trbv polymorphism predicts adverse events during...

Thermo Fisher Scientific • 5791 Van Allen Way • Carlsbad, CA 92008 • www.thermofisher.com For Research Use Only. Not for use in diagnostic procedures. The content provided herein may relate to products that have not been officially released and is subject to change without notice.

T Looney1, E Linch2 D Topacio-Hall2, G Lowman2, J Conroy3, C Morrison3, F Hyland1 (1) Thermo Fisher Scientific, South San Francisco CA (2) Thermo Fisher Scientific, Carlsbad CA (3) OmniSeq, Inc, Buffalo, NY.

TRBV polymorphism predicts adverse events during checkpoint blockade immunotherapy

METHODS

Figure 5: Sample annotations and repertoire features for study cohort consisting of 55 Caucasians who developed adverse events grades 1-4 following CPI monotherapy. Samples derive from the Roswell Park Cancer Research Institute. Libraries were prepared and sequenced by OmniSeq using 25ng cDNA derived from peripheral blood leukocyte total RNA.

ABSTRACT Identifying predictive biomarkers for immune related adverse events (IRAEs) during checkpoint blockade immunotherapy (CPI) is a key objective of current immuno-oncology research. Polymorphism within the TCRB variable gene (TRBV) has been implicated in autoimmune disease and may be mechanistically linked to IRAEs [1]. Efforts to evaluate TRBV polymorphism by traditional approaches such as whole genome sequencing (WGS) have been hampered by the repetitive nature of the TCRB locus and incomplete genome assembly. Here we present a novel long-amplicon TCRB repertoire sequencing approach to evaluate the link between TRBV polymorphism and adverse events in 55 Caucasians receiving CPI for cancer. Three lines of reasoning support the notion that germline encoded TRBV polymorphism could be a key determinant of IRAEs. First, the TCR locus is repetitive and structurally complex, impeding the measurement of variation by traditional short read WGS or microarray based methods; second, single amino acid substitutions within the framework or CDR 1 and 2 regions of the rearranged TCRB chain are know to significantly alter TCR affinity for HLA; and third, adverse events during CPI may manifest as acute versions of chronic autoimmune diseases that have been separately linked to TRBV polymorphism. Identifying predictive biomarkers for IRAEs may be critical for combination and neoadjuvant use of CPI for cancer.

•  Long amplicon TCRB repertoire sequencing may be used to perform haplotype analysis of the repetitive and structurally complex TRB locus.

•  We detected four major haplotype groups in a cohort consisting of 55 Caucasians.

•  Members of haplotype group 2 showed no severe events during immunotherapy.

•  Expanding the cohort size would reveal additional haplotypes and improve prediction of adverse events.

•  TRBV polymorphism is germline encoded and therefore may serve as a true predictive biomarker for IRAEs following CPI for cancer.

Figure 4. Cartoon of the TRB locus for two distinct haplotypes. The TRB locus contains tandemly arranged polymorphic variable genes. The Oncomine TCRB-LR assay allow one to detect all of the variable gene alleles present in a repertoire. This data can be used to infer the haplotype of the TRB locus.

Figure 6. Heatmap of TRBV allele profiles for study cohort. TCRB repertoires were used to construct variable gene allele profiles for each individual. The sets of alleles detected in each individual are displayed in heatmap form, where each row represents a different individual and each column a different variable gene allele. Red tiles indicate that an allele was detected in an individual while blue tile indicate allele absence. Columns are arranged via hierarchical clustering, while rows are arranged according to haplotype group classification produced by k-means clustering.

Figure 3. Strategy for identification of non-IMGT variable gene alleles. Bone fide novel alleles will present as systematic mismatches to IMGT across a plurality of clones, each possessing a distinct CDR3 nucleotide sequence.

RESULTS RESULTS

Figure 8. Principal component analysis of TRBV allele profiles was used to subdivide samples into four haplotype groups. Each point represents a different individual, while symbol indicates grade of adverse event. Haplotype group 2, comprising ~33% of this cohort, appears to be protected against severe adverse events following CPI.

CDR loops

T Cell Receptor

HLA presenting peptide

Workflow

Figure 2 Overview of Workflow. The Oncomine TCRB-LR assay utilizes AmpliSeq multiplex PCR primers to target the TCRβ Framework 1 and Constant regions, enabling clonotyping and detection of TRBV gene polymorphism linked to adverse events during immunotherapy.

Introduction

Abstract Methods Results Results

Conclusions

References

1.  Looney et al. Haplotype Analysis of the TRB Locus by TCRB Repertoire Sequencing (2018). bioRxiv 406157

2.  Ye et al. IgBLAST: an immunoglobulin variable domain sequence analysis tool (2013). Nucleic Acid Research W34-40.

Figure 1. Percentage of Melanoma Subjects Having Severe (Grade 3 or higher) Adverse Events. Adaptive from Larkin et al. 2015

1 For Research Use Only. Not for use in diagnostic procedures.

Figure 1A and B | Methods Overview and Amplification Strategy

Constant

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.


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

RNA/cDNA Variable Diversity Joining Constant

N1 N2








Leader FR1 FR2

Diversity(D) Joining (J) Variable (V)

CDR3

FR3 CDR1

CDR2

AmpliSeq Primers ~325-400 bp Non-FFPE RNA

Allelic variants

Primer Digestion

Ligation of bidirectional sequencing adaptors

FR1-C multiplex PCR from 25ng PBL cDNA

NGS library templating and sequencing

Annotate rearrangements by alignment to IMGT

Identify and eliminate indel and PCR errors

Report VDJ rearrangements

Identify non-IMGT alleles and construct allele profiles

K-means clustering to identify haplotype groups

Associate TRBV haplotypes with phenotypes of interest

1A. 1B.


Figure 1A and B | Methods Overview and Amplification Strategy

Constant


N1 N2








RNA/cDNA Variable Diversity Joining Constant

N1 N2








Leader FR1 FR2

Diversity(D) Joining (J) Variable (V)

CDR3

FR3 CDR1

CDR2

AmpliSeq Primers ~325-400 bp Non-FFPE RNA

Allelic variants

Primer Digestion

Ligation of bidirectional sequencing adaptors

FR1-C multiplex PCR from 25ng PBL cDNA

NGS library templating and sequencing

Annotate rearrangements by alignment to IMGT

Identify and eliminate indel and PCR errors

Report VDJ rearrangements

Identify non-IMGT alleles and construct allele profiles

K-means clustering to identify haplotype groups

Associate TRBV haplotypes with phenotypes of interest

1A. 1B.


Figure 1E | Strategy for Detection of non-IMGT Alleles

Variable CDR3

Clones having canonical IMGT variable gene

Clones having allele absent from IMGT database

Systematic mismatch to reference

1E.

Figure 7. Example of a non-synonymous IMGT variant. IgBLAST alignment of an allele having two amino acid substitutions compared to the best matching IMGT allele. This particular allele was detected in our samples and the Lym1K database derived from 1000 genomes data. Detection of non-IMGT alleles improves the resolution of uncommon haplotypes. 3 For Research Use Only. Not for use in diagnostic procedures.

Ipi+Nivo

Nivo (PD-1)

Ipi (CTLA-4)

Ipi+Nivo

Nivo (PD-1)

Ipi (CTLA-4)

0 10 20 30 40 50

Immune Mediated Adverse Events are a Key Obstacle for Combination Immunotherapy

Percentage of Melanoma Subjects Having Severe (Grade 3 or 4) Adverse Events

Ipilimumab (CTLA-4)

Nivolumab (PD-1)

Ipilimumab + Nivolumab 56%

16%

27%

Adapted from Larkin et al 2015

Common Adverse Events by Organ System

Predicting adverse events could enable: •  Personalized drug selection and dosing •  Expanded use of combination therapies !Safer and more effective immunotherapy

BRAIN: Encephalopathy, aseptic meningitis, parathesias, weakness

GLANDS: Sicca syndrome

HEART: Myocarditis

COLON: Diarrhea, colitis, perforation, megacolon

BLOOD VESSELS: Vasculitis

PITUITARY GLAND: Hypophysitis

THYROID: Thyroiditis

LUNGS: Pneumonitis

KIDNEYS: Lupus nephritis, acute interstitial nephritis

LIVER: Hepatitis

MUSCLES: Myositis

JOINTS: Inflammatory arthritis


Ipi+Nivo

Nivo (PD-1)

Ipi (CTLA-4)

Ipi+Nivo

Nivo (PD-1)

Ipi (CTLA-4)

0 10 20 30 40 50

Immune Mediated Adverse Events are a Key Obstacle for Combination Immunotherapy

Percentage of Melanoma Subjects Having Severe (Grade 3 or 4) Adverse Events

Ipilimumab (CTLA-4)

Nivolumab (PD-1)

Ipilimumab + Nivolumab 56%

16%

27%

Adapted from Larkin et al 2015

Common Adverse Events by Organ System

Predicting adverse events could enable: •  Personalized drug selection and dosing •  Expanded use of combination therapies !Safer and more effective immunotherapy

BRAIN: Encephalopathy, aseptic meningitis, parathesias, weakness

GLANDS: Sicca syndrome

HEART: Myocarditis

COLON: Diarrhea, colitis, perforation, megacolon

BLOOD VESSELS: Vasculitis

PITUITARY GLAND: Hypophysitis

THYROID: Thyroiditis

LUNGS: Pneumonitis

KIDNEYS: Lupus nephritis, acute interstitial nephritis

LIVER: Hepatitis

MUSCLES: Myositis

JOINTS: Inflammatory arthritis


•  Heatmap highlighting alleles detected in 55 samples. Column indicates allele; red tiles indicate allele present in a particular sample.

•  Grouping of alleles into blocks indicates presence of four major allele haplotypes.

•  Haplotypes represent sets of alleles that tend to be inherited together owing to genetic linkage and population structure.

TRBV Allele Typing Reveals Four Major Haplotype Groups in 55 Sample Cohort

Alleles

TRBV7.7.01p85A

TRBV6.6.01p51T

TRBV5.6.01p86G

TRBV5.4.01p101A

TRBV5.3.01

TRBV5.1.01p85A

TRBV16.01p152A

TRBV28.01p15G

TRBV5.6.01p36T67A

TRBV15.01

TRBV18.01p200G

TRBV12.3.01p67T

TRBV5.6.01p28G

TRBV5.1.01p0T

TRBV30.02

TRBV20.1.05

TRBV10.3.03

TRBV11.3.01p169G

TRBV12.4.01p189T

TRBV14.02

TRBV12.5.01p157A

TRBV7.8.01p231G

TRBV5.8.01p36T217T

TRBV6.9.01p9G

TRBV9.02

TRBV6.4.02

TRBV7.3.03

TRBV10.1.02

TRBV10.2.01p81C

TRBV11.2.03

TRBV5.5.01

TRBV6.6.02

TRBV7.4.01p61A

TRBV7.8.02

TRBV7.2.01p15A

TRBV6.5.01p33A

TRBV7.9.03

TRBV30.01p14A

TRBV3.1.01p229C

TRBV23.1.01p8C

TRBV4.3.01p0T

TRBV12.4.01p49A

TRBV12.4.01p112G

TRBV11.1.01p133G

177T190GTRBV5.3.01p148T

TRBV10.3.02

TRBV10.1.01

TRBV7.2.02

TRBV24.1.01p142A

TRBV6.7.01

TRBV7.4.01

TRBV10.3.01

TRBV20.1.02

TRBV2.01p144A

TRBV30.01

TRBV5.5.02

TRBV11.2.01

TRBV6.6.01

TRBV6.4.01

TRBV7.3.01

TRBV16.01

TRBV7.9.01

TRBV7.8.01

TRBV7.7.01

TRBV7.6.01

TRBV6.5.01

TRBV6.2.01

TRBV6.1.01

TRBV5.8.01

TRBV5.6.01

TRBV5.4.01

TRBV5.1.01

TRBV4.2.01

TRBV4.1.01

TRBV3.1.01

TRBV29.1.01

TRBV28.01

TRBV27.01

TRBV25.1.01

TRBV23.1.01

TRBV20.1.01

TRBV2.01

TRBV19.01

TRBV14.01

TRBV13.01

TRBV12.5.01

TRBV12.4.01

TRBV12.3.01

TRBV11.1.01

TRBV11.3.01

TRBV15.02

TRBV18.01

TRBV10.2.01

TRBV9.01

TRBV24.1.01

TRBV4.3.01

TRBV7.2.01

TRBV7.3.02

TRBV6.8.01

TRBV6.9.01

1215323538424547485359616311161920212324252628343637394041444649511718305657606291013142227293133435052545558

0

0.2

0.4

0.6

0.8

1

Haplotype Group


Cohort from Roswell Park Cancer Research Institute Sequenced by OmniSeq Category) Subdefini(on) Overall) Grade)1A2)AE) Grade)3A4)AE) PAValue)!! Number!of!Individuals! 55! 44! 11! NA!Age) !!!! Median!(range)! 65!(34<84)! 64!(42<84)! 67!(34<77)! 0.97!Sex) !!!! M! 34! 27! 7! 1!!! F! 21! 17! 4!Ethnicity) !!!! Caucasian! 54! 44! 10! 0.20!!! Unknown! 1! 0! 1!Cancer) !!!! Melanoma! 26! 19! 7!

0.73!

!! Adenocarcinoma! 15! 12! 3!!! Urothelial!Carcinoma! 5! 4! 1!!! Renal!Carcinoma! 1! 1! 0!!! Squamous!cell!carcinoma! 7! 7! 0!!! Unknown! 1! 1! 0!Treatment) !!!! Ipilimumab!(CTLA<4)! 25! 18! 7!

0.18!!! Nivolumab!(PD<1)! 18! 17! 1!!! Pembrolizumab!(PD<1)! 12! 9! 3!Response) !!)) PD! 14! 11! 3!

0.77!)) SD! 9! 8! 1!)) PR! 7! 7! 0!)) CR! 3! 3! 0!)) Unknown! 22! 15! 7!Repertoire)Features) !!

!!Reported!reads!per!sample!(thousands)! 552!(94<1718)!

567!(94<1247)! 533!(159<1718)! 0.65!

!! Clones!Detected!(thousands)! 32!(5<70)! 32!(5<70)! 30!(14<62)! 0.64!!! Clone!Size!Evenness! .86!(.46<.96)! .84!(.56<.96)! .88!(.46<.94)! 0.87!

•  Cohort of 55 individuals from Roswell Park Cancer Research Institute who developed adverse events (grades 1-4).

•  Various cancer types •  Patients treated with 1 of 3 immune

checkpoint blockade inhibitors •  Response annotations for 44

individuals

•  Sequencing performed through a third party (OmniSeq)

•  Sample type = pre-treatment blood (PBL)


•  Principal component analysis of sample TRBV allele types (n=55) followed by k-means clustering was used to subdivide samples into four haplotype groups.

•  The incidence of severe (grade 3 or 4) adverse events varies markedly between haplotype groups.

•  Haplotype Group 2 appears to be protected from severe adverse events.

Assay Identifies TRBV Haplotypes Associated with Adverse Events

Haplotype Group

Grade 1 or 2 AE

Grade 3 or 4 AE

1 9 4 2 20 0 3 3 4 4 12 3 p = .0024, Fisher’s Exact Test

-2 -1 0 1

-2-1

01

Clustering of subjectsbased on TRBV allele haplotype

PC1 of TRBV haplotype

PC

2 of

TR

BV

hap

loty

pe

Grade1234

3/15 (20%)

4/7 (57%)

4/13 (31%)

0/20 (0%)

Clustering of subjects by TRBV allele haplotype


•  Principal component analysis of sample TRBV allele types (n=55) followed by k-means clustering was used to subdivide samples into four haplotype groups.

•  The incidence of severe (grade 3 or 4) adverse events varies markedly between haplotype groups.

•  Haplotype Group 2 appears to be protected from severe adverse events.

Assay Identifies TRBV Haplotypes Associated with Adverse Events

Haplotype Group

Grade 1 or 2 AE

Grade 3 or 4 AE

1 9 4 2 20 0 3 3 4 4 12 3 p = .0024, Fisher’s Exact Test

-2 -1 0 1

-2-1

01

Clustering of subjectsbased on TRBV allele haplotype

PC1 of TRBV haplotype

PC

2 of

TR

BV

hap

loty

pe

Grade1234

3/15 (20%)

4/7 (57%)

4/13 (31%)

0/20 (0%)

Clustering of subjects by TRBV allele haplotype


For Research Use Only. Not for use in diagnostic procedures.

•  A haplotype is a set of DNA variations (i.e. polymorphisms) that tend to be co-inherited. •  Owing to human population structure and the small genetic distance between TRBV genes, we

expect to observe groups of co-inherited variable gene alleles.

The TRB Locus Contains Tandemly-Arranged Polymorphic Genes

TRB locus for two distinct haplotypes Color indicates variable gene allele

Haplotype 1

Haplotype 2

200 kb


For Research Use Only. Not for use in diagnostic procedures.

•  A haplotype is a set of DNA variations (i.e. polymorphisms) that tend to be co-inherited. •  Owing to human population structure and the small genetic distance between TRBV genes, we

expect to observe groups of co-inherited variable gene alleles.

The TRB Locus Contains Tandemly-Arranged Polymorphic Genes

TRB locus for two distinct haplotypes Color indicates variable gene allele

Haplotype 1

Haplotype 2

200 kb

PC1

PC

2

Predictive biomarkers for IRAEs could enable:

trbv polymorphism predicts adverse events during...

Documents