declined to elaborate, noting the early and proprietary nature of the work.