The Seattle-based company, Adaptive Biotechnologies, grew out of research collaborations at the Fred Hutchinson Cancer Research Center. The core technology offered by Adaptive is a high-throughput method for sequencing and quantification of rearranged antigen receptors on T- and B-cells.
Coauthors Dr Anna Sherwood and Dr Ryan Emerson talked to us, along with Dr Mark Rieder, VP of Operations at Adaptive Biotechnologies, about their recently published paper, Using synthetic templates to design an unbiased multiplex PCR assay , and their advancements in massively parallel sequencing for profiling of the diversity of an individual’s T-cell receptor (TCR) repertoire. This immunosequencing assay is a new approach to quantifying unknown T- and B-cell input templates and has many potential diagnostic applications, including tracking the presence of cancer cells in leukemia patients.
Genotyping diversity of T- and B-cell receptors
The diversity of T- and B-cell populations is vast and until recently it has not been possible to know the quantitative composition of the many receptor types in a given sample. For example, just for TCR-β receptors there are potentially trillions of possible rearrangements that can be produced from the germ-line receptor gene, meaning that the TCR-β gene acts as a unique sequence tag for each T cell. A single T-cell can be identified by the TCR that it displays, as well as any cells that came from the same parent cell. Cells displaying the same TCR due to common descent are known as clones, and they are central to the adaptive immune response. In total, at any one point in time profiling the antigen receptor of circulating T- and B-cells represents a sequencing space nearly 30x greater than the human genome (3 Gb).
Identifying a high-copy clonal population of cells that is expanded due to some pathological process is relatively easy to do by sequencing of a small sampling of T- or B-cells (e.g., in acute lymphocytic leukemia), and looking for an abnormally high proportion of cells sharing a single receptor sequence. The problem arises when investigators are interested in low-copy, or single-copy, clonal cell populations (e.g., in tracking the residual burden of a leukemic T cell clone after most of the cells have been killed off by therapy).
Until the development of next generation sequencing, only low-copy clonal populations that had been previously identified, and that could be amplified for sequence analysis, could be tracked. The resolution for other low level clones of potential interest, or the overall sequence representation of all the receptor types in a sample could not easily be determined.
The adaptive immune system is a very dynamic system that is time-varying and constantly responding to environmental interactions. Understanding in detail these dynamic changes in immune cell diversity, even for very low copy clones, will significantly increase our knowledge regarding the adaptive immune response.
Next generation sequencing and multiplex PCR
In their publication , researchers at Adaptive Biotechnologies examine the makeup of T-cell receptor gamma (TCRG) V- and J-gene sequences, as a representative example of their approach. The first step in sequencing using the Illumina platform involves isolating the DNA of interest. The V- and J-gene combinations are isolated from samples using multiplex PCR to amplify rearranged receptor genes, using primers that recognize all of the possible combinations of these randomly mixed gene regions.
Using multiplex PCR to isolate TCRG sequences is a powerful method, but requires careful optimization in order to avoid changing the quantitative representation of the receptor genes in the pool, and to prevent low-level or under-amplifying sequences from being left out of the final data set.
When performing multiplex PCR, the primers need to have similar amplification efficiencies in order to make comparisons of data quantitative; this can only be done experimentally by conducting a multiplex PCR on a sample with known concentrations of input target sequences and quantitatively comparing the results to output values. Learn more about multiplex qPCR in the DECODED article, Multiplex qPCR—how to get started.
Adaptive has developed a very powerful tool, taking advantage of recent advances in synthetic biology which allow them to use multiplex PCR and sequencing to quantitatively look at the adaptive immune response with high accuracy and throughput. According to Dr Rieder, “if other companies do not have the same type of definitive standards, where they can identify their amplification bias and adjust for it, they cannot make the same claims about their assay being comprehensive and quantitative.”
Adaptive multiplex PCR optimization
In order to use multiplex PCR to quantify TCRG receptors in clinical samples, Drs Sherwood and Emerson explained how they address the amplification bias of the multiplex primers. The key to quantifying the amplification bias of all of these potentially interacting primers is to have an input in which the exact quantitative composition of the pool is known. From there they can experimentally quantify the output from that pool, and see how it differs from the expected results. For most genes, obtaining a sample with known input quantity is trivial (each gene amplified in the multiplex PCR is generally present at the same copy number, two per genome). When considering multiplex PCR for rearranged immune receptor genes, however, finding a precisely-measured pool of input templates has proven elusive.
Depending on the receptor type in question, the authors designed amplification primers, and corresponding synthetic receptor sequences, using gBlocks® Gene Fragments. The gBlocks Gene Fragments allowed them to create a “gold standard synthetic immune system” where the binding sites for every possible forward and reverse primer combination for a given receptor type were represented. In order to represent the human TCRG locus with 14 V domains and 5 J domains, the authors used 56 gBlocks Gene Fragments—J1 and J2 domains were amplified with the same primer sequence . Figure 1A illustrates the V and J sequence locations in the TCRG locus in the human genome .
The synthetic templates, shown in Figure 1, were designed with terminal, universal adaptors A and B, and therefore, would amplify with the same primers. The universal adaptors were important because they served as an anchor for adaptors needed to quantify each template using a simplex PCR approach. After the templates were pooled and amplified from the universal adaptors, a quantitative representation of each template was determined by Illumina NGS sequencing analysis.
Internal to the universal adaptor sequences were combinations of forward and reverse priming sites for the antigen receptor V- and J-gene combinations, recognized by the multiplex primers. Using the multiplex primers recognizing these sequences, the researchers could determine whether the primers over- or under-amplified any of the TCRG receptor sequences.
Figure 1. Design of synthetic TCRG templates. (A) Shows the organization of the human TCRG locus on chromosome 7, and the relative positions of the 14, V and 5, J motifs . (B) gBlocks Gene Fragments are used to create synthetic templates for an artificial immune receptor pool. A template is created for every possible V- and J- gene combination that is found in T-cell receptor gamma (TCRG) sequences. Universal adaptor sequences (UA and UB) are included for quantification of the synthetic template pool. A unique barcode (BC) is also included in each template for each V-gene/J-gene pair to provide additional verification beyond the gene sequence, as well as to make it possible to sequence the artificial template alongside the experimental sample. IM1 and IM2 are artificial sequences that further identify the gBlocks Gene Fragments as artificial templates during sequence analysis .
Adaptive used this data to iteratively adjust every single primer in their mix. Even if a particular primer combination was only off of the average amplification efficiency by a small amount, it still would be adjusted. Dr Emerson explained, “Since we know what the exact right answer is, we can adjust that primer concentration in the mix to get it as close to the correct answer as possible. If a primer is severely outcompeted in the multiplex, then we redesign it. We included a large segment of the V- and J-gene segments in the synthetic templates so that we can adjust priming sites as necessary without changing the gBlocks Fragments.”
After the final adjustments were made, one more round of quantification was performed with the same reagents and methods. At this point, remaining bias was reduced towards the mean using computational methods (Figure 2).
Figure 2. Example of multiplex primer optimization for human IGHV (hsIGHV) sequences. Similar to the published TCRG example, a synthetic pool of IGHV targets was created with gBlocks Gene Fragments. The data shows the experimentally-observed amplification bias for the multiplex primers, using the synthetic receptor sequences (black bars). Following this initial multiplex experiment, the amplification bias of each primer is determined as the difference from an average amplification efficiency, or normalized value of 1. The concentration or design of the multiplex primers is then adjusted and the experiment repeated (red bars). Finally, much of the residual bias is removed by computation methods (blue bars). The adjusted multiplex primer pool is then ready to quantify even low level clones in unknown, complex biological samples. Unpublished data (source: Adaptive Biotechnologies).
The remaining features of the template that were not gene or adaptor related were: unique barcodes for each primer-pair combination, and two short sequences labeled IM1 and IM2 that contained sequences that identified these as synthetic templates (Figure 1). These sequence features helped with quantification of the pool, and elimination of any contaminating sequences. Such sequence features could also be useful in experiments where investigators want to have inline, positive controls.
"gBlocks Gene Fragments are a game-changer”
Adaptive Biotechnologies initially tried to make all of their synthetic templates themselves. They did this using overlapping primer and PCR assembly, but these templates lacked a lot of the features used in the gBlocks Gene Fragments templates described here. Following the initial prototyping, the researchers realized that they had spent a lot of time and resources, without achieving their desired result.
Adaptive Biotechnologies was an early adopter of gBlocks Gene Fragments, because they knew that gBlocks fragments would be a “game-changer” for their assays. gBlocks Gene Fragments allow them to benchmark their assays, and reduce amplification bias; and these successes have really set the Adaptive technology apart from its competition. Adaptive Biotechnologies knew they could make a better assay faster, and the sequence verification that IDT performs on gBlocks Gene Fragments meant that the Adaptive Biotechnologies scientists didn’t have to worry about assembly errors.
Looking at clinical samples
After optimization, Adaptive Biotechnologies has a multiplex PCR and sequencing assay that is able to give a very accurate picture of the quantity of clonal lymphatic cells, in this case T-cells that are circulating in the bloodstream. In their paper , the authors focused specifically on leukemia samples. In leukemia, there is typically one malignant clone circulating in the blood from which all other cancer cells descend. Using their assay, the Adaptive Biotechnologies scientists can get an idea of the number of those clones before, during, and after treatment.
Rigorous multiplex optimization ensures that even low level clones do not drop out because of experimental error. The necessary amount of multiplex PCR and sequencing varies depending on where an individual is in the progression of their disease. Before treatment, only a minimal amount of material is required to detect a highly expanded clone or set of clones, in the initial sample where the expanded clone will often consist of >50% of the identified sequences.
Following treatment, these clones may be as infrequent as 1:100,000, and analyzing 500,000–1,000,000 blood cells would be required to detect them. Even a single malignant cell left circulating can be a source for a new pool of cancer cells. In their paper, the scientists at Adaptive Biotechnologies showed that they were able to detect and quantify clones against “a complex biological background,” that were present at a level of only 1 T-cell per 100,000 total input cells . This level of quantification opens up a whole new level of sensitivity when trying to monitor long-term minimum residual disease (MRD) of hematological cancers.
Adaptive uses IDT
In addition to gBlocks Gene Fragments, Adaptive Biotechnologies uses several other IDT products. Adaptive purchases their custom oligos from IDT, and takes advantage of the available custom formulation and mixing services to get their primers delivered in a ready-to-use format, which saves them a lot of time. Overall, the researchers’ experience with IDT has been positive, and their early use of gBlocks Gene Fragments has been beneficial to Adaptive Biotechnologies.