Experimental Hematology 2009;37:728–738

Comprehensive assessment of the TCRBV repertoire in small T-cellsamples by means of an improved and convenient multiplex PCR method

Sara Mariania,b, Seung Yae Hwanga,b, Myriam Fogliettaa,b, Lisa Bonelloc,Candida Vitalea, Marta Cosciaa,b, Francesca Fiorea,b, Benedetto Brunoa, and Massimo Massaiaa,b

aDivisione di Ematologia dell’Universita di Torino, Torino, Italy; bLaboratorio di Ematologia Oncologia, Centro di Ricerca in

Medicina Sperimentale (CeRMS), Ospedale San Giovanni Battista, Torino, Italy; cDipartimento di Scienze Biomediche edOncologia Umana, Universita di Torino, Centro di Ricerca in Medicina Sperimentale (CeRMS), Ospedale San Giovanni Battista, Torino, Italy

(Received 17 January 2009; revised 6 March 2009; accepted 9 March 2009)

Offprint requests t

dell’Universita di To

francesca.fiore@unito

0301-472X/09 $–see

doi: 10.1016/j.exph

Objective. Overall diversity of the T-cell receptor (TCR) repertoire can be regarded as a reca-pitulatory signature of a host’s immunocompetence status. We aimed to establish a time- andcost-saving multiplex polymerase chain reaction (PCR) method for determining the TCRrepertoire of conventional ab T cells in small T-cell samples.

Materials and Methods. The method estimates the length distribution of the complementarity-determining regions 3 (CDR3) of b variable (BV) gene segments (TCRBV repertoire) bymultiplex PCR, followed by fluorescent run-off reactions to visualize BV-BC and/or BV-BJ re-arrangements. Run-off products are separated on a capillary sequencer and subsequentlyanalyzed with GeneScan or Genotyper programs. Detection-limit studies with normal T cells,KMS27 cells, and regulatory T cells were carried out to evaluate sensitivity andreproducibility.

Results. Head-to-head comparison of the method with conventional immunoscope assay hasshown that it is a time- and cost-saving approach to characterize TCRBV and TCRBJ reper-toires, including the presence of oligoclonal T cells in samples containing as few as 1 3 105 Tcells.

Conclusion. We have developed a multiplex PCR method that allows comprehensive assess-ment of the TCRBV repertoire at the BV-BC and BV-BJ levels, and saves a considerableamount of time, reagents, and cell input. � 2009 ISEH - Society for Hematology andStem Cells. Published by Elsevier Inc.

Conventional ab T cells are strictly dependent on T-cellreceptor (TCR) diversity to exert their helper and effectorfunctions, given their need to clonotypically recognize anti-gens in association with major histocompatibility complexmolecules on antigen-presenting cells, target cells, andpathogens [1]. The naıve TCR repertoire is initially assem-bled in the thymus and shaped over time by transient expan-sion, long-term accumulation, or deletion of clonotypic Tcells elicited by pathogens, autoimmune events, or tumorcells, and by numerical variations of polyclonal T cells inresponse to homeostatic mechanisms [2–5]. As a result,the overall degree of TCR diversity at any given time canprovide a recapitulatory record of host immune

o: Francesca Fiore, M.D., Divisione di Ematologia

rino, Via Genova 3, 10126 Torino, Italy; E-mail:

.it

front matter. Copyright � 2009 ISEH - Society for Hemat

em.2009.03.003

competence, including ongoing immune reactions andcumulative effects of previous antigen exposures.

The highest degree of TCR diversity is reached by thecomplementarity-determining regions 3 (CDR3) of theb variable (BV) chain owing to the availability of D genesegments, reading frame infidelity, and transferase-mediatedmodifications at different V-D or D-J junctions (TCRBVrepertoire). Given its higher degree of combinatorial diver-sity, the TCRBV repertoire is the most sensitive for fine char-acterization of TCR diversity during immune responses [2].

Multiparameter flow cytometry is often used to measurethe TCRBV repertoire at the membrane protein level [6],but this approach does not provide any information aboutT-cell clonality. Polymerase chain reaction (PCR)–basedapproaches, like the spectratyping [7] or immunoscope[8] methods, explore the CDR3 length distribution profileof each TCRBV subfamily and reveal the relative frequencyof T cells using any possible functional CDR3 transcript.

ology and Stem Cells. Published by Elsevier Inc.

729S. Mariani et al. / Experimental Hematology 2009;37:728–738

The immunoscope assay, when applied at the BV-BJ level,also predicts whether oligoclonal T-cell expansionscontribute to the emergence of altered CDR3 length profiles[8]. PCR-based approaches have been further improved bythe introduction of multiplex PCR methods aimed at dimin-ishing the number of reactions in favor of time, cost, andcell input without impairing specificity [9–12].

Here, we describe a novel time- and cost-saving multi-plex PCR method for determination of TCRBV and TCRBJrepertoires in normal and pathological samples containingas few as 1� 105 ab T cells.

Table 1. Grouping and concentration of primers used in the multiplex

polymerase chain reaction analysis

PCR tubes Primers*

Final

concentrations (mM)

Set 1y BV1 0.1

BV2 0.1

BV3 0.2

BV4 0.4

BV5 0.4

Set 2y BV6 0.1

BV7a 0.2

BV7b 0.2

BV8 0.3

BV9 0.05

Set 3y BV11 1.5

BV12a 1.5

BV12b 1.5

BV13 0.4

BV13.5 0.4

BV14 0.01

BV15 0.2

Set 4y BV16 1.5

BV17 0.1

BV18 0.4

BV20 0.05

Set 5y BV21 0.3

BV22 1.5

BV23 1.5

BV24 2.0

All BC1z 0.5

PCR 5 polymerase chain reaction.

*TCRBV numbering is derived from Wei et al. [data from 28].ySequence from Akatsuka et al. [data from 11].zSequence from Pannetier et al. [data from 8].

Materials and methods

Cell samplesTwenty milliliters freshly heparinized peripheral blood weredrawn from healthy donors (n 5 12) or multiple myeloma (MM)patients (n 5 8), and peripheral blood mononuclear cells(PBMC) were separated by density gradient centrifugation (Fi-coll-Hypaque; Invitrogen Life Technologies, Paisley, Scotland,UK). Informed consent was obtained upon enrollment in clinicalprograms approved by the Piemonte Institutional Review Board.Cell viability was assessed by trypan blue dye exclusion test.Clonally TCR-positive Jurkat acute lymphoid leukemia T-cellline and TCR-negative MM-derived KMS27 cell line were usedas positive and negative controls, respectively.

Regulatory T cells (Tregs) were purified from 10 PBMCsamples with a pan-T isolation kit or double-step immunomag-netic cell sorting according to manufacturer’s instructions (Milte-nyi Biotec, Bergisch-Gladbach, Germany). Purity was checked byflow cytometry using the appropriate monoclonal antibodies andanalyzed with CellQuest software (Becton Dickinson, MountainView, CA, USA).

RNA extraction, complementaryDNA synthesis, and quality controlsTotal RNA was extracted from PBMC, T cells, and Tregs, or fromthe Jurkat and KMS27 cell lines after lysis in Trizol Reagent (In-vitrogen Life Technologies), and processed as reported previously[13]. Sample cell counts ranged from 0.16� 106 to 7� 106. RNAsamples of 2.5 mg (or less for suboptimal extractions) were reversetranscribed at 42�C with the Reverse Transcription System (Prom-ega, Madison, WI, USA), according to manufacturer’s instruc-tions. Complementary DNA (cDNA) quality controls werecarried out by PCR analysis of b2-microglobulin transcripts usingthe following 50-CTCGCGCTACTCTCTCTTTCTGG-30 and 30-CTTACATGTCTCGATCCCACTTAA-50 as forward and reverseprimers, respectively [14].

The amount of T-cell–derived cDNA per sample was deter-mined by amplification of a BC-specific sequence of 306 basepairs (bp) [14] and comparison with the band intensity of a -predetermined reference control sample (see also Results: Speci-ficity and sensitivity of the multiplex PCR method).

Multiplex BV-BC amplificationFrom four to seven BV-specific primers and one consensus BCprimer were multiplexed in five PCR reactions, as reported inTable 1. The method described by Akatsuka et al. [11] was modi-

fied as follows: BV25-specific primer was excluded from set 5;concentrations of BV-specific primers were modified to allowa balanced amplification of all BV subfamilies, irrespective oftheir relative frequency; BC-consensus primer (defined as BC1)was derived from Pannetier et al. [8] (Table 1).

The PCR mixture contained 1� Buffer (Promega), 4 mMMgCl2 (Promega), 200 mM deoxyribonucleotide triphosphates(Pharmacia, Uppsala, Sweden), 0.5 mM reverse consensus BC1primer (MWG-Biotech AG, High Point, NC, USA), and 2 IUTaq DNA polymerase (Promega). BV primers (MWG-BiotechAG), DNAase- and RNAase-free distilled water, and the appro-priate amounts of T-cell–derived cDNA were added to a finalvolume of 50 mL. Samples were amplified in a PTC-100 ThermalCycler (MJ Research, Watertown, MA, USA) with the followingPCR conditions: 10 minutes denaturing time at 95�C, 40 amplifi-cation cycles each consisting of 30 seconds at 94�C, 90 seconds at50�C, and 90 seconds at 68�C, followed by a final extension timeof 10 minutes at 72�C and cooling at 4�C. PCR products were thenstored at –20�C.

Visualization of BV-BC andBV-BJ rearrangements by fluorescent run-off reactionsTwo microliters multiplex PCR product from each set were addedto 10 mL fresh PCR reagent mix. BV-BC run-off reactions wereperformed using 1.3 mM BC2 primer originally described byAkatsuka et al. [11]. This was labeled with Ned, 6-Fam (Fam),

Table 2. Expected length of center peaks of b variable (BV) subfamilies

with normal CD3 length distribution profile

PCR tubes PCR products*

CDR3 length:

center peaks (bp)

Set 1 BV1-BC2 174

BV5-BC2 224

BV4-BC2 273

BV2-BC2 329

BV3-BC2 378

Set 2 BV7-BC2 165

BV9-BC2 215

BV8-BC2 268

BV6-BC2 330

Set 3 BV15-BC2 164

BV14-BC2 220

BV11-BC2 272

BV12-BC2 326

BV13-BC2 374

Set 4 BV20-BC2 166

BV17-BC2 218

BV18-BC2 268

BV16-BC2 374

Set 5 BV21-BC2 161

BV24-BC2 222

BV23-BC2 325

BV22-BC2 372

bp 5 base pairs.

*Gene segments were listed according to the mean

complementarity-determining region 3 (CDR3) length in nucleotides

they produced after multiplex polymerase chain reaction (PCR),

and according to their set of origin.

730 S. Mariani et al./ Experimental Hematology 2009;37:728–738

or Hex and used as follows: Ned-BC2 primer for sets 1 and 3,Fam-BC2 primer for sets 2 and 4, and Hex-BC2 primer for set5. BV-BJ run-off reactions were performed using 0.2 mM 13 sepa-rate Fam-BJ primers [15]. Run-off reactions were run for 15 cyclesat the same conditions described previously.

Capillary electrophoresis and software analysesEqual volumes of fluorescent BV-BC run-off products werepooled from sets 1 and 2, and from sets 3, 4, and 5 for capillaryelectrophoresis. Four microliters of these pooled products or4 mL monoplex BV-BJ run-off products were diluted in 10.5 mLdeionized formamide and 0.5 mL GeneScan-500 ROX size marker(both from Applied Biosystems, Applera Italia, Monza, Italy),heated at 95�C for 10 minutes for cDNA denaturation and imme-diately cooled on ice. Electrophoresis was run for 24 minutes at60�C for each tube through a POP-4 polymer on an ABI Prism310 capillary sequencer (Applied Biosystems). Data wereacquired on GeneScan platform and compensated with a standardD-Set matrix (Applied Biosystems) that specifically distinguishesNed, Fam, Hex, and Rox colors. Finally, CDR3 length profileswere determined by GeneScan or Genotyper software analyses(Applied Biosystems) through appropriate fluorophore channels.Center peaks of the Gaussian CDR3 length distributions fromeach BV-BC run-off are illustrated in Table 2.

Specificity and sensitivityThe TCRBV8þ Jurkat cell line and the TCR-ab–negative KMS27cell line were used as positive and negative controls. PBMC fromhealthy donors and patients with MM were head-to-head analyzedwith the multiplex PCR method and the standard immunoscopeassay [8]. All BV subfamilies were compared, with the exceptionof BV6 and BV13 subfamilies, because different sequences aretargeted by the specific primers in the two methods.

Detection limit studies were carried out to evaluate sensitivityand reproducibility. The positive and negative controls containedonly normal T cells and KMS27 cells, respectively. Five moresamples with total cDNA content of 400 ng were obtained byserial dilutions of normal T-cell cDNA with KMS27 cell cDNAas follows: 200/200 ng; 100/300 ng; 80/320 ng; 30/370 ng; 15/385 ng. The intensity of a BC-specific 306-bp band was assessedby monoplex PCR, and that corresponding to the 80/320 ngsample was set as the threshold value to predetermine whetheror not a sample with an unknown amount of T-cell derivedcDNA is suitable for the multiplex PCR analysis (see Results:Specificity and sensitivity of the multiplex PCR method).

Results

Amplification of TCRBV repertoire by multiplex PCRThe multiplex PCR method described by Akatsuka et al.was used as a building block for the first step of our method[11]. Twenty-two functional BV subfamilies were dividedinto five sets, each containing from four to sevensubfamilies (Table 1). Each set was amplified with four toseven BV primers under optimal conditions. Primer BV25was excluded from set 5 because the high concentrationneeded to detect the BV25 subfamily, poorly representedin the TCRBV repertoire [11], affected annealing of the

neighboring primers (data not shown). The BC primer(BC1) was the same as reported in the immunoscope assay[8] (Table 1). This choice was imposed by its relative posi-tion compared to the BC primer described by Akatsukaet al. [11] (BC2). In fact, the former and the latter primerswere ideal to perform a first-step of PCR and the followingsecond steps of emi-nested PCR reactions, respectively.

To optimize the sensitivity of the method, we also modifiedthe BV primer concentrations to obtain comparable peakheights in each set, despite the relative frequencies of BVsubfamilies (Table 1), and we increased the concentrationsof MgCl2 and Taq polymerase. Lastly, the annealing was per-formed at 50�C and the temperature and time of the extensionphase were set at 68�C and 90 seconds, respectively.

Effective visualization of BV-BCrearrangements by fluorescent PCR run-off reactionsAs a second step, we applied the fluorescent run-off PCRapproach of the immunoscope assay to the multiplex PCRproducts [8]. Two microliters of each set were amplifiedwith the fluorescent-nested primer BC2, which effectivelyrun at a temperature of 50�C in all sets rather than at58�C, as reported previously [11], probably because of anoverestimation of the primer melting temperature andconsequent unsuccessful annealing to any BV-BCtranscripts.

731S. Mariani et al. / Experimental Hematology 2009;37:728–738

BC2 was labeled with Ned, Hex, or Fam fluorophores tooptimally visualize run-off PCR reactions of multiplex PCRproducts. The most appropriate labelings were Ned for sets 1and 3, Fam for sets 2 and 4, and Hex for set 5. It was thuspossible to combine sets 1 and 2, and sets 3, 4, and 5 to reducethe number of samples to be run on the capillary sequencerwithout jeopardizing sensitivity and specificity (Fig. 1).

The multiplex PCR method was validated by head-to-head comparison with the immunoscope assay using samplesfrom normal donors and MM patients. The two methods gavesimilar results in all the BV subfamilies (Fig. 2). However,the multiple-member subfamilies BV6 and BV13 were notconsidered, because they comprise multiple genes differentlyamplified by the two methods. In fact, while the multiplexPCR approach targets all possible members of one subfamilyin the same PCR tube, and produce a single result for any BVsubfamily, the immunoscope method splits some members indifferent tubes, producing multiple results for the samesubfamily, as many as the members of the subfamily are[8,11].

Specificity and sensitivity of the multiplex PCR methodSpecificity was tested using the ab TCR-negative KMS27and the BV8-positive Jurkat cell lines [16] and sensitivityusing serial dilutions of highly purified normal ab T cellsmixed with KMS27 cells. Each dilution step compriseda total count of 2� 106 cells in the appropriate T cell toKMS27 cell ratio, with samples containing only ab T cellsor KMS27 cells as positive and negative controls, respec-tively. The multiplex PCR analysis was run on samplescontaining 400, 200, 100, 80, 30, 15, and 0 ng ab T-cell–derived cDNA. All BV subfamilies showed a polyclonalprofile in the undiluted T-cell cDNA sample. This patternwas still detected in 19 of 22 (86%) BV subfamilies inthe 80-ng T-cell cDNA sample (containing 1� 105 T cells).The method ceased to provide information when the T-cellcDNA amount was #30 ng, because the profile of O50%of BV subfamilies could not be visualized under theseconditions (Fig. 3a).

To predict whether a sample with an unknown number ofab T cells is suitable for the multiplex PCR method, wepredetermined the expression of a 306 BC gene segmentas a tracking marker to quantify the amount of T-cell–derived cDNA [14]. Figure 3b shows that a BC band inten-sity equivalent to at least 80 ng T-cell cDNA in thereference control sample is required to run an informativeassay.

Effective visualization of BV-BJrearrangements by fluorescent PCR run-off reactionsMost multiplex PCR methods have been applied at the BV-BC level only, whereas the immunoscope assay is alsoapplicable at the BV-BJ level. To merge the advantagesof the two methods, the BV-BC products generated instep 1 were used as templates for the fluorescent BV-BJ

run-off reactions and allowed a full TCRBV characteriza-tion in cell samples containing as few as 1� 105 T cells.This approach was validated in samples from MM patientsby determining the usage of BJ gene segments in BV fami-lies with an oligoclonal CDR3 length distribution profile. Arepresentative example is shown in Figure 4. The BV-BCrun-off reaction of set 1 displayed a predominant 273-bppeak in the BV4 subfamily consistent with the presenceof oligoclonal T cells (Fig. 4a). The same peak was visual-ized when the template was generated with a dedicatedmonoplex BV4-BC PCR (Fig. 4b). Two microliters fromset 1 were dispensed into different tubes and amplifiedwith 13 Fam-labeled BJ primers covering all functionalBJ gene segments. This showed that the 273-bp peak ispredominantly composed of oligoclonal T cells using theBJ1.4 and BJ1.5 gene segments, but also revealed the pres-ence within the BV4 subfamily of unsuspected polyclonal Tcells using the BJ2.5 and BJ2.7 gene segments. This usageof BJ segments was identical to that observed in run-offreactions of the monoplex BV4-BC PCR product(Fig. 4d). These results indicate that the presence oftemplates from multiple BV subfamilies does not affectthe specificity of BJ primers, and that the multiplex PCRmethod allows full characterization of BJ segment usagein multiple peaks of multiple BV subfamiliessimultaneously.

Validation of the multiplexPCR method in small T-cell samplesTregs displaying the CD4þ CD25high Foxp3þ phenotypewere selected to further validate our multiplex PCR assay.In our hands, Tregs represent 5% of circulating peripheralblood mononuclear cells in healthy individuals and theirtotal counts are approximately 40 cells/mL. After immuno-magnetic enrichment, we obtained a total number of 0.16and 1.8� 106 Tregs from two normal donors. The immuno-regulatory activity of Tregs was confirmed by showing theiranergy to CD3 stimulation, together with their ability tosuppress proliferation of autologous CD4þ CD25– T cellsstimulated by anti-CD3 monoclonal antibody in the pres-ence of irradiated accessory cells [17]. The multiplexPCR method was successfully applied to these samplesand clearly confirmed that the TCRBV repertoire of Tregsis polyclonal in normal donors (Fig. 5) [18].

DiscussionIts very high degree of combinatorial diversity makes theTCRBV repertoire the target of choice for T-cell molecularclonality studies. However, molecular assays estimating theCDR3 length polymorphism of each individual BV chaincan recognize more subtle antigen-driven variations than themere emergence of malignant T-cell clones. Our aim was toset up a suitable assay for full characterization of the TCRBVrepertoire (22 of 25 BV subfamilies) in small T-cell samples.

Figure 1. Multiplex polymerase chain reaction analysis of a normal donor TCRBV repertoire. The complementarity-determining region 3 (CDR3) length

distribution profiles of 22 BV subfamilies are grouped in five sets. In the successfully amplified BV subfamilies, the polyclonal profile is distinguished by

a bell-shaped distribution of BV-BC transcripts or by one or more peaks slightly above the normal bell-shaped background. Numbers in squares indicate the

length of the central peak of each individual BV subfamily in base pairs (bp).

732 S. Mariani et al./ Experimental Hematology 2009;37:728–738

Several multiplex PCR approaches have been devised for thispurpose and, in order to save time, cost, and cell input incomparison with monoplex PCR approaches. Major chal-lenges of multiplex PCR methods are careful selection of

BV primers to avoid annealing interference and over- or under-representation during the amplification phase; fine-tuning ofBV primer concentrations to compensate for the relativefrequency of BV subfamilies in the normal TCRBV repertoire;

Figure 2. Comparison between the multiplex polymerase chain reaction (PCR) method and the monoplex immunoscope assay. The complementarity-deter-

mining region 3 (CDR3) length distribution profiles of BV subfamilies from two representative sets (A) are compared with the corresponding profiles

obtained by monoplex PCR with the immunoscope assay (B).

733S. Mariani et al. / Experimental Hematology 2009;37:728–738

a sufficient amount of T-cell–derived cDNA to minimize thecompetition of multiplexed BV primers for the template andallow a balanced amplification of all BV subfamilies, irrespec-tive of their relative frequency.

So far, multiplex PCR methods have pooled togetherfrom 2 to 23 BV primers in 2 to 12 reaction sets yieldingto the analysis from 22 to 23 BV subfamilies [9–12]. OurPCR method multiplexed from four to seven BV primersin five sets and provided a full representation of 22 BVsubfamilies in samples containing as little as 80 ng T-cellderived cDNA, calculated to be equivalent to approximately1� 105 T cells per sample. This is an important resultbecause it means that overall characterization of TCRdiversity can be achieved in most cell samples withoutany prior T-cell purification step. So far, the lowest T-cell–derived cDNA amount has been 250 ng, and this wasobtained from purified normal or pathological T-cellsamples as opposed to unfractionated cell samples [10].

Most multiplex PCRs investigate the TCRBV repertoireby determining the CDR3 length polymorphism at the BV-BC level only. The BV-BC products generated in ourmultiplex PCR method could be successfully used forfluorescent BV-BJ run-off reactions to allow better charac-terization of BV subfamilies that already display an

abnormal CDR3 length distribution profile. This is the firsttime that the BV-BJ run-off reactions have been applied toa multiplex PCR method. This second level of acuitymimics the progressive magnification power of a micro-scope [8], and provides further information on the oligo-or monoclonality of predominant peaks already detectedat the BV-BC level. Indeed, head-to-head experimentshave shown that BJ usage of predominant peaks in BVsubfamilies with an oligoclonal CDR3 length distributionprofile is detected with the same accuracy, irrespective ofwhether monoplex or multiplex PCR products are used astemplates for fluorescent BV-BJ run-off reactions.

van Dongen et al. [12] multiplexed 23 BV primers, 13BJ primers, and 2 BD primers in three reaction sets,covering all functional BV and BJ gene segments [12].This method has been developed primarily for the clonalityassessment of T-cell malignancies, but not for the fine char-acterization of TCR diversity during immune responses,because it does not indicate the specific BV or BJ genesegments used by clonally restricted PCR products.

Chen et al. [19] reported recently on an oligonucleotide-based microarray to monitor immune reconstitution afterhematopoietic stem cell transplantation. The microarrayincluded 28 BV-specific and 13 BJ-specific probes and

Figure 3. Detection limit of the multiplex polymerase chain reaction (PCR) method. TCRBV repertoires obtained by multiplex PCR analysis of three serially

diluted representative samples containing 400 ng, 80 ng, and 30 ng T-cell cDNA (A). Eighty nanograms is the lower threshold needed to obtain an informative

assay in 19 of 22 BV families. (B) Results of a monoplex PCR assessing the intensity of 306-base pair band derived from the BC gene segment in serial

dilutions of T-cell–derived complementary DNA (cDNA). This has been elaborated as a preliminary assay to determine whether a sample with an unknown

number of T cells is suitable for the multiplex PCR analysis. A BC band intensity equivalent to 80 ng is the minimal amount to run an informative assay.

CDR3 5 complementarity-determining region 3.

734 S. Mariani et al./ Experimental Hematology 2009;37:728–738

effectively provided quantitative and qualitativeinformation on TCR diversity, including the ability to iden-tify clonal T-cell expansions. However, the clonality studieswere performed using T-cell lines with well-known BV-BJrearrangements, a technical approach that facilitates detec-tion of increased signals from the corresponding oligonu-cleotide probes. More importantly, clonal T-cellexpansions, to be detected in the test sample, should belarge enough to emerge from the normal boundaries of

the cohybridized reference sample (represented by pooledhealthy donors). Unfortunately, the biological and clinicalrelevance of immune-reactive T-cell clones is very oftenunrelated to their size. We have previously used a combina-tion of monoplex immunoscope and multiparameter flowcytometry methods to show that in MM patients incomplete remission after autologous transplantation, manynormally or underexpressed BV subfamilies (up to 20%of the total TCRBV repertoire) contain a restricted

Figure 4. Multiplex polymerase chain reaction (PCR) products can effectively be used for the analysis of BJ gene segment usage. Multiplex PCR products of

set 1 (A) and the monoplex BV4-BC PCR product (B) from a representative myeloma patient were head-to-head subjected to run-off reactions using 13

fluorescent BJ primers to characterize the BJ gene segment usage of a predominant 273-base pair (bp) peak detected by both methods within the BV4

subfamily (arrows). The BV-BJ run-off profiles are shown in (C) and (D). Both show that the 273-bp peak is mostly composed of oligoclonal T cells using

the BJ1.4 and BJ1.5 gene segments (arrows), and that the BV4 subfamily is also composed to a much lesser extent by polyclonal T cells using the BJ2.5 and

BJ2.7 gene segments (arrowheads). Besides providing the same accuracy, the multiplex PCR method offers additional information on BJ gene segment usage

by the other BV subfamilies included in set 1. CDR3 5 complementarity-determining region 3.

735S. Mariani et al. / Experimental Hematology 2009;37:728–738

Figure 5. Validation of the multiplex polymerase chain reaction method in small-sized T-cell samples. The figure shows the polyclonal TCRBV repertoire of

regulatory T cells (Tregs) isolated from a normal donor. CDR3 5 complementarity-determining region 3.

736 S. Mariani et al./ Experimental Hematology 2009;37:728–738

repertoire of T-cell clones [14]. Likewise, tumor-specificidiotype-reactive T cells have been identified in normallyrepresented oligoclonal as opposed to phenotypicallyexpanded BV subfamilies [20]. Likewise, caution isrequired when interpreting a BV or BJ expansion asa synonym of clonality. For instance, we have detected upto 7% of phenotypically expanded polyclonal BV subfam-ilies in MM after autologous transplantation [14].

Sequencing the CDR3 region achieves the greatest defi-nition and allows unequivocal identification ofmonoclonality and eventually the specificity for the

peptide/major histocompatibility complex, leading to char-acterization of reactive or pathogenetic T-cell clonotypes inpatients with various diseases, such as aplastic anemia,graft-vs-host disease, and others [21]. Likewise, CDR3sequencing of sorted single tetramerþVbþCD8þ T cellshas been used to provide insights into how antigen-specificT-cell responses are generated and maintained. Usinga single-cell TCR repertoire approach, it has been possibleto clonally dissect T-cell immunodominance hierarchies,including identification of ‘‘high’’ vs ‘‘low’’ avidity, or‘‘central’’ vs ‘‘effector’’ memory clonotypes, which is also

737S. Mariani et al. / Experimental Hematology 2009;37:728–738

essential for development of cellular-based vaccines andimmunotherapy [2,22,23].

These approaches are highly informative, but their appli-cation is limited by their complexity and biological and tech-nical requirements, such as knowledge of antigen-specificityand Vb usage, tetramers availability, and access to cell-sort-ing facilities. Pyrosequencing is a DNA-sequencing methodbased on the principle of sequencing-by-synthesis and pyro-phosphate detection through a series of enzymatic reactions[24]. This bioluminometric, real-time DNA sequencing tech-nique offers unique applications that are cost-effective anduser-friendly. The pyrosequencing technology has beenused in a broad range of applications, such as single nucleo-tide polymorphism genotyping, de novo mutation detection,gene identification, and microbial genotyping, but thisapproach remains unexplored in the analysis of TCR diver-sity. Our multiplex PCR method can be prodromal to identi-fication of recurrent T-cell clonotypes of clinical relevance by‘‘fishing’’ predominant peaks with identical BJ gene segmentusage in skewed BV subfamilies.

We have validated our multiplex PCR method by deter-mining the overall degree of TCR diversity expressed byTregs. These cells play a central role in a variety ofdiseases, ranging from autoimmunity and cancer to obesityand atherosclerosis [25]. Their low peripheral counts haveso far necessitated withdrawal of significant amounts ofblood for their purification and subsequent functional andmolecular analyses. We have been able to characterize thefull TCRBV repertoire in a sample containing as little as1.6� 105 Tregs. Our results confirmed that Tregs fromnormal donors use a large, unrestricted TCRBV repertoirevery similar to that of CD4þ CD25– T cells [26]. Theseresults are consistent with the functional properties ofTregs, which are activated through their respective TCR,but are then capable of suppressing immune responses ina nonspecific manner at a very low Treg-to-effector cellratio. These properties compensate for their low numberand allow them to exert suppressor activity in the absenceof antigen-driven clonal expansions.

In conclusion, we have developed a multiplex PCRmethod that allows comprehensive assessment of theTCRBV repertoire at the BV-BC and BV-BJ levels, andsaves a considerable amount of time, reagents, and cellinput. Compared to previous reports, we have been ableto visualize all functional BV-BC rearrangements in1 hour by using only 2 tubes per sample instead of 5 to12 tubes as reported previously [8–11,27]; and visualizeall BV-BJ rearrangements with 67 fluorescent run-off reac-tions as compared with 324 of the immunoscope assay [8].By using a single-channel capillary sequencer, we havealready cut down from 162 to approximately 33 hours thetotal time required to run the full TCRBV repertoire. Thissaving is open to further improvement by using multi-channel sequencer without any further modification of themultiplex PCR method.

AcknowledgmentsSupported in part by Ministero dell’Universita e della RicercaScientifica (M.I.U.R.) (Rome, Italy); Regione Piemonte (RicercaSanitaria e Ricerca Scientifica) (Torino, Italy); Compagnia SanPaolo di Torino (Torino, Italy), Fondazione Cassa di Risparmiodi Torino (C.R.T.) (Torino, Italy); Comitato Regionale PiemonteseGigi Ghirotti (Progetto Vita Vitae) (Torino, Italy); FondazioneNeoplasie Sangue (Fo.Ne.Sa) (Torino, Italy).

References1. Nikolich-Zugich J, Slifka MK, Messaoudi I. The many important fac-

ets of T-cell repertoire diversity. Nat Rev Immunol. 2004;4:123–132.

2. Kedzierska K, La Gruta NL, Stambas J, Turner SJ, Doherty PC.

Tracking phenotypically and functionally distinct T cell subsets via

T cell repertoire diversity. Mol Immunol. 2008;45:607–618.

3. McCloskey TW, Haridas V, Pahwa R, Pahwa S. T cell receptor V beta

repertoire of the antigen specific CD8 T lymphocyte subset of HIV in-

fected children. Aids. 2002;16:1459–1465.

4. Quaratino S, Feldmann M, Dayan CM, Acuto O, Londei M. Human

self-reactive T cell clones expressing identical T cell receptor beta

chains differ in their ability to recognize a cryptic self-epitope.

J Exp Med. 1996;183:349–358.

5. Alvarez CM, Opelz G, Giraldo MC, et al. Evaluation of T-cell receptor

repertoires in patients with long-term renal allograft survival. Am J

Transplant. 2005;5:746–756.

6. Hu H, Queiro MR, Tilanus MG, de Weger RA, Schuurman HJ.

Expression of T-cell receptor alpha and beta variable genes in normal

and malignant human T cells. Br J Haematol. 1993;84:39–48.

7. Gorski J, Yassai M, Zhu X, et al. Circulating T cell repertoire

complexity in normal individuals and bone marrow recipients

analyzed by CDR3 size spectratyping. Correlation with immune

status. J Immunol. 1994;152:5109–5119.

8. Pannetier C, Even J, Kourilsky P. T-cell repertoire diversity and clonal

expansions in normal and clinical samples. Immunol Today. 1995;16:

176–181.

9. Maslanka K, Piatek T, Gorski J, Yassai M, Gorski J. Molecular anal-

ysis of T cell repertoires. Spectratypes generated by multiplex poly-

merase chain reaction and evaluated by radioactivity or

fluorescence. Hum Immunol. 1995;44:28–34.

10. Fernandes S, Chavan S, Chitnis V, Kohn N, Pahwa S. Simplified

fluorescent multiplex PCR method for evaluation of the T-cell

receptor V beta-chain repertoire. Clin Diagn Lab Immunol. 2005;

12:477–483.

11. Akatsuka Y, Martin EG, Madonik A, Barsoukov AA, Hansen JA.

Rapid screening of T-cell receptor (TCR) variable gene usage by

multiplex PCR: application for assessment of clonal composition.

Tissue Antigens. 1999;53:122–134.

12. van Dongen JJ, Langerak AW, Bruggemann M, et al. Design and

standardization of PCR primers and protocols for detection of

clonal immunoglobulin and T-cell receptor gene recombinations

in suspect lymphoproliferations: report of the BIOMED-2

Concerted Action BMH4-CT98-3936. Leukemia. 2003;17:

2257–2317.

13. Chomczynski P, Sacchi N. Single-step method of RNA isolation by

acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Bi-

ochem. 1987;162:156–159.

14. Mariani S, Coscia M, Even J, et al. Severe and long-lasting disruption

of T-cell receptor diversity in human myeloma after high-dose chemo-

therapy and autologous peripheral blood progenitor cell infusion. Br J

Haematol. 2001;113:1051–1059.

15. Puisieux I, Even J, Pannetier C, Jotereau F, Favrot M, Kourilsky P. Oli-

goclonality of tumor-infiltrating lymphocytes from human melanomas.

J Immunol. 1994;153:2807–2818.

738 S. Mariani et al./ Experimental Hematology 2009;37:728–738

16. Halle JP, Haus-Seuffert P, Woltering C, Stelzer G, Meisterernst M. A

conserved tissue-specific structure at a human T-cell receptor beta-

chain core promoter. Mol Cell Biol. 1997;17:4220–4229.

17. Kuniyasu Y, Takahashi T, Itoh M, Shimizu J, Toda G, Sakaguchi S.

Naturally anergic and suppressive CD25 þ CD4 þ T cells as a func-

tionally and phenotypically distinct immunoregulatory T cell subpop-

ulation. Int Immunol. 2000;12:1145–1155.

18. Fujishima M, Hirokawa M, Fujishima N, Sawada K. TCRalphabeta

repertoire diversity of human naturally occurring CD4 þ CD25þregulatory T cells. Immunol Lett. 2005;99:193–197.

19. Chen X, Hale GA, Neale GA, et al. A novel approach for the analysis

of T-cell reconstitution by using a T-cell receptor beta-based oligonu-

cleotide microarray in hematopoietic stem cell transplantation. Exp

Hematol. 2007;35:831–841.

20. Halapi E, Werner A, Wahlstrom J, et al. T cell repertoire in patients

with multiple myeloma and monoclonal gammopathy of undetermined

significance: clonal CD8 þ T cell expansions are found preferentially

in patients with a low tumor burden. Eur J Immunol. 1997;27:2245–

2252.

21. Plasilova M, Risitano A, Maciejewski JP. Application of the molecular

analysis of the T-cell receptor repertoire in the study of immune-medi-

ated hematologic diseases. Hematology. 2003;8:173–181.

22. Kedzierska K, Venturi V, Field K, Davenport MP, Turner SJ, Doherty

PC. Early establishment of diverse T cell receptor profiles for influ-

enza-specific CD8(þ)CD62L(hi) memory T cells. Proc Natl Acad

Sci U S A. 2006;103:9184–9189.

23. Venturi V, Kedzierska K, Tanaka MM, Turner SJ, Doherty PC, Daven-

port MP. Method for assessing the similarity between subsets of the T

cell receptor repertoire. J Immunol Methods. 2008;329:67–80.

24. Ahmadian A, Ehn M, Hober S. Pyrosequencing: history, biochemistry

and future. Clin Chim Acta. 2006;363:83–94.

25. Brusko TM, Putnam AL, Bluestone JA. Human regulatory T cells: role

in autoimmune disease and therapeutic opportunities. Immunol Rev.

2008;223:371–390.

26. Kasow KA, Chen X, Knowles J, Wichlan D, Handgretinger R, Riberdy

JM. Human CD4 þ CD25þ regulatory T cells share equally complex

and comparable repertoires with CD4 þ CD25- counterparts. J Immu-

nol. 2004;172:6123–6128.

27. Ou Y, Tong C, Zhang Y, et al. An improved design of PCR primers for

detection of human T cell receptor b chain repertoire. Mol Biol Rep.

2009;36:145–152.

28. Wei S, Charmley P, Robinson MA, Concannon P. The extent of the

human germline T-cell receptor V beta gene segment repertoire.

Immunogenetics. 1994;40:27–36.