Clonality assessment via analysis of immunoglobulin (Ig) and T cell receptor (TCR) gene rearrangements has become an important and valuable adjunct in the diagnostic process of suspect lymphoproliferations. This holds for histopathologically suspect lymphoid lesions in tissues (lymph nodes, spleen, skin, and gastrointestinal tract) but is similarly true in case of a lymphocytosis or suspect lymphocyte population in bone marrow, peripheral blood, or other body fluids (such as spinal fluid). Initially, such clonality testings could only be done by means of Southern blot analysis [10, 11]. Meanwhile, multiple protocols and primer sets for PCR-based clonality assessment have been developed, initially mainly focusing on two targets: IGH (Ig heavy chain locus) for B cell clonality and TCRG (TCR gamma locus) for T cell clonality. These two targets were mainly chosen because of their broad applicability and because of the limited number of primers required. At the end of the 90s of the previous century, a European BIOMED-2 consortium started to redesign primers and optimize PCR protocols for Ig/TCR clonality testing [12]. This has resulted in standardized multiplex PCR protocols for IGH and TCRG analysis, which showed very good performance when evaluated in large series of WHO-defined lymphoma and leukemia entities [1, 3, 6]. Importantly, the BIOMED-2 consortium (now called EuroClonality) has also developed multiplex PCR protocols for alternative targets for B cell (IGK and IGL, Ig kappa and lambda light chain locus, respectively) and T cell (TCRB and TCRD, TCR beta and delta, respectively) clonality testing. Especially, the IGK and TCRB gene rearrangement assays have proven to be highly informative. Hence, these targets could serve as alternatives for IGH and TCRG, respectively, but also have important complementary value and unprecedented performance when used in combination with IGH and TCRG [1, 3].

Along with the technical standardization of the multiplex PCR protocols and their commercial availability in kit format, these tests have now found their way to routine clonality testing laboratories worldwide. Correct interpretation of the results of these multiplex PCR reactions is, therefore, now the challenge. Via educational workshops, publications, and a website-based support system, the EuroClonality group is actively contributing to optimal use and interpretation of these assays. One of the major issues in correct interpretation of clonality testing results appears to be the occurrence of multiple clonal PCR products. This is especially true for the IGK and TCRB targets due to the intrinsic complexity of the configuration of these two loci. Here, we address the issue of multiple clonal products, evaluate situations in which multiple rearrangements are generated in the IGK and TCRB loci, and discuss the interpretation of multiple products in the context of clonality assessment diagnostics.

Multiple rearrangements can result in multiple clonal products

Both B and T lymphocytes undergo several stages of differentiation. For B lymphocytes, the initial differentiation process takes place in the bone marrow, whereas early T lymphocyte differentiation is in the thymus. A common feature of both differentiation pathways is the fact that the precursor B cells and precursor T cells (also called thymocytes) undergo hierarchical gene rearrangements in their Ig and TCR loci, respectively [2, 13]. B cell precursors start rearranging the IGH locus with an initial D–J gene rearrangement, followed by coupling of a V gene to the D–J recombination; following evaluation of the expressed IgH chain in a preB cell receptor (preBCR) complex, the next step is to create a proper Ig light chain. To this end, the IGK locus first starts to rearrange its V and J genes. If this does not lead to a good Igκ chain, the next step is to rearrange V and J genes in the IGL locus in order to generate a good Igλ chain (Fig. 1a). In an analogous way, thymocytes undergo hierarchical rearrangements in their TCR loci, starting with D–D, D–J, and V–DJ recombinations in the TCRD locus, followed by V–J rearrangements in the TCRG locus. Unless this leads to a proper TCRγδ receptor, the thymocytes will continue to rearrange their TCRB loci (initially D–J, followed by V–DJ rearrangement) and, finally, their TCRA loci (V–J recombinations) (Fig. 1b).

Fig. 1
figure 1

Gene rearrangements during human lymphoid differentiation. a Hierarchy of Ig gene rearrangements during human precursor B cell differentiation in the bone marrow. b Hierarchy of TCR gene rearrangements during human precursor T cell differentiation in the thymus. HSC hematopoietic stem cell, DN double negative thymocyte, DP double positive thymocyte, ISP immature single positive thymocyte, SP single positive thymocyte. a was adapted from Van Zelm et al. [13]. b was adapted from Dik et al. [2]

The mere purpose of the whole V(D)J recombination process is thus to create a unique antigen receptor (Ig or TCR molecule) that is positively selected to recognize antigen and negatively selected for showing too high affinity to autoantigens. However, V(D)J recombination is not a directed process. Rather, the coupling of V, (D), and J genes is purely random. Moreover, during the actual coupling process, the involved V, D and J genes are trimmed and nontemplated nucleotides are (randomly) inserted, which further adds to the diversity. Since a proper antigen receptor chain can only be formed if the reading frame of three nucleotides (triplet codons) is preserved, many rearrangements are unproductive (out-of-frame rearrangement or rearrangement with stop codon). This implies that for a given Ig/TCR locus, very frequently two rearrangement attempts have taken place: one on each chromosome (allele). Given that the theoretical chance for a rearrangements to be in the correct reading frame and lacking a stop codon is ~25–30%, this implies that biallelic Ig/TCR gene rearrangements (D–J and/or V–DJ) are thus more rule than exception in B and T lymphocytes.

Since leukemias and lymphomas are the malignant counterparts of normal lymphocytes, it can be appreciated that the tumor clones that derive from a single transformed lymphocyte mostly contain biallelic rearrangements as well. Hence, in PCR-based clonality testing, the combination of two V–(D)J rearrangement products or one incomplete D–J product and a complete V–(D)J product is frequently seen. Even though biallelic rearrangements are thus the most likely underlying biological phenomenon for having two PCR products, the presence of two different clones could also be compatible with multiple products. Indeed, the presence of two parallel clones has been described in leukemia and lymphoma with a frequency of ~5% of cases [9]. That notwithstanding, biallelic rearrangements in lymphoma are often a more logical explanation than biclonality, unless additional evidence is available on the existence of two phenotypically distinct cell populations from either immunohistochemistry or flowcytometry.

Detection of up to two rearrangements per Ig/TCR locus can thus be readily explained by mono- or biallelic rearrangements. However, if three or more PCR products are found, other explanations have to be considered including the existence of two clones (biclonality) or even multiple clones (oligoclonality). In order to distinguish between (biallelic) monoclonality versus biclonality versus multiple clones/oligoclonality, a careful evaluation of the number of distinct PCR products per locus is required (Table 1). However, it may occasionally be difficult to determine which PCR products truly reflect clonal rearrangements. It is, therefore, strongly recommend to always run clonality assays in duplicate. Results from duplicate analyses will disclose the reproducibility of the PCR products in terms of amplicon size and signal intensity, which facilitates proper interpretation of the number of true clonal products (one, two or more). Notably, in case of a limited number of cells (so-called paucicellular samples), often seemingly clear clonal PCR products are detected albeit with nonreproducible amplicon sizes. This is sometimes referred to as pseudoclonality, which is defined as the detection of nonreproducible clonal PCR products due to selective amplification of low amount of template (Table 1). When there is a broad and diverse repertoire, which in fact reflects many different lymphocytes (clones), the term polyclonality is used. Evaluation of two different Ig or TCR targets (IGH plus IGK or TCRG plus TCRB) might sometimes shed light on interpretation of the number of clones present in the lesion. It should be noted, however, that the IGK and TCRB loci are a bit exceptional. Due to their specific configuration, >1 rearrangement can be present on the same chromosome/allele, and hence, three or even four PCR products can easily be compatible with one clone (Table 1). The complexity of the IGK and TCRB loci and the expected type and number of PCR products will be addressed in the next sections in more detail.

Table 1 Spectrum of clonality

Multiple rearrangements in the IGK locus

When during its differentiation a precursor B cell is not successful in creating a productive IGK rearrangement, further recombination between one of the upstream Vκ genes and one of the downstream Jκ genes might occur. In theory, such ongoing recombination can be repeated as long as upstream Vκ genes and downstream Jκ genes are still present on that allele. Alternatively, the second IGK allele will undergo one or more Vκ–Jκ recombinations in an attempt to make a productive IGK rearrangement (in-frame, no stop codon). If all options fail, the precursor B cell will switch to recombination of the IGL locus. Prior to initiation of IGL recombination, however, a completely different type of IGK recombination occurs. This is due to the presence of additional elements in the IGK locus that can be involved in recombination (reviewed in Langerak and Van Dongen [5]) (Fig. 2a). One of these is the so-called Kappa deleting element (Kde), which is located 24 kb downstream of the Cκ region. This Kde can recombine to i) any of the Vκ genes upstream of the unproductive Vκ–Jκ exon (resulting in a Vκ–Kde rearrangement) or ii) the other nonclassical IGK rearrangeable sequence in the intron (intron RSS) between the Jκ genes and the Cκ region (leading to an intron-Kde rearrangement) (Fig. 2b). The effect of both types of rearrangements is a deletion of the Cκ region on the involved allele as well as of the iEκ and 3′Eκ enhancer regions that normally control expression of the IGK locus in cis. The net effect of these recombinations is thus a complete functional inactivation of the involved IGK allele. In the case of a Vκ–Kde rearrangement, the preexisting Vκ–Jκ recombination is thereby removed (Fig. 2b). However, in the other scenario in which an intron-Kde rearrangement is employed to inactivate the IGK locus, the prior Vκ–Jκ recombination will be preserved. As a consequence, two different IGK rearrangements now reside on the same allele: a Vκ–Jκ rearrangement plus an intron-Kde rearrangement (Fig. 2b).

Fig. 2
figure 2

Multiple rearrangements in the human IGK locus. a Overview of the human IGK locus on chromosome band 2p11.2. Rearrangeable Vκ genes are indicated in dark grey (productive genes) or light gray (unproductive genes). The cluster with inverted Vκ genes (designated with D) is located approximately 800 kb upstream of the noninverted proximal Vκ gene cluster. Adapted from ImMunoGeneTics database ( b A schematic representation of consecutive rearrangements in the human IGK locus. Following an initial Vκ–Jκ rearrangement, additional rearrangements can occur that inactivate the IGK locus by removing the Cκ region and the enhancers (iEκ and 3′Eκ). These inactivating rearrangements involve the Kde sequence that can recombine to one of the Vκ genes (Vκ–Kde rearrangement), thereby deleting the preexisting Vκ–Jκ rearrangement. Alternatively, Kde recombines to an isolated recombination signal sequence (RSS) in the Jκ–Cκ intron; in that situation, the earlier Vκ–Jκ and intron-Kde rearrangements are both present on the same allele. c The Vκ genes are organized in a proximal cluster and a distal cluster, the latter showing an inverted orientation. Whereas recombination between a Vκ proximal gene and a Jκ gene follows the normal rearrangement process with deletion of the intermediate region, recombination between a Jκ gene and a Vκ distal gene can only occur through an inversional rearrangement without loss of genomic DNA. Ongoing Vκ–Jκ recombination involving a Vκ distal gene will thus result in two different Vκ–Jκ rearrangements being present on the same allele

Depending on the exact configuration of its two IGK alleles, up to four rearrangements can thus be present in one cell (Table 2). This phenomenon has major implications for interpretation of multiple PCR products in IGK clonality testing. The BIOMED-2 multiplex IGK assay consists of two reaction mixtures, one targeting the classical Vκ–Jκ rearrangements (tube A) and the other targeting the Kde rearrangements (tube B). Both assays are very informative and have proven their clear complementary value in clonality testing [3]. Particularly, the Kde assay has an enormous advantage because it targets rearrangements that are free of somatic hypermutations, thus reducing the rate of false negativity. However, when applying both IGK assays (Vκ–Jκ rearrangements plus Kde rearrangements) cautious interpretation is thus warranted due to the biological complexity of the IGK locus. In fact, per assay, a total of two rearrangements can be detected per clone (similar to other Ig/TCR loci), but for both IGK multiplex assays, together this can mount to four distinct PCR products in a single B cell clone (Fig. 3; Table 2). A practical approach when dealing with complex IGK patterns from multiplex tubes A and B would be to interpret the pattern of a given case along with the configurations as mentioned in Table 2. This will help to evaluate if and how the clonal products could belong to the same clone. As IGH patterns will often be available as well, this will almost always help to confirm or exclude the monoclonal character of the suspect B cell population. Only in ambiguous cases might sequencing of the products be necessary to solve interpretation.

Table 2 Summary of different IGK rearrangement configurations that are compatible with a single clone
Fig. 3
figure 3

Example of a case with multiple IGK products belonging to one clone. In case 2011–463, an aberrant leukemic B cell population is diagnosed. GeneScan analysis of IGK tube A shows two different clonal V–J products (149 and 197 nt, respectively), whereas two additional clonal Kde products are seen upon analysis of IGK tube B (239 and 278 nt products). Based on the amplicon sizes, the most logical explanation of these IGK patterns is a Vκ–Jκ product (149 nt) on one allele, together with an intron-Kde rearrangement on the same allele (278-nt product), whereas the second allele contains a Vκ–Kde rearrangement (239-nt product). The 197-nt Vκ–Jκ product most probably results from known cross-priming of the Vκ3 primer to the true Vκ–Jκ rearranged product. Thus, the IGK pattern is still compatible with a single B cell clone (see also Table 2), which is confirmed by the GeneScan analysis results of IGH tubes A (330/340 nt), B (278 nt), and C (polyclonality)

For reasons of completeness, it is good to mention that yet other rearrangements involving the intron RSS element can occur in the human IGK locus [4]. However, as these so-called J–intron rearrangements are quite rare and not detectable with the IGK multiplex clonality assays, they are beyond the scope of this paper and thus not further discussed here.

One other complicating aspect of the human IGK locus is worth mentioning here as it may influence interpretation of multiple PCR products. Contrary to other Ig/TCR loci, the Vκ genes are organized in two different clusters: a proximal cluster just upstream of the Jκ genes and a distal cluster further upstream (Fig. 2a). This 3′ distal cluster has most probably arisen by duplication during evolution. The involved Vκ genes in this distal cluster show an inverted orientation, implying that an inversional rearrangement is required to form a V–J recombination. As during an inversional rearrangement no DNA is removed from the genome, this Vκ–Jκ recombination will be preserved upon a second inversional Vκ–Jκ rearrangement (Fig. 2c). Consequently, theoretically, two or more Vκ–Jκ rearrangements would be detectable on the same allele if ongoing recombinations occur multiple times in an attempt to obtain a productive IGK allele. This inversional rearrangement process would imply that even more Vκ–Jκ PCR products belonging to one clone might be detectable in a given case. It is quite hard to address this phenomenon experimentally or to estimate the extent of its occurrence, but it is something that should be considered when interpreting IGK clonality testing data.

Finally, some of the Vκ primers in the multiplex assay are known to show some level of cross-reactivity to particular genes of other Vκ families. As the bindings sites for the different Vκ family primers are not all clustered in the same area, this might sometimes give an additional (weak) PCR product next to the correct Vκ–Jκ PCR product. Collectively, this means that there is no direct relationship between the number of clonal IGK PCR products and the number of B cell clones, and that caution is needed in interpreting IGK clonality testing results.

Multiple rearrangements in the TCRB locus

Notably, all Ig and TCR loci roughly have a similar genomic configuration, with a cluster of V genes, a cluster of D genes (at least for some loci), and a cluster of J genes. However, the configuration of the TCRB locus is slightly different in the sense that the two Dβ genes are not clustered but rather dispersed. Moreover, each of these Dβ genes comes with its own set of Jβ genes (Fig. 4). This configuration of two Dβ–Jβ clusters has probably arisen through duplication during evolution and has clear implications for rearrangement detection.

Fig. 4
figure 4

Multiple rearrangements in the human TCRB locus. The human TCRB locus consists of two Dβ–Jβ clusters. As a result, two independent Dβ–Jβ rearrangements (one involving the β1 region, the other involving the β2 region) can be present on the same allele. In the second phase, a Vβ gene can rearrange to either one of the Dβ–Jβ joints. Depending on the Dβ–Jβ joint that is targeted, this will create multiple rearrangements (Vβ–Dβ–Jβ1 plus Dβ2–Jβ2) or just a single rearrangement (Vβ–Dβ–Jβ2) on that allele

If during differentiation a thymocyte has not productively rearranged its TCRD and TCRG genes to form a functional TCRγδ receptor, recombination will start in the TCRB locus (Fig. 1b). Normally, as a first step in TCRB recombination one of the two Dβ genes will be coupled to one of the Jβ genes. As a direct consequence of the specific TCRB configuration with two separated Dβ–Jβ regions (β1 and β2 areas), two different Dβ–Jβ rearrangements might be formed on the same allele. If during further differentiation the thymocyte continues to rearrange its TCRB locus, one of these two Dβ–Jβ recombinations will be completed to a Vβ–Dβ–Jβ rearrangement. In case this Vβ–Dβ–Jβ rearrangement involves the prior Dβ2–Jβ2 recombination, the entire intergenic region including the already formed Dβ1–Jβ1 recombination will be removed on the circular excision product. Conversely, if the Vβ–Dβ–Jβ rearrangement involves the prior Dβ1–Jβ1 recombination, the more downstream located Dβ2–Jβ2 recombination will now be preserved on the same allele (Fig. 4). Hence, based on the exact type of rearrangements that occur on the two TCRB alleles, multiple PCR products might be amplified from one clone. When evaluating only the two complementary Vβ–Jβ multiplex PCR assays (BIOMED-2 tubes A and B), the number of PCR products (of correct size) is limited to two. However, if also Dβ–Jβ rearrangements (BIOMED-2 TCRB tube C) are evaluated, the number of distinct PCR products compatible with one clone could mount to four (Table 3).

Table 3 Summary of different TCRB rearrangement configurations that are compatible with a single clone

Apart from the multiple rearrangements on one allele, another phenomenon might complicate TCRB analysis with respect to multiple PCR products. Both Vβ–Jβ and Dβ–Jβ rearrangements could give rise to the occurrence of additional PCR products of extended size. This is to be explained by the relatively small intergenic Jβ gene distances of sometimes just 100–150 bp. As both TCRB multiplex PCR mixtures consist of Jβ gene-specific primers, occasionally this would allow additional amplification of a particular Vβ–Jβ/Dβ–Jβ rearrangement from the Jβ gene immediately downstream of the Jβ gene that is involved in the actual rearrangement [7, 8]. When the primer for the downstream Jβ gene is present in the same mixture, this will give rise to double products (of distinct size) in the same PCR reaction (Fig. 5). Even more subtle is the situation that the Vβ–Jβ rearrangement gives rise to an amplicon of the correct size in the one reaction (amplified with the true Jβ primer) and to an extended amplicon in the other PCR reaction (amplified with a primer for the downstream Jβ gene). The phenomenon of extended PCR products (including TCRB) is discussed in much more detail in the article by Rothberg et al. (this issue of the journal).

Fig. 5
figure 5

Extended PCR products in the human TCRB locus. Because of the close proximity of the Jβ genes, extended PCR products can be generated due to priming from a downstream Jβ gene. In the example, the Vβ22–Jβ2.3 rearrangement is detected as a 246-nt product in GeneScan analysis, while an extended 398-nt product is detected that results from priming at the Jβ2.4 gene (figure from: [7])

Similar to what has been discussed for complex IGK patterns, the presence of multiple TCRB products should also be interpreted in comparison to the possible TCRB configurations (Table 3), while taking into account the phenomenon of extended TCRB products as well.

Concluding remarks on interpretation of multiple clonal products

Knowledge on the immunobiological phenomena leading to multiple clonal Ig or TCR products is a key issue for correct interpretation of clonality testing results. Although an obvious relation exists between the number of clonal products for a given Ig/TCR locus and the number of B/T cell clones (spectrum of monoclonality–biclonality–oligoclonality–polyclonality), the situation is not always that straightforward. First, the presence of two different PCR products could reflect biclonality but can often more easily be attributed to the occurrence of biallelic rearrangements in a single clone. Second, due to the specific configuration of the IGK and TCRB loci, multiple rearrangements can occur on one allele resulting in the possibility that up to four PCR products could be compatible with a single clone. Finally, when >2 clonal Ig/TCR products (>4 for IGK and TCRB) are detectable, this could reflect biclonality, oligoclonality, or even pseudoclonality. It is not always easy to distinguish between those options but the use of duplicates will be helpful to determine reproducibility and relevance of the detected clonal PCR products. In conclusion, straightforward interpretation of clonality testing results can be hampered by the occurrence of multiple clonal products. Only careful consideration of immunobiological and technical explanations will prevent incorrect interpretation in such cases.