We performed an immunogenetic analysis of 345 IGHV–IGHD–IGHJ rearrangements from 337 cases with primary splenic small B-cell lymphomas of marginal-zone origin. Three immunoglobulin (IG) heavy variable (IGHV) genes accounted for 45.8% of the cases (IGHV1-2, 24.9%; IGHV4-34, 12.8%; IGHV3-23, 8.1%). Particularly for the IGHV1-2 gene, strong biases were evident regarding utilization of different alleles, with 79/86 rearrangements (92%) using allele *04. Among cases more stringently classified as splenic marginal-zone lymphoma (SMZL) thanks to the availability of splenic histopathological specimens, the frequency of IGHV1-2*04 peaked at 31%. The IGHV1-2*04 rearrangements carried significantly longer complementarity-determining region-3 (CDR3) than all other cases and showed biased IGHD gene usage, leading to CDR3s with common motifs. The great majority of analyzed rearrangements (299/345, 86.7%) carried IGHV genes with some impact of somatic hypermutation, from minimal to pronounced. Noticeably, 75/79 (95%) IGHV1-2*04 rearrangements were mutated; however, they mostly (56/75 cases; 74.6%) carried few mutations (97–99.9% germline identity) of conservative nature and restricted distribution. These distinctive features of the IG receptors indicate selection by (super)antigenic element(s) in the pathogenesis of SMZL. Furthermore, they raise the possibility that certain SMZL subtypes could derive from progenitor populations adapted to particular antigenic challenges through selection of VH domain specificities, in particular the IGHV1-2*04 allele.
Splenic marginal zone lymphoma (SMZL) is a low-grade B-cell lymphoma, accounting for approximately 1–2% of non-Hodgkin lymphomas.1 Since its original description in 1992,2 the ontogenetic derivation and histogenesis of SMZL have been hotly debated. More recently, it became clear that SMZL shows distinct characteristics compared with other low-grade B-cell lymphomas, enabling its inclusion in the 2008 WHO classification of Tumors of the Haematopoietic and Lymphoid tissues as a distinct clinical and pathological entity.3 In addition to SMZL, the 2008 WHO classification recognizes a provisional category of splenic lymphoma/leukemia unclassifiable (SLLU), including entities such as splenic diffuse red pulp lymphoma (SDRL)4 and hairy cell leukemia-variant.5 Currently, the ontogenetic relationship of SMZL to these provisional entities remains undefined.
Immunogenetic analysis offers valuable insight into the ontogeny of B-cell malignancies and the assignment of the malignant B cells to their normal B-cell counterparts.6 Biases in the usage of certain immunoglobulin (IG) heavy variable (IGHV) genes constitute evidence for the involvement of antigens and/or superantigens in lymphomagenesis by stimulating the proliferation of B cells with distinctive IG receptors. Additionally, many lymphoma subtypes are characterized by somatic hypermutation (SHM) in IGHV genes, suggestive of antigen-driven affinity maturation.7, 8
Previous studies by us and others have shown that SMZL shows a biased IG gene repertoire and is characterized by heterogeneity with regard to the mutational status of the IG receptors.4, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 Individual studies have highlighted different IG sequence features, including repertoire differences between SMZL, SDRL and hairy cell leukemia-variant;4, 15 intraclonal diversification of IG genes through ongoing SHM;12, 17, 18, 23 potential analogies between SMZL and hairy cell leukemia-variant;15 and, very recently, the existence of cases sharing quasi-identical antigen-binding sites (stereotyped B-cell receptors).21 Furthermore, some studies have also examined whether particular molecular characteristics of the IG receptors might be associated with genetic or phenotypic features, and/or clinical outcome. However, owing to their retrospective nature, definitive conclusions cannot be drawn, except possibly for certain genomic aberrations, which have been detected with higher frequency among IGHV-unmutated cases (especially, deletion of chromosome 7q10, 20, 22 and/or cases expressing particular IGHV genes).20, 22
With few exceptions,21, 22 most relevant publications have been based on the analysis of relatively small series. Furthermore, the inclusion criteria and types of tissue specimens examined varied significantly, introducing classification uncertainties (for example, when spleen histology was unavailable) and/or selection biases (for example, by analyzing exclusively blood or spleen specimens). Finally, from a biological standpoint, the available information about the precise patterns of SHM in SMZL is limited, especially in subsets defined by the usage of certain IGHV genes.
We conducted the present study in order to systematically explore the IG gene repertoire in a series of 337 cases with primary splenic small B-cell lymphomas of marginal-zone origin, by far the largest series yet, consolidated in the context of a multicenter collaboration coordinated by the Splenic B cell Lymphoma Group. We demonstrate that over 30% of cases with a diagnosis of SMZL based on splenic histopathology express distinctive IG receptors that use a single polymorphic variant of the IGHV1-2 gene and also show restricted antigen-binding site motifs and precise targeting of SHM.
These findings document the existence of molecular subtypes of SMZL defined by immunogenetic analysis of the IG receptors. In addition, they allude to selection by specific (super)antigenic element(s) in the development of at least a major subset of SMZL. Finally, they indicate that certain subtypes of SMZL could derive from progenitor cell populations adapted to particular antigenic challenges through usage of a limited set of B-cell receptor with predetermined specificities.
Patients and methods
Overall 337 patients with splenomegaly and an immunophenotype suggestive of primary splenic small B-cell lymphomas of marginal-zone origin (mainly SMZL and, in a few cases, SLLU) from collaborating institutions in France, Greece, Italy, Spain, the United Kingdom and the United States of America were included in this study (Supplementary Information).
Splenectomy at diagnosis or during the course of the disease was performed in 226/337 cases (67%). Based on spleen biopsy histopathological findings and following the 2008 WHO classification criteria,3 205/226 cases with available information were classified as typical SMZL, 17/226 as SDRL and, finally, 4/226 as splenic small B-cell lymphoma not otherwise classifiable. For non-splenectomized cases (n=111), diagnosis was based on a combination of features, including clinical presentation, lymphocyte morphology, immunophenotype, cytogenetics and/or fluorescence in situ hybridization analysis (demonstrating absence of the t(11;14) and t(14;18) chromosomal translocations), and bone marrow histology, following the recently published Splenic B cell Lymphoma Group guidelines.1 Some cytogenetic, molecular and clinical information has been reported previously.10, 13, 16, 22
The study was conducted in accordance with the Declaration of Helsinki and approved by the local Ethics Committee of each institution.
PCR amplification of IGHV–IGHD–IGHJ rearrangements
The starting materials for analysis were genomic DNA (gDNA) and total cellular RNA, isolated from fresh or formalin-fixed, paraffin-embedded spleen specimens (150 cases), and fresh peripheral blood samples (187 cases), by standard methods.
PCR amplification of IGHV–IGHD–IGHJ rearrangements was performed following the protocols described in previous publications by our groups.10, 13, 14, 15, 16 In general, when fresh samples were available, reverse transcriptase-PCR or gDNA-PCR amplification of IGHV–IGHD–IGHJ rearrangements was performed using IGHV leader primers or consensus primers for the IGHV FR1 along with appropriate IGHJ genes, as described previously.10, 13, 14, 15, 16 For the formalin-fixed, paraffin-embedded material, gDNA-PCR amplification of IGHV–IGHD–IGHJ rearrangements was performed using a semi-nested approach, using the same IGHV FR1 or, in 19/337 cases (5.6%), IGHV FR2 5′ consensus primers in both amplification rounds, with two different IGHJ 3′ consensus primers, of which the second round primer was more internal.14
In order to evaluate the presence of intraclonal diversification, subcloning analysis of the IGHV–IGHD–IGHJ PCR amplicons from four cases expressing IGHV1-2*04 IG receptors was performed, as described previously.24
Sequence analysis and interpretation
PCR amplicons were subjected to direct sequencing on both strands. Overall, 345 productive IGHV–IGHD–IGHJ rearrangements were evaluated (eight cases carried double productive rearrangements): of these, 160 have been reported in previous publications by our groups (Supplementary information).
Sequences were analyzed using the IMGT databases25, 26 and the IMGT/V-QUEST tool (version 3.2.17, Université Montpellier 2, CNRS, LIGM, Montpellier, France).27 In order to ensure consistency, sequence data deriving from our previous studies were re-analyzed along with the new data as detailed in the Supplementary Information. Furthermore, interpretation of results for all cases was performed de novo in the context of this study, following the approaches reported in our recent publications.8, 24, 28
In order to explore the possible modifications in the antibody structure caused by a single amino-acid replacement—namely, the arginine (R)-to-tryptophan (W) substitution at position VH FR3-75 to obtain an IGHV1-2*04-rearranging antibody—we performed a molecular dynamics study of the human anti-polyhydroxybutyrate antibody Fv: IGHV1-2*02 IGKV1-39*01/1D-39*01 (Protein Data Bank code 2D7T). Replacement of the VH FR3-75 R (germline for IGHV1-2*02) by the W residue (germline for IGHV1-2*04) was executed by the COOT algorithm.29 Addition of the missing hydrogen molecules to the initial structures and verification of the protonation states were performed using the H++ online server (http://biophysics.cs.vt.edu/H++).30
Molecular dynamics simulations of the wild-type antibody, the mutant and the H-chain alone (in both its wild-type and mutant forms) were performed using the GROMACS 4.0.5 software package,31 using the Amber 99 SB force field.32, 33, 34 The relative orientation of the VH and VK domains was determined according to a previously described structure alignment protocol,35 and their positions through the Cα atoms of certain VH and VK residues are listed in detail in the Supplementary information. The same simulation protocol was followed for all molecules.
IGHV gene repertoire and SHM status
A total of 350 IG heavy chain rearrangements were amplified from 337 cases, as a second rearrangement was amplified in 13 cases. Eight of 337 cases (2.4%) carried two productive rearrangements, whereas for the remaining five (1.5%) the second rearrangement was unproductive, owing to the presence of stop codons or an out-of-frame junction. All unproductive rearrangements were excluded from further analysis and IG gene repertoire analysis was performed for a total of 345 productive rearrangements.
No significant differences regarding IG gene usage or other features of the clonotypic IGs were identified in relation to the starting material (gDNA or total cellular RNA), except for a higher frequency of double rearrangements when the starting material was gDNA, in keeping with what has been reported in other B-cell malignancies, namely chronic lymphocytic leukemia (CLL).36 Among 39 functional IGHV genes identified, the most frequent by far was IGHV1-2 (86/345 rearrangements, 24.9%), followed by IGHV4-34 (44/345 cases, 12.8%) and IGHV3-23 (28/345 cases, 8.1%) (Figure 1, and Supplementary Tables 1 and 2). Collectively, these three genes accounted for 45.8% of the cohort (158/345 rearrangements). Particularly for the IGHV1-2 gene, strong biases were also evident at the level of different polymorphic variants (alleles). In fact, of 86 IGHV1-2 rearrangements, 79 (92%) used allele *04 versus only 7 (8%) that used allele *02 (Supplementary Table 2). As a whole, this large and geographically diverse series confirmed and significantly extended previously published results in smaller series by us and others.4, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22
The identified IGHV gene repertoire biases were more pronounced when focusing on 213 productive IGHV–IGHD–IGHJ rearrangements obtained from 205 cases classified as SMZL based on spleen histopathological findings, according to 2008 WHO classification criteria.3 In particular, within this group (the ‘SMZL-Spleen histology’, SMZL-SH, group), the frequency of the IGHV1-2 gene peaked at 31.9% (68/213 rearrangements), with 66/213 cases (31%) using allele IGHV1-2*04 (Figure 1 and Supplementary Table 3). With this exception, no other significant differences were identified between the cohort and the SMZL-SH group. Noticeably, however, only 1/17 cases with a diagnosis of SDRL on the spleen was found to use the IGHV1-2*04 allele (P<0.03 for comparison to the SMZL-SH group).
Based on the percentage of IGHV gene identity to the germline, 46/345 sequences (13.3%) could be assigned to a ‘truly unmutated’ subgroup (100% gene identity), whereas the remaining sequences (299/345, 86.7%) showed some impact of SHM activity, ranging from minimal to pronounced. Mutated cases were further subdivided according to their mutational ‘load’ into successive bins of 1% difference from the closest germline gene (Supplementary Table 4). For statistical comparisons, sequences with 97–99.9% gene identity were classified as ‘borderline/minimally mutated’ (n=130, 37.7%), whereas those with <97% gene identity as ‘significantly mutated’ (n=169, 49%).
The IGHV gene repertoires of these three mutational subgroups differed significantly. In particular, the IGHV1-2*04 gene represented 43% of all ‘borderline/minimally’ mutated cases, whereas the IGHV4-34 gene accounted for 28% of all ‘truly unmutated cases’ (Supplementary Table 5). Pronounced differences were also identified by examining the mutational profile of subgroups of cases defined by IGHV gene usage. Focusing on selected genes frequent at the cohort level, distinct patterns were recorded, for example: (i) in the subgroup of IGHV1-2*04 rearrangements, 71% were ‘borderline/minimally mutated’; (ii) in the subgroup of IGHV3-23 rearrangements, 75% were ‘significantly mutated’; and (iii), finally, in the subgroup of IGHV4-34 rearrangements, an even distribution was noted among the three mutational categories (Figure 2 and Supplementary Table 6).
The nucleotide sequence of allele *04 encodes for a tryptophan (W) residue at position VH FR3-75 instead of the arginine (R) residue that is encoded by all remaining alleles of the IGHV1-2 gene. To assess the effect of the W residue on IGHV1-2*04-using IG molecules, we performed a molecular dynamics simulation based on the Protein Data Bank-available structure of the anti-polyhydroxybutyrate antibody Fv, which bears the IGHV1-2*02 heavy chain. We introduced the VH FR3-75 R-to-W replacement, selected two plausible conformations for this residue (in terms of lowest energy after minimization among several W rotamers) and considered them distinct for the purposes of the simulation. Indeed, in one conformation (hereafter denoted as R1) the side chain points toward the hydrophobic core of chain VH, whereas in the other (R2) it points toward the solvent, and transitions from one position to the other would require a substantial rearrangement of the surrounding residues. The simulation clearly showed that both ‘mutated’ versions undergo a structural rearrangement, indicated by both analysis of the conformation of residue VH FR3-75 and the overall root mean square distance between the original IGHV1-2*02 and the ‘mutated’ IGHV1-2*04 conformations (Figure 3a). In other words, the allelic mutation resulted in a significant alteration to the relative positions of the two chains (Figure 3b).
The R1 and R2 conformations were different, likely related to the different positioning of the side chain, as outlined previously. Indeed, in the inward-pointing R1 conformer, mobility was limited by the surrounding residues to a larger degree compared with the outwards-pointing R2. Nevertheless, the mobility of the tryptophan (‘IGHV1-2*04-like’) side chain was independent of the presence of the VK domain in the molecular dynamics runs, as similar effects were observed when the VH domain alone was subjected to the same simulation.
Further insights into the potential effect of the R-to-W replacement on the structure of the receptor can be inferred from the relative translation and rotation of the VH and VK domains, both influencing the receptor.35 From comparing the VH/VK rotation angles of the IGHV-1*04 with the IGHV-1*02 sequences during the 30-ns simulation (Figure 3c), it is clear that the IGHV-1*04 conformations, regardless of the rotamer at position VH FR3-75, showed larger values of rotation with respect to the wild-type, reaching up to ∼25 degrees. This results in a repositioning of the receptor residues that would be expected to influence the antigenic functionality and specificity of the molecule. In fact, we observed a substantial closure of the cavity between the complementarity-determining region-3 (CDR3) loops that would likely prevent the binding of the corresponding antigen (Figure 3d).
VH CDR3 characteristics
The most frequent IGHJ gene was IGHJ4 (156/345 cases, 45.2%), followed by IGHJ6 (86/345 cases, 24.9%) (Supplementary Table 7). IGHD genes were identified in 344/345 sequences. Genes belonging to the IGHD3 subgroup predominated, followed by IGHD6 subgroup genes; seven IGHD genes were used in 61.3% of the cohort (Supplementary Table 8).
Rearrangements using the IGHV1-2*04 allele showed biased recombination with just a handful of IGHD3 subgroup genes. Overall, 52/79 IGHV1-2*04 rearrangements (65.8%) used one of the IGHD3-3, IGHD3-9 or IGHD3-10 genes (32, 10 and 10 cases, respectively) (Figure 4). In 45/52 (86.5%) IGHV1-2*04/IGHD3 gene rearrangements, the IGHD gene was used in the same reading frame (RF), leading to VH CDR3s with common ‘IGHD-derived’ amino-acid motifs (Tables 1 and 2).
The VH CDR3 length ranged between 5 and 35 amino acids (median, 18). Significant differences were noted with regard to VH CDR3 length in rearrangements of different mutational status. In particular, ‘borderline/minimally mutated’ cases had longer VH CDR3s (median, 20 amino acids) than either ‘truly unmutated’ or ‘significantly mutated’ cases (median VH CDR3 lengths of 18 and 16 amino acids, respectively); the difference between ‘borderline/minimally mutated’ and ‘significantly mutated’ cases was statistically significant (t-test, P<0.0001). This difference was essentially accounted for by rearrangements using IGHV1-2*04 that had a median VH CDR3 length of 22 amino acids (range, 14–27), thereby differing significantly from all the remaining cases (median VH CDR3 length of non-IGHV1-2*04, 16 amino acids; t-test, P<0.0001) (Figure 5).
We also investigated the isoelectric point (pI) of the VH CDR3s of cases using the IGHV1-2*04 allele and found that the majority showed a basic pI (median pI value of 8.69). This was even more pronounced in IGHV1-2*04 rearrangements using the IGHD3-3 or IGHD3-10 gene in RF3, which showed median pI values of 9.11 and 9.43, respectively.
Molecular features of SHM
At the cohort level, the majority of mutations occurred in CDRs rather than framework regions; furthermore, most CDR mutations were R rather than S, leading to high R/S ratios (>3.0) in CDRs and, vice versa, lower R/S ratios (<3.0) in framework regions (Supplementary Table 9). Notably, however, clustering of R mutations was identified in the VH FR2 of IGHV1-2*04 and IGHV4-34 ‘borderline/minimally mutated’ sequences, leading to very high R/S ratios of 10.25 and 8, respectively (Supplementary Table 10).
Subgroups of cases using the five most frequent IGHV genes were searched for the existence of shared (recurrent) amino-acid replacements introduced by the SHM process across the whole IGHV sequence. The most striking case again concerned IGHV1-2*04 rearrangements, with codon VH FR2-39 standing out as the most frequently targeted position for amino-acid change throughout our series. In fact, 50/75 mutated IGHV1-2*04 sequences (66.7%) were found to carry a change at this codon, albeit of a conservative nature, in that the germline methionine was replaced exclusively by biochemically similar aliphatic residues. A list of the recurrent mutations observed within this subgroup is provided in Supplementary Table 11.
Intraclonal diversification analysis
A total of 79 subcloned IGHV–IGHD–IGHJ gene sequences from four IGHV1-2*04-expressing cases, all belonging to the ‘borderline/minimally mutated’ category, were analyzed. Confirmed mutations, that is, mutations present in more than one but less than all subcloned sequences, documenting intraclonal diversification, were discovered in three cases, whereas the remaining case carried only unconfirmed mutations, that is, mutations present in a single subcloned sequence (Supplementary Table 12). Interestingly, the M-to-I replacement at codon VH FR2-39, which is the predominant shared mutation within the IGHV1-2*04 group, emerged as the most frequent confirmed mutation in the subcloning analysis as well, as it was detected in subclones from all three cases with intraclonal diversification (Supplementary Table 13).
Clustering of VH CDR3 sequences
Using the sequence pattern-based methodology described previously by our group,28 25/345 sequences (7.2%). were assigned to 13 different level zero clusters with restricted (stereotyped) VH CDR3 sequences (Supplementary Table 14). Twelve clusters were composed by a pair of sequences, whereas one cluster was composed of three rearrangements. Six of 13 clusters included IGHV1-2*04/IGHD3-3 rearrangements exclusively; in all such cases, the sequences differed mainly at the 3′ part of the VH CDR3 owing to the usage of different IGHJ genes (Table 3). Four common sequences between level zero IGHV1-2*04/IGHD3-3 clusters were further grouped in a higher level cluster, based on the existence of common amino-acid patterns (Table 3).
Interestingly, of 28 IGHV1-2*04/IGHD3-3 rearrangements with ‘IGHD-derived’ amino-acid motifs within the VH CDR3 (Figure 4), only 10 could be assigned to ‘stereotyped’ clusters. The remaining, despite the ubiquitous presence of short shared ‘IGHD-derived’ motifs within the antigen-binding sites, failed to show significant overall VH CDR3 amino-acid likeness, that is, they did not fulfill the established criteria for VH CDR3 ‘stereotypy’.28
We investigated the IG gene repertoire in 337 cases with primary splenic small B-cell lymphomas of marginal-zone origin, by far the largest published series, consolidated in the context of a multi-institutional collaboration. Thanks to the large size of the present cohort and the application of novel analytical tools, we have obtained a comprehensive picture of the IG repertoire in SMZL, unprecedented by previous studies, with important ontogenetic implications.
Strong IGHV gene repertoire biases were identified, with 45.8% of cases using one of the following three genes: IGHV1-2, IGHV4-34 or IGHV3-23. The IGHV1-2 gene predominated by far, with a rearrangement frequency of 24.9%, peaking at a remarkable 31.9% among cases classified as SMZL based on spleen histopathological findings, following the 2008 WHO criteria.3 Noticeably, we found a marked under-representation of the IGHV1-2 gene among cases with spleen histopathology suggestive of SDRL, in keeping with our previous report.4 On these grounds, we propose that immunogenetic analysis may add discriminative power in the differential diagnosis between SMZL and other primary splenic small B-cell lymphomas of marginal-zone origin assigned by the 2008 WHO classification to the provisional category of SLLU.
The IG gene repertoire biases extended to the usage of polymorphic variants of the IGHV1-2 gene, with allele *04 markedly over-represented compared with the other alleles of this gene. Comparisons with the normal repertoire are hampered by the lack of a truly representative and comprehensive reference data set, especially for normal splenic marginal-zone B cells, with the available studies collectively reporting only 31 productive rearrangements.37, 38 However, no such bias is evident in other B-cell lymphomas, for example, CLL or MCL, where among the rearrangements using the IGHV1-2 gene, alleles *02 and *04 are detected at roughly similar frequencies.8, 24 This restricted usage of allele *04 in SMZL is noteworthy, given that it differs from allele *02 by a single amino acid, and can be considered as a molecular argument for selection.
Eight of 337 cases from our series (2.4%) carried double productive rearrangements, indicating that allelic exclusion may not be absolute in SMZL, similar to what has been reported previously for other B-cell malignancies, most notably CLL.8, 39, 40, 41 Admittedly, allelic exclusion might still be maintained at the level of transcription or translation, or both. However, from an ontogenetic perspective, it is perhaps relevant that expression of dual IG receptors has been reported in autoreactive B cells, probably as an attempt to ‘dilute out’ a strongly autoreactive receptor and, thus, escape clonal deletion.42 Interestingly, in mouse transgenesis models, autoreactive B cells with dual IG receptors accumulate in the marginal zone,43 which might be relevant for SMZL immunopathogenesis.
In previous studies of the IG gene repertoire in SMZL, assignment to the ‘mutated’ IGHV subset was based on the 2% cut-off value for deviation from the closest IGHV germline gene, which is widely used for prognostication in CLL.44 From a biological perspective, this may be an oversimplification, as a low mutational ‘burden’, even a single amino-acid replacement, can be functionally relevant.8, 45, 46 For this reason, we adopted a different approach and separated the truly unmutated cases (100% identity) from the remaining cases showing any level of SHM and, thus, considered as mutated. For statistical purposes, mutated cases were further subdivided according to their IGHV gene mutational status into ‘borderline/minimally mutated’ (97–99.9% identity to germline) and ‘significantly mutated’ subgroups (<97% identity). The 97% cut-off was an educated choice, and, although still essentially arbitrary, allowed to cluster the large majority (71.8%) of IGHV1-2*04 rearrangements within the ‘borderline/minimally mutated’ group.
Given the generally low level of SHM among IGHV1-2*04 rearrangements, it is noteworthy that three of four such cases with ‘borderline/minimally mutated’ status showed intraclonal diversification with restricted distribution, alluding to precisely targeted ongoing SHM activity. This finding indicates that the SHM mechanism may continuously operate in SMZL even after transformation, in keeping with previous observations.18, 23 Although the precise timing of interactions with antigen(s) and their functional implications for SMZL evolution will likely remain difficult to define accurately, a role for persistent antigenic stimulation could be invoked, at least for certain subsets of SMZL.
Altogether, not only is IGHV1-2*04 used by a striking multitude of cases, indicating strong selective pressures, but, in addition, the selective process is more often than not coupled with limited, though restricted, changes in the germline sequence. The clinical significance as well as other potential associations of this distinctive mutational profile would be hard to define accurately on the present material, mainly owing to limitations inherent in a retrospective study as ours (including missing data and heterogeneity of therapeutic regimens), and should be further evaluated in a prospective manner.
From a sequence perspective, IGHV1-2*04 is an outstanding gene: whereas almost all alleles of human IGHV genes (including the remaining alleles of IGHV1-2) carry an arginine (R; basic, hydrophilic) at position VH FR3-75 (R-75), in IGHV1-2*04 this position is occupied by a tryptophan (W; hydrophobic, non-polar). R-75 has been identified as one of a small set of conserved residues in both IGHV and IGKV/IGLV sequences, regardless of the specific gene considered, and therefore is expected to be critical for the structure of the IG fold.47 In this context, substitution of R for W in IGHV1-2*04 may be considered as ‘exceptional’.
In fact, the molecular dynamics simulation indicates that the R-to-W replacement at position VH FR3-75 could result in a substantial alteration in antibody conformation. Notably, in the specific antibody that we used as a model, the replacement led to the closure of the CDR3 cavity. The magnitude of the change observed, along with the strong bias in the usage of IGHV1-2*04 rearrangements, indicate that W-75 likely has an important structural role, and allude to the recognition of specific antigenic epitopes through the conferred structural conformation. It remains to be elucidated if and how the R-to-W replacement at position VH FR3-75 might result in the same significant conformation change in any other antibody structure with different underlying sequence from the one used in our simulation.
Currently, the functional effects of this ‘exceptional’ molecular feature and its potential implications for SMZL ontogeny remain speculative. However, the biased expression of a highly distinctive and potentially structure-altering, germline-encoded VH sequence might be considered as evidence for heavy-chain dominance in the clonogenic IG receptors of SMZL. The fact that the IGHV gene can be the specificity-defining sequence is now supported by structural, functional and transgenesis studies.48, 49, 50 Some of these studies have also indicated the complementary importance of VH CDR3 sequence and length restrictions, likely related to a proper positioning of key structural determinants, or to stability requirements of the heavy chain.51 Are these observations relevant to SMZL?
The IGHV1-2*04 rearrangements carried exceptionally long, electropositive VH CDR3s and showed biased usage of the IGHD3-3 and IGHD3-10 genes, mostly in RF3, resulting in VH CDR3s with common ‘IGHD-derived’ motifs. Despite these shared features, the vast majority of IGHV1-2*04 rearrangements did not fulfill the established criteria for VH CDR3 stereotypy, as defined previously in other lymphoid malignancies, namely CLL,8, 28 mainly because of variable IGHJ gene usage and heterogeneous junctional residues.
Taken together, the molecular features of IGHV1-2*04 B-cell receptors in SMZL argue for antigenic interactions through highly conserved residues, located throughout the VH domain and mainly outside the VH CDR3, thus raising the possibility that superantigens might be implicated in the selection of the clonogenic progenitors. The case for superantigenic interactions is also supported by the examination of SMZL heavy chains using IGHV4-34, the second-ranking gene in our series. All IGHV4-34 SMZL sequences analyzed here maintained the VH FR1-binding motifs for the N-acetyllactosamine antigenic determinant52 intact and, in principle, were able to interact with N-acetyllactosamine-containing self- and exogenous antigens, thus, receiving stimulation signals independent of the classical antigen-binding site.
According to the 2008 WHO classification, the postulated normal counterpart of SMZL is a B cell of ‘unknown differentiation stage’, with ‘the presence of IG gene somatic hypermutations in 50% of cases suggesting exposure to antigen in the germinal center microenvironment’.3 The results presented here may contribute to refining this view. In fact, somatic mutations were detected at a much higher frequency (almost 88% of the cohort), indicating exposure to antigen in the vast majority of cases. Admittedly, seminal questions remain, concerning the identity of the progenitor cells and the nature of the selecting antigens.
Definitive answers to these questions will admittedly be provided only through multidisciplinary examination of the physiology and immune functions of normal splenic marginal-zone B cells combined with detailed histopathological and molecular studies of SMZL and related entities. However, our immunogenetic findings seem to justify the claim that SMZL may have multiple cells of origin, thus reflecting the heterogeneity of B-cell populations residing within the normal splenic marginal zone.53, 54, 55 In addition to the previously postulated GC-derived progenitor, we propose that at least one-third of SMZL cases could derive from various other types of splenic marginal-zone B cells that express distinctive IG receptors. On the evidence presented here, it could perhaps be speculated that these receptors likely reflect the innate usage of a series of genes selected over evolution to provide ‘germline immune memory’.
We express our gratitude to Professor Marie-Paule Lefranc (Laboratoire d’Immunogenetique Moleculaire, LIGM, Universite Montpellier II, Montpellier, France) for valuable support, suggestions and guidance in immunoglobulin gene sequence analysis. The analysis of French cases was supported by grants from the Comité du Rhône de la Ligue Nationale contre le Cancer. The analysis of the Spanish cases was supported by grants from the AECC and the Ministerio de Ciencia e Innovación, Spain (SAF2008-03871, RETICS). The analysis of cases from the Institute of Cancer Research (London, UK) was supported by the UK Cancer Research Fund. The antibody molecular dynamics simulation was supported by a research grant from the Italian Ministry for University and Research (FIRB) and the CARIPLO Foundation, Milano, Italy.
Vasilis Bikos performed research, analyzed data and wrote the paper. Nikos Darzentas and Anastasia Hadzidimitriou performed research and analyzed data. Zadie Davis, Sarah Hockley, Alexandra Traverse-Glehen, Patricia Algara, Alessandra Santoro, David Gonzalez, Manuela Mollejo, Antonis Dagklis and Pascale Felman performed research. George Bourikas, Achilles Anagnostopoulos and Athanasios Tsaftaris supervised research. Fabrizio Gangemi and Massimo Degano performed the antibody molecular dynamics simulation. Emilio Iannitto, Maurilio Ponzoni, Francoise Berger, Chrysoula Belessi, Paolo Ghia, Theodora Papadaki, Ahmet Dogan, Estella Matutes, Miguel Angel Piris and David Oscier provided samples and associated clinicopathological data, and supervised research. Kostas Stamatopoulos designed the study, supervised research and wrote the paper.
About this article
Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)