sequencing the circulating and infiltrating t-cell repertoire on the ion s5tm

1
Read Length (bp) Counts For Research Use Only. Not for use in diagnostic procedures. Thermo Fisher Scientific • 5781 Van Allen Way • Carlsbad, CA 92008 • thermofisher.com Geoffrey Lowman 1 , Elizabeth Linch 1 , Lauren Miller 1 , Denise Topacio-Hall 1 , Tim Looney 2 , Yongming Sun 2 , Mark Andersen 1 , Fiona Hyland 2 , Ann Mongan 2 (1) Thermo Fisher Scientific, 5781 Van Allen Way, Carlsbad, CA 92008 (2) Thermo Fisher Scientific, 180 Oyster Point Blvd. South San Francisco, CA 94080 Sequencing the circulating and infiltrating T-cell repertoire on the Ion S5 TM Results Introduction/Background Abstract 3130 T-Cell receptor (TCR) repertoire sequencing by next-generation sequencing (NGS) is a valuable tool for building a deeper understanding of the adaptive immune system. As immunotherapy, particularly T-cell therapies, show increasing potential in treating cancer, the ability to gain a detailed, unbiased view of the TCR repertoire becomes imperative for biomarker discovery, immune response to treatment, and study of tumor microenvironments. A key question the field seeks to understand is the relationship between circulating T-cells and infiltrating T-cells at the tumor site. Here, we present a novel AmpliSeq approach for TCR repertoire sequencing using the Ion Torrent S5 sequencer which leverages simplified workflows and offers up to 600 bp reads which allow for a more complete characterization of the entire V(D)J region of TCRβ. With a unique long read length capability, this method can leverage mRNA as input, which minimizes requirement as starting materials (10-500ng for typical use cases) and focusing sequencing to productive TCRβ arrangements. This AmpliSeq approach targets the constant (C) and the FR1 regions, minimizing the potential for primer bias and greatly increasing the phylogenetic information content compared to techniques that exclusively characterize the CDR3 domain. Our results show that the circulating T-cell repertoire size is approximately 2 orders of magnitude higher than the infiltrating T-cell repertoire. Accordingly, while it is difficult to fully capture the complete repertoire of circulating T-cells due to its vast diversity, we show that it is possible to reliably capture the complete infiltrating T-cell repertoire with multiplexing as high as 10 samples on the Ion 530 chip. Replicate runs of infiltrating T- cells offers correlation of ~0.9, indicating that the results were reproducible and the samples were sequenced to appropriate depth. In summary, we believe that this high multiplex workflow and fast turn- around time (<48 hours RNA-to-answer) will allow researchers to more routinely characterize infiltrating T-cell repertoire and offer the field a better understanding of the impact of repertoire diversity on tumor elimination. High throughput: >50,000 clones/ sample from blood, cell populations, or fresh-frozen tissue generated from a single tube workflow. Input cDNA Amplify targets using Oncomine TM TCR-Seq β Primer Panel Partially digest primer sequences Ligate Adapters Unbiased output: AmpliSeq TM V- and C-region primers are optimized to reproduce results from 600bp 5’- RACE protocols. Comprehensive: 400bp read length offers complete characterization of CDR1,2,3. Highly accurate: sequencing and amplification errors are corrected with statistical models leveraging TCR mRNA as input. Diversity view V gene CDR3 length (nt) Clonotype identification V-gene identification Read count Proportion of different alleles Analysis solution: Complete Ion Reporter analysis solution including clonotype identification and diversity, CDR3 length, and multisample analysis for longitudinal tracking for clones of interest. Sequencing lung tumor biopsy revealed 589 unique TCR clones. Oligoclonal repertoire with a small number of dominating clones Shannon Diversity - 6.78. Sequencing PBL revealed 45305 unique TCR clones. Diverse, polyclonal repertoire with few highly expanded T cells. Shannon Diversity - 13.95. Figure 2. Sequencing read length histogram from library generated using the Oncomine TM TCR-Seq β assay run on an S5-530 chip. Total read counts are typically between 15-20M reads, of which 70-80% are “productive” (blue+light blue) when run through our analysis pipeline. For Research Use Only. Not for use in diagnostic procedures. © 2017 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified. 2-day workflow using the Oncomine TM TCR-Seq β assay - Immune Repertoire sequencing using AmpliSeq TM library construction N1 N2 FR1 FR2 Diversity (D) Joining (J) Constant (C) Variable gene (V) CDR3 FR3 CDR1 CDR2 Figure 1. Schematic of a section of T-cell receptor which has undergone V(D)J rearrangement, identifying the CDR1, 2, and 3 regions, and the Framework (FR1, 2, and 3) regions. Using RNA/cDNA as input allows design of primers targeting the (V) and (C) regions. Figure 3. 10 FF tumor infiltrating lymphocyte (TIL) samples (NSCLC) sequenced on 530 chips. We see high concordance between identified clones (95.8-99.6%) comparing sequencing replicates – which indicates sequencing to adequate depth. The Oncomine TM TCR-Seq β assay harnesses the power of AmpliSeq TM library construction chemistry which allows for primer design targeting the FR1 region, and includes primers targeting all known (V) region rearrangements. The resulting sequencing libraries are generally 325-350bp in length. The AmpliSeq TM workflow allows for library construction in under 5 hours with 30 minutes of hands on time. The Oncomine TM TCR-Seq β assay enables sequencing on the Ion S5 TM 530 chip. The capacity of this sequencing arrangement allows for multiplexing of up to 16 samples per chip for samples with limited repertoire size (cultured cells, tissue). The ability to multiplex also enables sequencing runs which contain sample comparisons, i.e. – tissue and blood samples run on the same sequencing run (Figures 4-6). Figure 3 above depicts correlation plots generated for technical sequencing replicates for 10 fresh-frozen tumor infiltrating leukocyte (TIL) samples taken from non-small cell lung carcinoma (NSCLC) biopsies. High concordance between identified clones (95.8-99.6%) indicates sequencing to adequate depth. Immune repertoire sequencing using Ion Torrent sequencing includes a full analysis suite which outputs highly accurate assignment of clonotype populations. Repertoire sequencing is sensitive to base-call (substitution) sequencing errors, which can mimic the natural variation in the repertoire, making error correction strategies very difficult to implement. Ion Torrent sequencing has a very low substitution error rate, and the errors that are produced are predominantly InDel errors. This error type allows for correction using the knowledge that by using mRNA as an input to the assay, all reads should code for productive amino acid sequences, and InDel errors will present as frame shift errors. This permits for the streamlined identification of sequencing errors, and in most cases, error correction, leading to higher rates of productive reads from each sample. For more information on the Oncomine TM TCR-Seq β assay please see poster #3567: “Sequencing the human TCRβ repertoire on the Ion S5™ System” Session Category: Bioinformatics and Systems Biology Session Title: Sequencing Analysis and Algorithms Session Date and Time: Tuesday Apr 4, 2017 8:00 AM - 12:00 PM Location: Convention Center, Halls A-C, Poster Section 23 Poster Board Number: 3 Acknowledgements Figure 5. Spectratype plot of V-gene primer usage versus CDR3 length, showing a comparison of TCRβ clones identified to be unique to the PBL (left panel – 91.78% of all clones identified) and TCRβ clones shared in both the PBL and tumor repertoire (right panel – 8.22% of total). Circle size denotes level of clonal expansion. Figure 4. Diagram outlining sample comparison between FF tumor infiltrating leukocyte and peripheral blood leukocytes from a patient with NSCLC sequenced on a single 530 chip. Unique to peripheral blood: 91.78% Shared with tumor repertoire: 8.22% CDR3 length (nt) V-gene primer V-gene primer Figure 6. Correlation plot showing the same dataset shown in Figure 5. Data along the x-axis represents clones unique to the PBL, data along the y- axis shows those unique to tumor, as well as 219 clones that are shared between PBL and tumor. To test the ability of the assay to differentiate repertoire sample types, total RNA was extracted from peripheral blood leukocytes (PBL) and fresh- frozen tumor biopsy from an individual with Stage 1B squamous cell carcinoma of lung. 100ng of total RNA from each sample was used as template input for TCRβ sequencing on a single Ion 530 TM Chip. The results of this study are outlined in Figure 4; sequencing of the tumor biopsy revealed 589 unique TCR clones, an oligoclonal repertoire with a small number of dominant clones, resulting in a Shannon Diversity value of 6.78. Sequencing PBL led to 45305 unique TCR clones. Next, we carried out correlation to determine the populations of clones that were unique to either the tumor biopsy or the PBL, and which clonal populations are shared between sample types. Correlation data is presented using two views shown in Figures 5 and 6. Figure 5 displays a spectratype plot of V-gene primer useage versus CDR3 length, split into two windows: L) clones unique to the PBL and R) clones shared between the PBL and the tumor biopsy. The spectratype plot allows for a fast view of the differences between repertoires unique to PBL versus the shared clone population. Namely, the expected size and diversity of the PBL repertoire compared to the relatively less diverse repertoire of the tumor biopsy which includes fewer, more highly expanded clones. Figure 6 displays the same data which separates into three natural populations; along the x- axis, clones unique to the PBL, along the y-axis, clones unique to the tumor biopsy, and along the diagonal, clones which are shared between sample types. This view allows quick identification of clonal populations which may be interesting routes for further experiments, for instance, tracking of clones from the blood which are identified in the solid tumor, or expansion of tumors which are found only in the tumor sample. log 10 clone frequency in PBL log 10 clone frequency in tumor 370 clones unique to tumor 45086 clones unique to PBL 219 shared clones The authors would like to acknowledge the work of all who participated in this program: Xinzhan Peng, Alice Zheng, Alex Pankov, Lifeng Lin, Grace Lui, Gauri Ganpule, Sonny Sovan, Larry Fang, Tyler Stine, Laura Nucci, Matthew Cato, Yuan-Chieh (Jeffrey) Ku, Janice Au-Young, Rob Bennett, and Jim Godsey.

Upload: thermo-fisher-scientific

Post on 07-Apr-2017

19 views

Category:

Science


0 download

TRANSCRIPT

Page 1: Sequencing the circulating and infiltrating T-cell repertoire on the Ion S5TM

Read Length (bp)

Cou

nts

For Research Use Only. Not for use in diagnostic procedures. Thermo Fisher Scientific • 5781 Van Allen Way • Carlsbad, CA 92008 • thermofisher.com

Geoffrey Lowman1, Elizabeth Linch1, Lauren Miller1, Denise Topacio-Hall1, Tim Looney2, Yongming Sun2, Mark Andersen1, Fiona Hyland2, Ann Mongan2

(1) Thermo Fisher Scientific, 5781 Van Allen Way, Carlsbad, CA 92008 (2) Thermo Fisher Scientific, 180 Oyster Point Blvd. South San Francisco, CA 94080

Sequencing the circulating and infiltrating T-cell repertoire on the Ion S5TM

Results

Introduction/Background

Abstract

3130

T-Cell receptor (TCR) repertoire sequencing by next-generation sequencing (NGS) is a valuable tool for building a deeper understanding of the adaptive immune system. As immunotherapy, particularly T-cell therapies, show increasing potential in treating cancer, the ability to gain a detailed, unbiased view of the TCR repertoire becomes imperative for biomarker discovery, immune response to treatment, and study of tumor microenvironments. A key question the field seeks to understand is the relationship between circulating T-cells and infiltrating T-cells at the tumor site. Here, we present a novel AmpliSeq approach for TCR repertoire sequencing using the Ion Torrent S5 sequencer which leverages simplified workflows and offers up to 600 bp reads which allow for a more complete characterization of the entire V(D)J region of TCRβ. With a unique long read length capability, this method can leverage mRNA as input, which minimizes requirement as starting materials (10-500ng for typical use cases) and focusing sequencing to productive TCRβ arrangements. This AmpliSeq approach targets the constant (C) and the FR1 regions, minimizing the potential for primer bias and greatly increasing the phylogenetic information content compared to techniques that exclusively characterize the CDR3 domain. Our results show that the circulating T-cell repertoire size is approximately 2 orders of magnitude higher than the infiltrating T-cell repertoire. Accordingly, while it is difficult to fully capture the complete repertoire of circulating T-cells due to its vast diversity, we show that it is possible to reliably capture the complete infiltrating T-cell repertoire with multiplexing as high as 10 samples on the Ion 530 chip. Replicate runs of infiltrating T-cells offers correlation of ~0.9, indicating that the results were reproducible and the samples were sequenced to appropriate depth. In summary, we believe that this high multiplex workflow and fast turn-around time (<48 hours RNA-to-answer) will allow researchers to more routinely characterize infiltrating T-cell repertoire and offer the field a better understanding of the impact of repertoire diversity on tumor elimination.

High throughput: >50,000 clones/sample from blood, cell populations, or fresh-frozen tissue generated from a single tube workflow.

Input cDNA

Amplify targets using OncomineTM TCR-Seq β Primer Panel

Partially digest primer sequences

Ligate Adapters

Unbiased output: AmpliSeqTM V- and C-region primers are optimized to reproduce results from 600bp 5’-RACE protocols. Comprehensive: 400bp read length offers complete characterization of CDR1,2,3.

Highly accurate: sequencing and amplification errors are corrected with statistical models leveraging TCR mRNA as input.

Diversity view

V gene

CD

R3

leng

th (n

t)

Clonotype identification

V-gene identification

Rea

d co

unt

Proportion of different alleles

Analysis solution: Complete Ion Reporter analysis solution including clonotype identification and diversity, CDR3 length, and multisample analysis for longitudinal tracking for clones of interest.

Sequencing lung tumor biopsy revealed 589 unique TCR clones.

•  Oligoclonal repertoire with a small number of dominating clones

•  Shannon Diversity - 6.78.

Sequencing PBL revealed 45305 unique TCR clones.

•  Diverse, polyclonal repertoire with few highly expanded T cells.

•  Shannon Diversity - 13.95.

Figure 2. Sequencing read length histogram from library generated using the OncomineTM TCR-Seq β assay run on an S5-530 chip. Total read counts are typically between 15-20M reads, of which 70-80% are “productive” (blue+light blue) when run through our analysis pipeline.

For Research Use Only. Not for use in diagnostic procedures. © 2017 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher Scientific and its subsidiaries unless otherwise specified.

2-day workflow using the OncomineTM TCR-Seq β assay - Immune Repertoire sequencing using AmpliSeqTM library construction

Variable Diversity Joining Constant

N1 N2

A. Adult IGH or TCRBeta chain rearrangement

In adult B and T cells, the process of VDJ rearrangement very often involves exonucleotide chewback of VDJ genes and the addition of non-templated bases, forming N1 and N2 regions in the B cell receptorheavy chain CDR3 and the T cell receptor Beta chain CDR3. These processes vastly increase IGH and TCRB CDR3 diversity.

Variable Diversity Joining Constant

B. Fetal IGH or TCRBeta chain rearrangement

In the fetus, the process of VDJ rearrangement often occurs withoutexonucleotide chewback of VDJ genes and addition of non-templated bases, resulting in a restricted IGH and TCRB CDR3 repertoire that is distinct from the adult repertoire.

These structural differences can be used to distinguish fetal B and T cellCDR3 receptors from maternal B and T cell CDR3 receptors in cell freeDNA present in maternal peripheral blood. In this way, fetal B and Tcell health and development may be monitored in a non-invasivemanner.

Figure 1. Structural differences between fetal and adult B and T cell receptors

Variable Diversity Joining Constant

N1 N2

A. Adult IGH or TCRBeta chain rearrangement

In adult B and T cells, the process of VDJ rearrangement very often involves exonucleotide chewback of VDJ genes and the addition of non-templated bases, forming N1 and N2 regions in the B cell receptorheavy chain CDR3 and the T cell receptor Beta chain CDR3. These processes vastly increase IGH and TCRB CDR3 diversity.

Variable Diversity Joining Constant

B. Fetal IGH or TCRBeta chain rearrangement

In the fetus, the process of VDJ rearrangement often occurs withoutexonucleotide chewback of VDJ genes and addition of non-templated bases, resulting in a restricted IGH and TCRB CDR3 repertoire that is distinct from the adult repertoire.

These structural differences can be used to distinguish fetal B and T cellCDR3 receptors from maternal B and T cell CDR3 receptors in cell freeDNA present in maternal peripheral blood. In this way, fetal B and Tcell health and development may be monitored in a non-invasivemanner.

Figure 1. Structural differences between fetal and adult B and T cell receptors

FR1   FR2  

Diversity  (D)   Joining  (J)   Constant  (C)    

Variable  gene  (V)   CDR3  FR3  

 CDR1  

 CDR2  

Figure 1. Schematic of a section of T-cell receptor which has undergone V(D)J rearrangement, identifying the CDR1, 2, and 3 regions, and the Framework (FR1, 2, and 3) regions. Using RNA/cDNA as input allows design of primers targeting the (V) and (C) regions.

Figure 3. 10 FF tumor infiltrating lymphocyte (TIL) samples (NSCLC) sequenced on 530 chips. We see high concordance between identified clones (95.8-99.6%) comparing sequencing replicates – which indicates sequencing to adequate depth.

The OncomineTM TCR-Seq β assay harnesses the power of AmpliSeqTM library construction chemistry which allows for primer design targeting the FR1 region, and includes primers targeting all known (V) region rearrangements. The resulting sequencing libraries are generally 325-350bp in length. The AmpliSeqTM workflow allows for library construction in under 5 hours with 30 minutes of hands on time.

The OncomineTM TCR-Seq β assay enables sequencing on the Ion S5TM 530 chip. The capacity of this sequencing arrangement allows for multiplexing of up to 16 samples per chip for samples with limited repertoire size (cultured cells, tissue). The ability to multiplex also enables sequencing runs which contain sample comparisons, i.e. – tissue and blood samples run on the same sequencing run (Figures 4-6). Figure 3 above depicts correlation plots generated for technical sequencing replicates for 10 fresh-frozen tumor infiltrating leukocyte (TIL) samples taken from non-small cell lung carcinoma (NSCLC) biopsies. High concordance between identified clones (95.8-99.6%) indicates sequencing to adequate depth.

Immune repertoire sequencing using Ion Torrent sequencing includes a full analysis suite which outputs highly accurate assignment of clonotype populations. Repertoire sequencing is sensitive to base-call (substitution) sequencing errors, which can mimic the natural variation in the repertoire, making error correction strategies very difficult to implement. Ion Torrent sequencing has a very low substitution error rate, and the errors that are produced are predominantly InDel errors. This error type allows for correction using the knowledge that by using mRNA as an input to the assay, all reads should code for productive amino acid sequences, and InDel errors will present as frame shift errors. This permits for the streamlined identification of sequencing errors, and in most cases, error correction, leading to higher rates of productive reads from each sample.

For more information on the OncomineTM TCR-Seq β assay please see poster #3567:

“Sequencing the human TCRβ repertoire on the Ion S5™ System” Session Category: Bioinformatics and Systems Biology

Session Title: Sequencing Analysis and Algorithms Session Date and Time: Tuesday Apr 4, 2017 8:00 AM - 12:00 PM

Location: Convention Center, Halls A-C, Poster Section 23 Poster Board Number: 3

Acknowledgements Figure 5. Spectratype plot of V-gene primer usage versus CDR3 length, showing a comparison of TCRβ clones identified to be unique to the PBL (left panel – 91.78% of all clones identified) and TCRβ clones shared in both the PBL and tumor repertoire (right panel – 8.22% of total). Circle size denotes level of clonal expansion.

Figure 4. Diagram outlining sample comparison between FF tumor infiltrating leukocyte and peripheral blood leukocytes from a patient with NSCLC sequenced on a single 530 chip.

Unique to peripheral blood: 91.78%

Shared with tumor repertoire: 8.22%

CD

R3

leng

th (n

t)

V-gene primer V-gene primer

Figure 6. Correlation plot showing the same dataset shown in Figure 5. Data along the x-axis represents clones unique to the PBL, data along the y-axis shows those unique to tumor, as well as 219 clones that are shared between PBL and tumor.

To test the ability of the assay to differentiate repertoire sample types, total RNA was extracted from peripheral blood leukocytes (PBL) and fresh-frozen tumor biopsy from an individual with Stage 1B squamous cell carcinoma of lung. 100ng of total RNA from each sample was used as template input for TCRβ sequencing on a single Ion 530TM Chip. The results of this study are outlined in Figure 4; sequencing of the tumor biopsy revealed 589 unique TCR clones, an oligoclonal repertoire with a small number of dominant clones, resulting in a Shannon Diversity value of 6.78. Sequencing PBL led to 45305 unique TCR clones. Next, we carried out correlation to determine the populations of clones that were unique to either the tumor biopsy or the PBL, and which clonal populations are shared between sample types. Correlation data is presented using two views shown in Figures 5 and 6. Figure 5 displays a spectratype plot of V-gene primer useage versus CDR3 length, split into two windows: L) clones unique to the PBL and R) clones shared between the PBL and the tumor biopsy. The spectratype plot allows for a fast view of the differences between repertoires unique to PBL versus the shared clone population. Namely, the expected size and diversity of the PBL repertoire compared to the relatively less diverse repertoire of the tumor biopsy which includes fewer, more highly expanded clones. Figure 6 displays the same data which separates into three natural populations; along the x-axis, clones unique to the PBL, along the y-axis, clones unique to the tumor biopsy, and along the diagonal, clones which are shared between sample types. This view allows quick identification of clonal populations which may be interesting routes for further experiments, for instance, tracking of clones from the blood which are identified in the solid tumor, or expansion of tumors which are found only in the tumor sample.

log10 clone frequency in PBL

log 1

0 cl

one

frequ

ency

in tu

mor

370 clones unique to tumor

45086 clones unique to PBL

219 shared clones

The authors would like to acknowledge the work of all who participated in this program: Xinzhan Peng, Alice Zheng, Alex Pankov, Lifeng Lin, Grace Lui, Gauri Ganpule, Sonny Sovan, Larry Fang, Tyler Stine, Laura Nucci, Matthew Cato, Yuan-Chieh (Jeffrey) Ku, Janice Au-Young, Rob Bennett, and Jim Godsey.