Splenic Marginal Zone B-Cell Lymphoma: Clinical Clustering of Immunoglobulin Heavy Chain Repertoires.

Abstract

Abstract Splenic marginal zone B-cell lymphoma (SMZL) is a rare clinical and pathological entity recognized by the WHO classification. SMZL usually presents with isolated splenomegaly, bone marrow involvement, and leukemic picture. Nodal disease is generally rare at diagnosis. When lymphocytes with villous projections are found in peripheral blood, the disease is termed splenic lymphoma with villous lymphocytes (SLVL). High HCV-seroprevalence is frequently reported in SMZL. The disease commonly pursues an indolent course; however, about a third of pts follow a more aggressive course. SMZL is also heterogeneous with respect to the preferential usage of IgVH genes, as well as the percentage of mutation load. No previous studies correlated IgVH rearrangement features with the clinical characteristics of SMZL. The aim of the present study was to determine the tumor-related IgVH gene rearrangement and compare the gene usage and the mutation status with clinical features of 59 pts. Diagnosis was made according to the WHO classification and to the diagnostic criteria for SMZL (Matutes et al., Leukemia 2008). Paraffin sections from lesional tissues (14 spleens, and 59 bone marrow trephines) were available for all pts. Histology and blood smears were reviewed. Total RNA was extracted from PB (n=13) or BM (n=46) mononuclear cells. IgVH gene rearrangements were amplified using 6 family-specific VH leader primers coupled with a 3′ heavy chain joining (JH) primer or with a constant region of cμ chain primer. The sequence obtained from direct sequencing of the clonal PCR product was compared with germline in the IMGT database. Sixty VDJ rearrangements were amplified and 54 were functional. VH1 family was used in 19 rearrangements (32%), VH3 family in 33 (55%) and VH4 family in 8 (13%). The most frequent VH genes were VH1-02 (n=13), VH3-23 (n=15), VH3-30 (n=7) and VH4-34 (n=5) (67% of all sequences). VH was unmutated in 25%. DH segments were assignable in 56/60 rearrangements (all of the VH unmutated and 41/45 of VH mutated). The most frequent DH families were DH2 (n=9) and DH3 (n=21); the most frequent DH genes were DH3-03 (n=8), DH3-22 (n=6) and DH2-02 (n=6). Most rearrangements used JH4 family (n=21), JH5 (n=8), and JH6 (n=19), accounting for 80% of all rearrangements. JH4b (n=14), JH6b (n=10), and JH6c (n=8) were the most common JH genes. Median length of HCDR3 region was 17 aa (range 11–32). For antigen selection analysis P value was < 0.05 in 41% for CDRs and in 55% for FRs. No correlation was found among VH and DH families (p=0.1), among VH and JH families (p=0.09) and among DH and JH families (p=0.4). Forty-two % of pts were unmutated in VH1 family, 15% in VH3 family and 25% in VH4 family (p=0.09). Unmutated cases were 7% in VH3-23 group, 46% in VH1-02 group and 14% in VH3-30 group (p=0.03). For clinical correlations we considered only the 54 functional rearrangements. Villous lymphocytes >10% were detected in 53% of VH1 pts, in 57% of VH4 pts, and in 17% of VH3 pts (p=0.01). Villous lymphocytes >10% were detected in 21% of VH3-23 pts, in 50% of VH1-02 pts and in no VH3-30 pt (p=0.05). Liver involvement was present in 9% of VH3-23 pts, in no VH1-02 pt and in 50% of VH3-30 pts (p=0.009). HCV serology was positive in 7% of VH3-23 pts, 17% of VH1-02 pts and 50% of VH3-30 pts (p=0.04). Serum MC was detected in 50% of VH3-23 pts, in 9% of VH1-02 pts and in 17% of VH3-30 pts (p=0.06). BM was involved in 86% of VH3-23 pts, in 100% of VH1-02 pts and in 50% of VH3-30 pts (p=0.02). A sinusal localization was detected in 67% of VH3-23 pts, in 20% of VH1-02 and in 100% of VH3-30 (p=0.03). Unmutated status was significantly related to a higher percentage of BM infiltration (p=0.003). The proportion of intermediate and high risk patients according to the SMZL score (Arcaini et al. Blood 2006) was higher in the unmutated respect to the mutated group (69% vs 32%, p=0.05). After a median F-UP of 3 yrs, 10 pts died (7 of lymphoma, 3 of causes non related to lymphoma). The median OS was 12 yrs and the median PFS was 2 yrs. OS did not differ among VH families (p=0.9). Median PFS was 2 yrs for mutated pts and 1 yr for unmutated pts. We performed clustering of pts by applying a 2-means clustering algorithm, using as parameters nodal disease, villous lymphocytes >10%, HCV serology positive, BM involvement: in cluster 1 VH1 family pts accounted for 20%, VH3 family for 71%, VH4 family for 9%; in cluster 2 VH1 family pts accounted for 47%, VH3 family for 29% and VH4 family for 24% (p=0.01). In conclusion, VH rearrangement analysis in SMZL reveals a non-random preference for VH1-02, VH3-23 and VH3-30 genes, whose use differs according to distinctive clinical features at presentation.