VDJ-Seq: Deep Sequencing Analysis of Rearranged Immunoglobulin Heavy Chain Gene to Reveal Clonal Evolution Patterns of B Cell Lymphoma

Summary

This protocol describes an approach to interrogate the recombined immunoglobulin heavy chain VDJ regions of lymphomas by deep-sequencing and retrieve VDJ rearrangement and somatic hypermutation status to delineate clonal architecture of individual tumor. Comparing clonal architecture between paired diagnosis and relapse samples reveals lymphoma relapse clonal evolution modes.

Abstract

Understanding tumor clonality is critical to understanding the mechanisms involved in tumorigenesis and disease progression. In addition, understanding the clonal composition changes that occur within a tumor in response to certain micro-environment or treatments may lead to the design of more sophisticated and effective approaches to eradicate tumor cells. However, tracking tumor clonal sub-populations has been challenging due to the lack of distinguishable markers. To address this problem, a VDJ-seq protocol was created to trace the clonal evolution patterns of diffuse large B cell lymphoma (DLBCL) relapse by exploiting VDJ recombination and somatic hypermutation (SHM), two unique features of B cell lymphomas.

In this protocol, Next-Generation sequencing (NGS) libraries with indexing potential were constructed from amplified rearranged immunoglobulin heavy chain (IgH) VDJ region from pairs of primary diagnosis and relapse DLBCL samples. On average more than half million VDJ sequences per sample were obtained after sequencing, which contain both VDJ rearrangement and SHM information. In addition, customized bioinformatics pipelines were developed to fully utilize sequence information for the characterization of IgH-VDJ repertoire within these samples. Furthermore, the pipeline allows the reconstruction and comparison of the clonal architecture of individual tumors, which enables the examination of the clonal heterogeneity within the diagnosis tumors and deduction of clonal evolution patterns between diagnosis and relapse tumor pairs. When applying this analysis to several diagnosis-relapse pairs, we uncovered key evidence that multiple distinctive tumor evolutionary patterns could lead to DLBCL relapse. Additionally, this approach can be expanded into other clinical aspects, such as identification of minimal residual disease, monitoring relapse progress and treatment response, and investigation of immune repertoires in non-lymphoma contexts.

Introduction

Cancer is a clonal disease. Since thirty years ago when Peter C. Nowell proposed the cancer clonal evolution model¹, many studies have tried to dissect clonal populations within tumor samples and reconstruct clonal expansion and evolution patterns that underlie the tumorigenesis process². Recently, whole-genome sequencing has enabled investigators to take a deep look at the clonal heterogeneity and evolution^3,4. However, due to the lack of tractable markers in many cell types, it is difficult to infer the precise clonal architecture and evolutionary path. Fortunately there is a natural clonality marker in mature B cells from which many lymphoid malignancies, including DLBCL, originate. In response to antigen stimulation, each B-cell can form a single productive IgH VDJ sequence by joining a V_H (variable), a D (diversity), and a J_H (joining) segment together from a large pool of these segments. During this process, small portions of the original sequence may be deleted and additional non-templated nucleotides may be added to create a unique VDJ rearrangement. This specific VDJ rearrangement can be inherited in all the progeny of this B-cell, therefore tagging individual mature B-cell and its offspring⁵. Furthermore, SHM occurs on the recombined VDJ sequences in the subsequent germinal center (GC) reaction to introduce additional mutations for the expansion of the antibody pool and the enhancement of antibody affinity⁶. Therefore, by comparing and contrasting VDJ and SHM patterns of lymphoma samples that have undergone these processes, intra-tumor heterogeneity could be delineated and clonal evolution path of the disease may be deduced.

Previously, VDJ rearrangement and SHM could be identified by PCR amplifying the recombined regions, cloning the PCR products, and subsequently Sanger sequencing to obtain sequence information. This approach is low-throughput and low yield, retrieving only a very small portion of the entire recombined VDJ repertoire, and hindering the characterization of the overall representation of the clonal population within a given sample. A modified approach was created by generating NGS indexed sequencing libraries from VDJ PCR products and performing PE 2x150 bp sequencing to obtain more than half a million recombined VDJ sequences per sample. In addition, a custom pipeline was developed to perform quality control (QC), align, filter VDJ sequencing reads to identify rearrangements and SHMs of each read, and perform phylogenetic analysis on the clonal architecture of each sample. In addition, a new approach has been established to further characterize the clonal evolution patterns for samples collected at various disease stages.

We have applied this technique to DLBCL patient samples. DLBCL is an aggressive form of non-Hodgkin lymphoma with frequent relapse in up to one third of the patients⁷. DLBCL relapses normally occur early, within 2 to 3 years of the initial diagnosis, although some do occur after 5 years⁸. Prognosis for relapsed patients is poor, with only 10% achieving 3 year progression-free survival due to limited treatment options. This is the basis to the urgent need for novel approaches to treat DLBCL relapse^9,10. However, molecular mechanisms associated with DLBCL relapse are still largely unknown. Particularly, the role of clonal heterogeneity at diagnosis and clonal evolution during DLBCL relapse development are presently uncharacterized, making it difficult to define an accurate and useful biomarker to predict relapse. To address these questions, we applied our VDJ-sequencing approach on multiple pairs of matched primary diagnosis-relapse DLBCL sample pairs. Two distinct clonal evolutionary scenarios of relapse emerged from the comparison of the clonal architectures between the diagnosis and relapse samples that suggests multiple molecular mechanisms may be involved in DLBCL relapse.

Protocol

1. VDJ Amplification

1.1) DNA Extraction from Tumor Samples

1. Extract DNA from thin sections (10-20 µm) of frozen O.C.T. embedded normal or malignant tissues.
   1. Digest 10-30 thin sections of embedded tissue cut by a cryostat microtome in 4 ml Nucleic Lysis Buffer (0.0075 M Tris HCl, pH 8.2; 0.3 M NaCl; 0.002 M Na₂EDTA) with Proteinase K (0.5 mg/ml, final concentration) and 0.625% SDS in a 15 ml centrifuge tube in a 37 °C water bath overnight.
   2. Add 1 ml of saturated NaCl (5 M) to the digestion mixture and shake vigorously for 15 sec.
   3. Centrifuge at 1,100 x g for 15 min at room temperature.
   4. Transfer the supernatant to a new 15 ml centrifuge tube and add two volumes of 100% Ethanol.
   5. Mix by inverting the tube 6-8 times. Centrifuge at 5,000 x g for 60 min at 4 °C to collect the precipitated DNA.
   6. Wash the DNA pellet twice with 70% ethanol. Centrifuge at 5,000 x g for 15 min each time to collect the pellet.
   7. Dissolve DNA in 100-400 µl TE buffer at room temperature on a shaker overnight. The DNA yield is 5 to 200 µg (final concentration between 50-500 ng/µl) depending on the size of the tissue
2. Extract DNA from thin sections (10-20 µm) of formalin-fixed, paraffin- embedded (FFPE) normal or malignant tissue.
   1. Incubate paraffin sections in 1 ml xylene twice at room temperature for 10 min each time in a 1.5 ml microcentrifuge tube to de-paraffinize. Collect the tissue sections by spinning at 13,000 x g for 5 min at room temperature.
   2. Incubate sections in 1 ml 100% ethanol twice at room temperature, 10 min each time to remove residue xylenes. Collect the tissue sections by spinning at 13,000 x g for 5 min at room temperature. The paraffin is completely dissolved at this point and only the tissue section is remaining.
   3. Air-dry the sections at room temperature for 10-15 min.
   4. Make a 0.5 mg/ml Proteinase K solution with 1x PCR buffer (diluted from 10x PCR buffer with nuclease-free water). Add the Proteinase K solution to the samples in a final volume of 50-100 µl and incubate overnight at 37 °C.
   5. Heat the samples at 95 °C for 10 min to inactivate Proteinase K.
      Note: At this step, sample DNA is dissolved in the PCR buffer and can be used in Steps 1.2 and 1.3 directly. The DNA yield is 0.5 to 20 µg (final concentration between 10-200 ng/µl) depending on the size of the tissue.

1.2) DNA Quality Assessment

1. Mix 0.25 µl Taq DNA polymerase with 45 µl of master mix from the commercial ladder kit in a PCR tube.
2. Add 5 µl DNA prepared from 1.1.1.7 or 1.1.2.5 into the PCR tube and mix well by pipetting up and down for at least 5 times.
3. Use the following conditions to amplify the DNA: 95 °C for 7 min; followed by 35 cycles of 45 sec at 95 °C, 45 sec at 60 °C, and 90 sec at 72 °C; then 72 °C for 10 min and hold at 15 °C.
4. Prepare a 2% agarose gel in TBE (Tris/Borate/EDTA).
5. Mix 20 µl PCR reaction with 4 µl 6x loading dye, and load onto a 2% agarose gel.
6. Stain the agarose gel with 0.5 µg/ml ethidium bromide solution and detect PCR products with a gel imagining system. Note: the samples that yield 5 PCR products at sizes of 100, 200, 300, 400, and 600 bp will be continued to generate VDJ amplicons.
   CAUTION: Ethidium bromide is a potent mutagen. Please handle with extreme caution by wearing protective gears, i.e. gloves, and dispose into specific containers per institution’s guidelines.

1.3) VDJ PCR

1.3.1) Amplify Recombined IgH VDJ Segment from Framework Region 1 (IgVHFR1)

1. Mix 45 µl master mix from the tube labeled as “Mix 2” of a commercial somatic IGH Hypermutation Assay for Gel Detection kit, 0.25 µl Taq DNA polymerase, and 5 µl sample DNA prepared from 1.1.1.7 or 1.1.2.5 in a PCR tube.
2. Use the following conditions to amplify the DNA: 95 °C for 7 min; followed by 35 cycles of 45 sec at 95 °C, 45 sec at 60 °C, and 90 sec at 72 °C; then 72 °C for 10 min and hold at 15 °C.
3. Resolve the entire PCR product in a 2% agarose gel by electrophoresis.
4. Stain the agarose gel with 0.5 µg/ml ethidium bromide solution and detect PCR products with a gel imaging system. A monoclonal amplicon is expected with the size range of 310-380 bp.
5. Excise the gel portion containing the monoclonal amplicon between 310-380 bp.
6. Purify DNA from excised gel using a standard Gel Extraction Kit according to manufacturer’s protocol. Note: for samples that FR1 fragments could be obtained, there is no need to amply IgVHFR2 and IgVHFR3.

1.3.2) Amplify Recombined IgH VDJ Segment from Framework Region 2 (IgVHFR2)

1. Mix 45 µl master mix from the tube labeled as “Tube B” of a commercial IGH Gene Clonality Assay for Gel Detection kit, 0.25 µl Taq DNA polymerase, and 5 µl sample DNA prepared from 1.1.1.7 or 1.1.2.5 in a PCR tube.
2. Use the following conditions to amplify the DNA: 95 °C for 7 min; followed by 35 cycles of 45 sec at 95 °C, 45 sec at 60 °C, and 90 sec at 72 °C; then 72 °C for 10 min and leave at 15°C.
3. Resolve the entire PCR product in a 2% agarose gel by electrophoresis.
4. Visualize PCR product(s) by 0.5 µg/ml ethidium bromide staining. A monoclonal amplicon may be observed within the size range of 250-295 bp.
5. Excise the gel portion containing the monoclonal amplicon between 250-295 bp.
6. Purify DNA from excised gel using a standard Gel Extraction Kit according to manufacturer’s protocol.

1.4) Optimize VDJ PCR

1. Use the following modified PCR conditions to amplify IgVHFR1 and IgVHFR2 using suboptimal DNA: 95 °C for 7 min, 40 cycles of 95 °C for 60 sec, 60 °C for 60 sec, and 72 °C for 90 sec, final extension at 72 °C for 10 min.

2. VDJ Amplicon Library Preparation and Sequencing

2.1) Library Preparation

2.1.1) End-repair

1. Transfer VDJ PCR product from 1.3 into a PCR tube and add Resuspension Buffer from DNA Sample Preparation Kit to bring up the volume to 60 µl.
2. Add 40 µl of End Repair Mix from DNA Sample Preparation Kit and mix thoroughly.
3. Incubate the reaction at 30 °C for 30 min in a pre-heated thermocycler (30 °C) with a pre-heated lid at 100 °C.
4. Mix 136 µl magnetic beads and 24 µl PCR grade water first in a 1.5 ml tube, then transfer the entire End-repair reaction from 2.1.1.3 into the 1.5 ml tube and mix well with the beads solution.
5. Place tubes on a magnetic stand for 2 min to allow the separation of the beads from the solution. Aspirate the supernatant, and wash the beads with 80% fresh prepared EtOH twice while the tube is on the magnetic stand.
6. Aspirate the ethanol solution completely and allow the beads to air dry for 15 min at room temperature for 15 min.
7. Take tube off the magnetic stand and resuspend beads in 17.5 µl Resuspension Buffer.
8. Place tubes back onto the magnetic stand for 2 min to separate beads from the Resuspension Buffer. Remove the Resuspension Buffer (End-repair product is resuspended in the Resuspension buffer now) into a clean PCR tube.

2.1.2) A-tailing

1. Add 12.5 µl A-tailing mix into the PCR tube containing End-repair product and mix thoroughly.
2. Incubate the PCR tube at 37 °C for 30 min in a pre-heated thermocycler (37 °C) with a pre-heated lid at 100 °C.

2.1.3) Adaptor Ligation

1. Add 2.5 µl Resuspension Buffer, 2.5 µl Ligation Mix, and 2.5 µl DNA Adaptor Index into the A-tailing reaction.
2. Incubate the reaction at 30 °C for 30 minl in a pre-heated thermocycler (30 °C) with a pre-heated lid at 100 °C.
3. Add 5 µl Stop Ligation Buffer into each reaction and mix thoroughly.
4. Mix 42.5 µl well-mixed magnetic beads to clean the reaction following steps 2.1.1.5 to 2.1.1.8. Add 50 µl Resuspension Buffer to elute the adaptor-ligated product.
5. Clean the adaptor-ligated product for a second time by using 50 µl well-mixed AMPure XP beads, and elute the product in 25 µl Resuspension Buffer into a clean PCR tube.

2.1.4) Amplify DNA Fragments

1. Add 5 µl PCR Primer Cocktail and 25 µl PCR Master Mix to the PCR tube containing adaptor-ligated product and mix thoroughly.
2. Perform amplification in a pre-programmed thermocycler with the following conditions: 98 °C for 30 sec; 10 cycles of 98 °C for 10 sec, 60 °C for 30 sec and 72 °C for 30 sec; then 72 °C for 5 min and hold at 10 °C.
3. Clean up the reaction by using 50 µl well-mixed magnetic beads, and elute the final product in 30 µl Resuspension Buffer.

2.1.5) Library Validation

1. Quantify the final product with a fluorometer using manufacturer’s protocol. Final library concentration is between 2.5 to 20 ng/µl.
2. Assess the quality of the final product by using an analytical instrument with High Sensitivity DNA chip according to manufacturer’s protocol. Expect a single band with the expected size for either IGVHFR1, IGVHFR2 or IGVHFR3. The expected size of the library product equals the size of the original VDJ amplicon plus the size of the adaptors (~125 bp).

2.2) VDJ-PCR Library Pooling and Sequencing

1. Calculate the molarity of each library fraction using the following formula: nM = [(ng/1,000)/bp*660] x 10⁹ where ng is the concentration of the VDJ-PCR library measured in step 2.1.5.1 and bp is the peak size of the VDJ-PCR library measured in step 2.1.5.2.
2. Dilute the library to 2 nM (10 µl total) with DNase-free water.
3. Combine the 10 µl of the diluted 2 nM libraries together into a 1.5 ml tube.
4. Add an equal volume of PhiX spike-in control into the 1.5 ml tube.
5. Load the pool at 7 pM concentration onto a flowcell.
6. Sequence the libraries using a paired-end 150 cycle sequencer according to manufacturer’s protocol.

3. Data Analysis

Note: A summary of the bioinformatics scripts used in this section can be found as a Supplementary Code File.

3.1) Alignment and QC

1. Map paired-end sequencing reads against a human IGH V, D and J region database downloaded from the IMGT website¹¹ using an adapted nucleotide blast algorithm based upon ‘blastn’ available from NCBI with gap open penalty of 2, gap extension penalty of 1, word length of 7, and e-value threshold of 10^-4. Discard the read-pair does not map to both an IGH V and a J region.
2. Count the remaining read-pairs that have IGH V and J mappings to obtain the frequencies of matching VJ combinations across all read pairs obtained from the sequencing run on the sample. Count a read-pair if it is mapped to both an IGH V and J gene, and then add its allele to the corresponding count for the particular VJ combination.
3. Rank the counts for each recombined VDJ region obtained from 3.1.2 from the highest to the lowest. The VDJ region has the highest reads count is defined as major rearrangement combination.
4. Discard aligned sequences that cover less than 35% of the domain major rearrangement identified in 3.1.3.

3.2) SHM Profile Identification

1. Count all SHM patterns across the reads from step 3.1.3. Define each SHM pattern as a subclone.
2. Count the number of the subclones that are corresponding to different unique SHM patterns, and the number of reads that are mapped to individual subclones.

3.3) Graphical Representation of Results

1. Perform the phylogenetic analysis on the subclones using their corresponding SHM patterns using a neighborhood joining method from R package “ape” according to manufacturer’s protocol. Calculate a distance matrix from the individual distinct alignments to recreate the subclone phylogeny rooted to the germline sequence of the VJ combination in question for each paired sample set.
2. Use the nucleotide sequence of each subclone as character vector and calculate the approximate string distance between all subclones in the major VJ rearrangement for both diagnosis and corresponding relapse samples using a Levenshtein distance measure where mathematically the distance between two alignments is given by d_x,y(i,j) where d_x,y(i,j) = 1 if i ≠ j and 0 otherwise, and then sum this function over the length of the string corresponding to nucleotide sequences i and j.
3. Graphically display the resulting matrix of subclone distances and their corresponding subclone counts in the following two ways:
   1. Use R package “MASS” according to manufacturer’s protocol to apply multidimensional scaling to the subclone distance matrix and generate a two dimensional principle coordinates map, which is then plotted along with the logarithm of subclonal counts as the radius of the circle.
   2. Construct an undirected graph based upon the Levenshtein distance, where the vertices correspond to each distinct subclone arising from the major VJ rearrangement.
      Note: If the distance between two distinct clones is equal to one then there exists an edge between these two vertices. If the distance is greater than one, then there is no edge connecting them.
   3. Plot the graph of 3.3.3.2 using the tkplot function in the “iGraph” R package with the Kamada Kawai layout according to manufacturer’s protocol. The radius of the vertices is proportional to the cubed root of the total number of clones mapped to it and the shading uses a red and blue range to denote the proportion of clones that either belong to the diagnosis (blue) or relapse (red) samples.

3.4) Heterogeneity (Entropy) Measurement

1. Examine the numbers and subclone counts of the individual patterns of somatic hypermutation occurring in the major VJ rearrangement for each paired set of samples.
2. Calculate the empirical entropy and residual empirical entropy adjusted for the number of distinct clones in each sample using the standard histogram based entropy estimation.

Results

The overall procedure of VDJ sequencing (VDJ-seq), including DNA extraction, recombined VDJ region amplification and purification, sequencing library construction, reads processing, and phylogenetic analysis, is represented in Figure 1. Routinely 5-200 μg DNA can be retrieved from frozen solid tissue sections or 0.5-20 μg DNA from formalin-fixed paraffin-embedded tissue sections. Depending on the quality, rearrangement pattern, and SHM degree of individual samples, a variety of VDJ PCR products from diff...

Discussion

Because of the nearly unlimited number of iterations of sequence information coded by VDJ rearrangement and SHM at the IGH locus of human B cells, examination of the entire IGH repertoire by high-throughput deep-sequencing proved to be an efficient and comprehensive way to delineate clonal and sub-clonal B-cell populations. Furthermore, this strategy can be used to study the clonal evolution path of B cell tumor development, remission, and relapse by comparing the clonal and sub-clonal architectures of patient samples co...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Ethanol	VWR	89125-170	200 proof, for molecular biology
TE buffer	Life Technologies	12090-015	10 mM Tris·Cl, pH 8.0; 10 mM EDTA
Xylenes	VWR	EM-XX0055-6	500 ml
Proteinase K	Life Technonogies	25530-015	100 mg
Deoxynucleotide triphosphate (dNTP) Solution Mix	Promega	U1515	10 mM each nucleotide
10x PCR Buffer	Roche	11699105001	Without MgCl2
AmpliTaq Gold DNA Polymerase with Gold Buffer and MgCl2	Life Technonogies	4311806	50 µl at 5 U/µl
Specimen Control Size Ladder	Invivoscribe Technologies	2-096-0020	33 reactions
IGH Somatic Hypermutation Assay v2.0 - Gel Detection	Invivoscribe Technologies	5-101-0030	33 reactions, Mix 2 for IGVHFR1 detection
IGH Gene Clonality Assay - Gel Detection	Invivoscribe Technologies	1-101-0020	33 reactions, Tube B for IGVHFR2 detection
GoTaq Flexi DNA Polymerase	Promega	M8291	20 µl at 5 U/µl
10X Tris-Borate-EDTA (TBE) buffer	Corning (cellgro)	46-011-CM	6x1 L
50x TAE BUFFER	VWR	101414-298	1 L
Ethidium bromide solution	Sigma-Aldrich	E1510	10 mg/ml
25 bp DNA Ladder	Life Technologies	10597011	1 µg/µl
100 bp DNA Ladder	Life Technologies	15628-019	1 µg/µl
UltraPure Agarose	Life Technologies	16500-500	500 g
40% Acrylamide/Bis Solution	Bio-Rad Laboratories	1610144	500 ml
QIAquick Gel Extraction Kit	Qiagen	28704	50 columns
Agencourt AMPure XP	Beckman Coulter	A63881
Qubit dsDNA High Sensitivity Assay Kit	Life Technologies	Q32854
High Sensitivity DNA Kit	Agilent Technologies	5067-4626
2100 Bioanalyzer	Agilent Technologies
PhiX Control v3	Illumina	FC-110-3001
MiSeq	Illumina
Qubit 2.0 Fluorometer	Life Technologies	Q32872
Resuspension buffer (Illumina TruSeq DNA Sample Preparation Kit v2)	Illumina	FC-121-2001
End repair mix (Illumina TruSeq DNA Sample Preparation Kit v2)	Illumina	FC-121-2001
A-tailing mix (Illumina TruSeq DNA Sample Preparation Kit v2)	Illumina	FC-121-2001
Ligation mix (Illumina TruSeq DNA Sample Preparation Kit v2)	Illumina	FC-121-2001
DNA adaptor index (Illumina TruSeq DNA Sample Preparation Kit v2)	Illumina	FC-121-2001
Stop ligation buffer (Illumina TruSeq DNA Sample Preparation Kit v2)	Illumina	FC-121-2001
PCR primer cocktail (Illumina TruSeq DNA Sample Preparation Kit v2)	Illumina	FC-121-2001
PCR master mix (Illumina TruSeq DNA Sample Preparation Kit v2)	Illumina	FC-121-2001
Magnetic stand	Life Technologies	4457858
Gel imaging system	Bio-Rad Laboratories	170-8370