Mutated V_H genes and preferential V_H3-21 use define new subsets of mantle cell lymphoma

Mantle cell lymphoma (MCL) is generally an aggressive and incurable malignant lymphoma with a median survival time of 3 to 4 years.1,2 Most patients are diagnosed with widespread disease with disseminated lymphadenopathy and frequent involvement of the bone marrow (BM), spleen, and gastrointestinal (GI) tract.1,2 Tumor cells typically express CD5, CD19, and CD20 but lack CD23 expression.3 Almost all MCLs exhibit the (11;14) translocation, leading to rearrangement of theBCL1 gene and overexpression of cyclin D1.4,5 The up-regulation of this cell cycle regulatory protein contributes to the pathogenesis of MCL, but alone it is insufficient for tumor progression.6 Despite this characteristic diagnostic marker, the disease entity includes a heterogeneous group of patients with various clinical courses. Few known markers can further define prognostic subgroups within the disease, with the exceptions of the morphologically distinct blastoid variant of MCL (MCL-BV) and the high proliferation index that have been reported to identify patients with aggressive disease.1 7Thus, the identification of more predictors of outcome will be the first step in revealing biologic differences in the disease, which may ultimately be exploited to enable improved treatment strategies.

Analysis of somatic hypermutation of the immunoglobulin variable heavy chain (IgV_H) genes has provided valuable insight into the origin of B-cell lymphomas. This process occurs during a T-cell–dependent immune response in the germinal center (GC) of secondary lymphoid follicles.8 Therefore, a mutated V_H gene in a B-cell lymphoma indicates GC or post-GC B-cell origin, whereas an unmutated V_H gene indicates derivation from a pre-GC B cell that has proliferated independently of T-cell help. In MCL, the postulated cell of origin is a B cell deriving from the mantle zone of a lymphoid follicle that has not entered the GC and, therefore, lacks hypermutation.9 This fact has been questioned because MCL patients with mutated V_H genes have been reported by our group and others.10-13 Recently, we investigated 51 MCL patients and showed that approximately 20% of them displayed hypermutated V_H genes (more than 2% sequence deviation from the corresponding germline gene was defined as mutated); the remaining had unmutated V_Hgenes.13 This led us to suggest that MCL may consist of 2 subsets—one larger unmutated group and one smaller mutated group. Similarly, studies of chronic lymphocytic leukemia (CLL) have convincingly revealed that this disease comprises 2 subsets but the proportion of mutated and unmutated cases is approximately equal.14-16 Although V_H gene mutation status has been shown to be an independent prognostic marker in CLL, with mutated cases having better clinical outcomes, our pilot survival analysis (of 48 MCL patients) did not reveal a significant association between V_H gene mutation status and overall survival in MCL.13 However, the 2% mutation cut-off level empirically chosen for CLL17 may not necessarily be correct for MCL, especially because the true biologic border that distinguishes between mutated and unmutated lymphomas is unknown.

V_H gene use has been of great interest in B-cell malignancies—a biased use of individual V_H gene family members has been shown in several lymphoma and leukemia entities.18-26 It has been hypothesized that B-cell receptors encoded by certain V_H genes may predispose those B cells to malignant transformation as they may have a greater tendency or capacity to proliferate more frequently due to the recognition of many self- or nonself-antigens.19,21,22,27,28 We have recently shown restricted V_H gene use in MCL, demonstrating an overuse of 3 different V_H genes: V_H3-21, V_H4-34, and V_H5-51.13 The 2 former V_H genes have a known, strong association with autoimmunity.29-33 Additionally, we demonstrated that the V_H3-21 gene is overrepresented in CLL with mutated V_H genes26 and is associated with a significantly worse overall survival than other mutated cases.

In the present study, we investigated the prognostic impact of somatic hypermutation and restricted V_H gene use in a large cohort of MCL patients. We could confirm our recent finding that MCL comprises a mutated (16%) and an unmutated (84%) group, but no significant difference in survival was found between the 2 subsets even when choosing different mutation cut-off levels (1%-5%). Analysis of the corresponding unrearranged germline gene was performed in T cells from 5 mutated cases and showed 100% homology to the published sequence, verifying that the incorporated mutations in the rearranged V_H genes corresponded to hypermutations. A novel and interesting finding was that V_H3-21⁺ MCL patients showed almost exclusive use of the V_λ light chain gene, V_λ3-19.

Tumor samples from 110 MCL patients diagnosed between 1982 and 2002 were identified from the registries of the Departments of Pathology at Uppsala University Hospital, Umeå University Hospital, Lund University Hospital, Karolinska Hospital, and Huddinge University Hospital, Sweden, where tumors were registered as either centrocytic lymphoma or MCL. Approval for this study (Ups 01-182) was obtained from the ethical committee at Uppsala University; informed consent was not required since the analyses were performed on retrospective patient material. The only denominator that the patients chosen from the 5 centers shared was the availability of frozen tumor material from which DNA could be successfully extracted and amplified, though we initially included 10 patients from whom paraffin-embedded tissue was used. Besides that, there was no obvious bias in our patient selection, which is also reflected by the typical clinical characteristics and prognoses found in these MCL patients. Tumor material was obtained from lymph nodes (n = 79), BM (n = 9), spleen (n = 8), peripheral blood (PB) (n = 5), GI tract (n = 3), tonsil (n = 3), and other sources (n = 3). MCL diagnosis was based on morphologic and immunophenotypic criteria according to the World Health Organization (WHO) classification34; cyclin D1 overexpression was confirmed by immunohistochemistry in 103 patients, and t(11;14) was identified by fluorescence in situ hybridization (FISH) analysis in the other 7 patients. All MCL patients were diagnosed at the individual institutions and were centrally reviewed by an experienced hematopathologist (C.S.). Medical records were reviewed for all but one patient. Staging was performed according to Ann Arbor staging methods, including computed tomography (CT) of the abdomen/thorax or ultrasound of the abdomen, chest x-ray examination, and BM biopsy/aspiration. Age, sex, stage, nodal/extranodal disease, BM/spleen involvement, blood lymphocyte count, serum lactate dehydrogenase (s-LDH) level, treatment, and disease course were recorded, and the international prognostic index (IPI) was calculated for each age group.35 Survival information was obtained from medical records and local Swedish cancer registries.

DNA was prepared from frozen or paraffin-embedded tumor material using standard protocols as previously described13 or with the QIAamp DNA Mini Kit (Qiagen, Valencia, CA), according to manufacturer's instructions. V_H and V light-chain (V_L) gene family-specific polymerase chain reaction (PCR) amplification was performed as outlined previously.13,36 Monoclonality of the V_H and V_L PCR products was distinguished from a polyclonal product by single-strand conformation polymorphism analysis, as detailed earlier.13 

In most samples (n = 78), clonal products from the V_Hgene PCR were sequenced directly using the BigDye Terminator Cycle Sequencing Reaction Kit version 2.0 (Perkin-Elmer, ABI, Foster City, CA). Cloning of the V_H gene PCR products was performed in 32 patients as previously described,37 and 3 to 10 colonies from each PCR product were sequenced. To investigate the presence of ongoing mutations in patients with mutated V_Hgenes, further subcloning of the amplified rearrangements was performed in 5 patients using the Zero Blunt TOPO PCR Cloning kit, which includes the proofreading enzyme Pfu (Invitrogen, Paisley, United Kingdom). At least 10 colonies were subsequently sequenced from each PCR product. This cloning kit was also used for the analysis of germline V_H gene sequences amplified from T-cell DNA. V_Lgene PCR products of V_H3-21⁺ patients were cloned using the previously described method37 or the Zero Blunt TOPO PCR Cloning kit. All sequences were analyzed using an automated DNA sequencer (ABI 377 or ABI 3700; Applied Biosystems, Foster City, CA).

T cells were isolated from frozen tumor samples from 5 patients with mutated V_H genes (from PB in 3 patients, from lymph node or spleen in the other 2 patients) using a CD4 Positive Isolation Kit (Dynal ASA, Oslo, Norway). Briefly, magnetic beads coated with mouse anti-CD4 IgM monoclonal antibody were incubated with the sample, which allowed the CD4⁺ T cells in the sample to bind to the magnetic beads. Subsequent washing with the use of a magnet enabled isolation of the bound cells.

Primers were designed to amplify the unrearranged V_Hgermline gene for V_H3-23, V_H3-72, V_H3-74, V_H4-34, and V_H4-59 using the Primer3 software program.38 Forward primers hybridized to a region upstream of framework region (FR)1, and the reverse primer hybridized to a region downstream of FR3, hence excluding a rearranged germline V_H gene from being amplified. Primer sequences (5' to 3') were: V_H3-23 forward, AGTTTGGGCTGAGCTGGCTTTTTC; V_H3-23 reverse, AATCTGCATTTGGTGCGTGTGAG; V_H3-72 forward, CCCCAGAATTCCCAGGTGTTTTC; V_H3-72 reverse, AAGGCTCCAGAGTCCTGCAAAAA; V_H3-74 forward, TTGCTGATCAGGACTGCACACAG; V_H3-74 reverse, ACTTGATGAATCAGCCCCAGATG; V_H4-34 forward, ATGGACCTCCTGCACAAGAACAT; V_H4-34 reverse, CAGACCACTGAGCTGTTGAGGAA; V_H4-59 forward, CACCTCTCCATACAAAGGCACCAC; V_H4-59 reverse, GGCGCTGAGCAGCACCTG. The PCR reaction was carried out using the same conditions as for V_H gene family-specific PCR except that annealing temperatures of 65°C for V_H3-23, V_H3-72, V_H4-59 and 64°C for V_H3-74 and V_H4-34 were used. PCR products were cloned, and at least 10 colonies were sequenced per sample.

The obtained sequences were aligned to immunoglobulin sequence databases—Ig BLAST (National Center for Biotechnology Information, USA), V-BASE (MRC Centre for Protein Engineering, Cambridge, United Kingdom), and International ImMunoGeneTics (IMGT).39V_H gene sequences containing mutations resulting in less than 98% homology to the germline gene were defined as mutated. The presence of mutations in hotspot regions was evaluated by analyzing the number of mutations that occurred in the germline sequence motif RGYW/WRCY.40 Ongoing somatic hypermutation was defined as the presence of a replacement mutation in at least 2 of 10 sequences analyzed.

Survival was calculated from the date of diagnosis until last follow-up or death. Kaplan-Meier survival analysis and log-rank analysis were performed to study the prognostic impact of various parameters in our MCL cohort, and P was calculated as 2-sided. Cox proportional hazards analysis was used to study the interrelationship between univariate significant parameters. Fisher exact test with 1-tailed P and McNemar χ² test were used to calculate the significance of differences in V_H gene use between MCL and healthy B cells. The Mann-Whitney U test was applied to calculate the significance of differences in distribution of age at diagnosis between groups. Statistica 6.0 software (StatSoft, Tulsa, OK) was used for all calculations.

We investigated 110 patients; 90 were men and 20 women. Eighty-four patients were diagnosed with classical MCL and 26 with MCL-BV. Median age at diagnosis was 68 years (range, 34-85 years). At follow-up, 56 patients had died and 53 were alive. Median follow-up time (n = 109) was 23 months (range, 0-162 months), and median survival time was 38 months (range, 2-162 months). Adequate staging was performed for 102 patients, revealing that 9 were in stage I-II and 93 were in stage III-IV. One hundred patients had nodal disease; only 6 did not have any nodal involvement (information from 4 patients lacking). Of these 6 patients without nodal involvement, 3 had BM/blood involvement as the sole site. Seventy-four patients had BM involvement, 24 had splenic involvement, 39 patients had extranodal involvement, and 32 had leukemic disease (of which only 8 had lymphocyte counts exceeding 20 × 10⁹/L). The distribution of IPI scores (n = 89) was as follows: 0 to 2, 44%; 3 to 5, 56%. The median s-LDH value (n = 93) was 6.6 μkatal (μkat)/L (range, 3.8-29 μkat/L; normal upper limit, 6.7 μkat/L). Treatment was recorded for 106 patients; 4 patients were untreated, 2 received radiotherapy alone, and the rest received 1 or more chemotherapy regimens or antibody treatment. Most patients received CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone)/CNOP (cyclophosphamide, mitoxantrone, vincristine, prednisone) treatment (n = 78), but chlorambucil/COP (cyclophosphamide, vincristine, prednisone) (n = 18), rituximab (n = 36), and nucleoside analogues (n = 40) were also used. Nineteen patients underwent high-dose chemotherapy with autologous stem cell transplantation (ASCT). Major clinical factors are summarized in Table 1.

One hundred seventeen IgH gene rearrangements were amplified and sequenced from 110 patients, including 7 patients with double rearrangements. Sequence results of the first 51 patients have been published.13 Of the 117 sequences, most had in-frame rearrangements, whereas 9 sequences had a stop codon or an out-of-frame rearrangement. Four of these nonfunctional sequences were detected in patients who had a double rearrangement though the other rearranged allele was in-frame. Of the other 5 nonfunctional sequences, 2 were V_H3-21⁺ rearrangements. Percentages of V_H gene family use of the 117 sequences were as follows: V_H1, 14%; V_H2, 3%; V_H3, 50%; V_H4, 25%; V_H5, 7%; V_H6, 2%. The most frequently used V_H genes were V_H3-21 (18%; n = 21), V_H4-34 (16%; n = 19), V_H3-23 (8%; n = 9), V_H5-51 (7%; n = 8), and V_H1-8 (6%; n = 7). In Table2, we compared the percentage use of these V_H genes with the use reported in CD5⁺ B cells from the PB of 2 healthy elderly donors41 and in PB B cells from 5 healthy elderly donors.42 This comparison revealed a higher frequency of the V_H3-21 gene and the V_H4-34 gene in MCL than in controls (P < .001). Use of the 2 most frequent V_Hgenes is referred to in “Discussion” in terms of percentage of 110 patients—19% for V_H3-21 and 17% for V_H4-34. The most commonly used diversity (D) genes were D6-13 (12%), D2-2 (10%), and D3-3 (9%). Percentages of use of the joining (J_H) genes were J_H1, 0%; J_H2, 3%; J_H3, 8%; J_H4, 34%; J_H5, 10%, and J_H6, 33%. The D and J_H genes could not be assigned for 8% and 12% of the sequences, respectively.

Of the 110 patients, 18 (16%) had somatic hypermutations in their V_H genes using the 2% cut-off level (Table3), and 92 (84%) were unmutated. Three patients with a double rearrangement had 1 mutated and 1 unmutated sequence; hence, they were regarded as MCL patients with mutated V_H genes (patients 41, 60, and 87). The number of mutations ranged from 5 to 15, corresponding to 97.8% to 93.3% homology to the germline gene, respectively. Five mutated V_H gene rearrangements were subcloned to investigate the presence of ongoing somatic hypermutation, but none of the mutated sequences showed evidence of this. Mutated V_H genes were also evaluated for the percentage of mutations that occurred in the known hypermutation hotspots, the RGYW/WRCY motifs. Of the total number of mutations (n = 121) in the 18 patients, 50% (n = 60) were in these known hotspots. In addition to the 18 MCL patients with more than 2% mutations, another 17 patients contained 1% to 2% mutations in their V_H genes, corresponding to 3 or 4 mutations, which may also represent somatic hypermutation. None of the V_H3-21⁺ MCL patients had mutated V_Hgenes (Table 4).

Immunophenotype data on the V_H3-21⁺ MCL patients showed that all but one patient (patient 58) expressed λ light chains (Table 4). V_L analysis was performed on all V_H3-21⁺ MCL patients, and in-frame light-chain rearrangements were detected and sequenced for 17 patients. Sixteen patients used the V_λ3-19 gene, whereas 1 used V_λ9-49. For 2 V_λ3-19⁺ patients, a second rearrangement was also detected; both were out-of-frame (patient 54, V_λ9-49; patient 97, V_λ2-11). Two out-of-frame rearrangements, V_λ7-43 and V_λ2-23, were detected for patient 23. The most frequently used J_λ genes were J_λ2/3 (sequence analysis did not distinguish between these genes in most patients because they are homologous). Use of V_λ3-19 in combination with the J_λ2/3 gene occurred in 11 patients. None of the V_λ genes had more than 2% mutations.

In 5 patients with mutated V_H genes displaying 96% to 97% homology to the published germline gene, we performed further analysis of the germline DNA from T cells obtained by cell sorting. In all 5 patients, V_H gene sequences from the T cells were 100% identical to the germline gene. For example, 1 of the 5 patients had a clonal V_H3-23 rearrangement in the MCL sample, and V_H gene analysis showed 6 mutations between FR1 and FR3, whereas the unrearranged V_H3-23 germline gene from the T cells of the same patient did not have any mutations. This demonstrated that the mutations in the rearranged V_H gene represented true hypermutations.

The impact of clinical and molecular parameters on survival in MCL was tested using the log-rank test (Table5). Patients 68 years and younger at diagnosis (n = 53) had significantly better survival times than those older than 68 years (n = 56); median survival times were 49 months (range, 2-162 months) and 23 months (range, 2-109 months), respectively. Other factors found to be significant predictors of survival were IPI and s-LDH level, whereas no significant differences with reference to overall survival were recorded between sex, classical MCL versus MCL-BV, stages I-II versus III-IV, nodal disease versus nonnodal disease, or the involvement of tonsil/spleen/GI tract versus noninvolvement. No association was found between mutation status and overall survival when cut-off levels of 1%, 1.5%, 2%, 2.5%, and 5% for defining a mutated case were applied. Figure1 shows the Kaplan-Meier curve comparing mutated and unmutated cases using the 2% cut-off level. Correlations were carried out between patients using specific V_H genes and the remaining patients. Significantly longer overall survival was found in V_H3-21⁺ patients compared with the rest of the material (log-rank test, P = .030); the V_H3-21⁺ subset had a median survival time of 53 months (range, 5-146+ months) compared with 34 months (range, 2-162 months) for the remaining patients (Figure2). This subset had a lower median age at diagnosis (median, 62 years; range, 44-85 years) compared with the rest of the MCL patients (median, 70 years; range, 34-84 years), but the Mann-Whitney U test did not reveal any significant difference regarding the ages of the subsets (P = .15). Median follow-up time for the V_H3-21⁺ patients was 33 months (range, 0-146 months), whereas the median follow-up time for the remaining MCL patients was 20 months (range, 0-162 months). Detailed clinical characteristics of the V_H3-21⁺ patients are shown in Table6. There was no prognostic association with any of the other 4 frequently used V_Hgenes—V_H4-34, V_H3-23, V_H5-51, and V_H1-8.

All parameters were tested in a univariate fashion using Cox proportional hazards analysis to determine the effect on prognosis. Age (P < .001), IPI (P = .008), and s-LDH level (P = .003) were all significant. When these factors were included in multivariate analysis, only age and s-LDH level retained their importance (P = .003 and P = .038, respectively).

For many years, it has been assumed that MCL harbors unmutated V_H genes and therefore originates from naive B cells. However, in a recent report from our group investigating 51 MCL patients, we showed that approximately 20% of patients with MCL had somatic hypermutations (more than 2%) within their V_Hgenes.13 In this follow-up study, we extended our cohort to 110 patients, and we demonstrate that MCL comprises a mutated (16%; n = 18) and an unmutated (84%; n = 92) group. These data shed light on the postulated cell of origin of MCL because the presence of somatic hypermutations in a subset of MCL implies that these lymphomas may originate from a B cell that has been exposed to the GC environment. To ensure that the observed mutations truly represent somatic hypermutations and not single nucleotide polymorphisms, we analyzed the unrearranged V_H gene in T cells from 5 selected patients whose corresponding rearranged clonal V_Hgene carried mutations in the range 2.2% to 3.6%. For each of these 5 patients, the unrearranged V_H gene was 100% homologous to the published germline sequence, confirming that the mutations corresponded to hypermutations in these patients. Additionally, analysis of the amount of mutations that occurred in the known hypermutation hotspot regions (RGYW/WRCY motifs) showed that 50% of the total number of mutations (n = 121) were in these sequence motifs, further substantiating the data. We also investigated whether there were ongoing somatic hypermutations but did not find evidence of this in the 5 mutated cases analyzed. This indicates that the clonal B cells in MCL patients with mutated V_H genes are not under prolonged antigenic stimulation in the GC and probably derive from a post-GC B cell. We believe that most MCLs derive from pre-GC B cells residing in the follicular mantle but that a small subset has been exposed to the GC environment.

An alternative explanation for mutated V_H genes could be that the precursor lymphoma cells have undergone hypermutation outside the GC. Recently, it was reported that patients with X-linked hyper IgM syndrome, which is characterized by a mutation in the CD40Lgene and the inability to form GCs, can still have a subset of B cells (CD27⁺IgM⁺IgD⁺) with mutated V_H genes.43 This is suggestive of the existence of a second diversification pathway to acquire hypermutations without the classical interaction between B and T cells. The rate of somatic hypermutations in these CD27⁺IgM⁺IgD⁺ B cells was generally low (approximately 1%-2%) in patients with hyper IgM syndrome. However, little is known about this separate route, and it remains to be investigated whether different B-cell subsets, such as MCL precursors, acquire hypermutations by this process.

In CLL, V_H gene mutation status may be used to group patients into 2 prognostic subsets, in which patients with mutated V_H genes have an overall survival time approximately twice as long as that of patients with unmutated V_Hgenes.14-16 In contrast, we could not find any significant differences in median survival times when comparing mutated and unmutated MCL at the 2% cut-off level. This mutation border, which has been empirically derived for CLL to avoid counting Taq polymerase errors and polymorphisms as hypermutations,14 may, however, not be appropriate for defining a mutated case in MCL. There is a large difference in the percentage of mutated cases in each disease (approximately 40%-50% in CLL vs 16% in MCL), and the mean mutation percentage of the mutated group is lower in MCL than in CLL14 (3% vs approximately 6%). In addition to the 18 patients with mutated V_H genes, another 17 patients had V_H genes with a mutation rate of 1% to 2%, which may also represent mutated MCL. Therefore, we performed survival analysis using other percentage cut-off levels for defining a mutated case (1%-5%), but similar results were obtained using different borders. Hence, in our cohort of MCL patients, V_H gene mutation status was not prognostically useful. Interestingly, it has been reported in an abstract by Garand et al44 that patients with nonnodal MCL (patients without clinical lymphadenopathy) have mutations more frequently (19 of 34 patients) than patients with nodal engagement (5 of 31 patients). They stated that survival did not differ significantly between mutated and unmutated MCL but that there was a tendency toward improved survival in the mutated subset. They also showed that patients with nonnodal MCL had a less aggressive disease than nodal patients.44 The proportion of MCL patients without lymphadenopathy seems generally small. Only 6 patients in our cohort had nonnodal disease, all of which were unmutated. Further analysis of this subset would be of interest to determine whether mutation status has a greater clinical impact within this rare entity.

Biased use of individual V_H genes has been demonstrated in several entities of B-cell lymphoma, as mentioned. Restricted use of the V_H4-34 gene in MCL has been indicated through an immunohistochemical detection method.21 This gene has also been found overused in a small series of MCL patients.12In our earlier study, we confirmed the overrepresentation of the V_H4-34 gene in MCL but could also show restricted use of V_H3-21.13 In this report, the 2 V_Hgenes were still overused, with V_H3-21 representing 19% of patients and V_H4-34 representing 17% of patients. In addition, our finding of skewed V_H3-21 gene use was recently supported by other groups, who have found increased use of V_H3-21 in MCL.45-47 Preferential use of the V_H4-34 gene has been found in several other B-cell malignancies, such as CLL, diffuse large B-cell lymphoma, and primary central nervous system (CNS) lymphoma,18,21,23,24 but this V_H gene has also been found to be involved in autoimmunity, encoding antibodies to self-antigens31,32 and nonself-antigens.21,28 The current hypothesis in lymphoproliferative disease is that B cells using specific V_H genes encode immunoglobulin molecules that may be stimulated by various antigens, resulting in an increased proliferation rate that makes them more susceptible to transformation.19,21,22,27 The V_H3-21 gene has also been associated with autoimmunity—it has been shown to be involved in the production of autoantibodies in rheumatoid arthritis and Sjögren syndrome.29,30 We have recently reported a biased use of the V_H3-21 gene in mutated CLL.26 Patients showed specific phenotypic and genotypic features, such as clonal λ light chain expression, restricted V_λ gene use (V2-14), and short, highly homologous CDR3s. Survival analysis also revealed that this subset of patients had worse overall survival than other mutated CLL.

In contrast to CLL, V_H3-21⁺ MCL patients (n = 21) had unmutated V_H genes and did not show any specific common features within CDR3. In parallel with CLL, they showed λ light-chain expression in all but one patient, and sequence analysis of 16 of 18 patients showed a strikingly skewed V_λ gene use of the V_λ3-19 gene (Table 4). This intriguing finding of restricted V_λ gene use in V_H3-21⁺ MCL implicates a possible role for antigen(s) in the pathogenesis of this subset of MCL. V_H3-21⁺ patients also had longer survival than the rest of the MCL patients, with a median overall survival time of 53 months versus 34 months (P = .03; Figure 2). Although these patients tended to be younger at diagnosis, they had no specific clinical characteristics that could account for the improved prognosis (Tables 1, 6). V_H3-21⁺ patients were typical MCL patients with extensive nodal and extranodal disease, and could not be considered indolent because they had the usual spectrum of adverse prognostic factors. However, a large proportion of patients in this group underwent ASCT, which could be one reason for their prolonged survival (Table 1). This still does not explain, however, why these patients had a tendency to present with MCL earlier than the other patients. Indeed, our molecular data implicate that V_H3-21⁺ patients constitute a new biologic entity considering the highly biased V_H/V_λuse, but it remains to be evaluated in prospective studies whether they also represent a separate clinical subset with a different prognosis.

We thank Carin Backlin and Eva Pallin for their skillful technical assistance.