Characteristics of the Immunoglobulin Heavy Chain (IgH) Rearrangement in Multiple Myeloma (MM): An Analysis On 308 Cases.

Characterization of the IgH receptor provides useful insights to understand the pathogenesis and natural history of lymphoid tumors. For example the recognition of stereotyped clusters of immunoglobulin receptors has been a major step forward to understand the pathogenesis of chronic lymphocytic leukemia (CLL). IgH rearrangements have been less extensively investigated in MM, mostly because of lack of large databases of IgH sequences. At our Institution a large number of MM patients has undergone IgH sequencing for minimal residual disease (MRD) evaluation. This database has been merged with MM IgH sequences available in the literature, resulting in 308 MM sequences which have been employed to comprehensively investigate the characteristics of the IgH rearrangement in this tumor.

131 IgH genes from MM patients were amplified and direct sequenced at our Institution (mostly for MRD detection purposes) from specific cDNA obtained at diagnosis, as already described (Voena et al, Leukemia 1997). 177 MM IgH sequences were derived from published databases (NCBI and EMBL). To further characterize the IgH repertoire, we have then compared the MM complementarity-determining region 3 (HCDR3) amino acid (AA) sequences to a panel of productive, non redundant 27413 HCDR3 AA sequences retrieved from public databases (EMBL, NCBI, IMGT/LIGM-DB: 25655 sequences) and from our unpublished laboratory database (1758 sequences), including sequences from malignant B-cell clones (3197 CLL, 1217 lymphomas) and from non malignant repertoire (2685 autoreactive, 4408 immunederegulation/immunodeficiency, 15429 normal and 477 phage display libraries). All the sequences have been analyzed using the IMGT database and tools (Lefranc et al., Nucleic Acid Res. 2005; http://imgt.cines.fr/) to identify IGHV, IGHD and IGHJ gene usage, to assess the somatic hypermutation (SHM) rate and to identify HCDR3 AA sequences. HCDR3 AA sequences were aligned together, in search of subsets of stereotyped receptors, using the ClustalX 2.0 software (http://www.clustal.org/). To define subsets of stereotyped IgH receptors, we followed the criteria proposed by Messmer et al. (J Exp Med. 2004) and Stamatopoulos et al. (Blood, 2007).

No significant differences were noted between our Institutional and published MM databases. Overall IGHV usage in MM appeared in keeping with the normal B cell repertoire with predominance of IGHV3 family (53.9%) followed by IGHV4 (slightly under-represented in MM, 17.2% vs 23.2%, p=0.02) and IGHV1 (12%); besides an over-representation of IGHV2 (7.8% vs 2.3%, p<0.001) was noticed. A modest but significant (p<0.05) over-representation of the IGHV3-9 (5.2% vs 2.6%), IGHV3-21 (4.5% vs 2%), IGHV5-51 (4.5% vs 2.2%) genes and under-representation of the IGHV3-23 (8.1% vs 12.2%) and IGHV4-34 (1% vs 6.5%, p<0.001) were observed. IGHD and IGHJ followed a distribution similar to that of normal IgH repertoire. The median SHM rate was 7.5% (range 0-28%): interestingly we found one single patient with 100% identity to the germline sequence and only three patients with >98% identity. The median length of the HCDR3 was 15 AA (range 7-29), again in line with normal IgH repertoire. Intra MM search for HCDR3 similarity showed no association which met minimal requirements to define stereotyped receptors. The comparison of MM sequences with non MM database showed that 98% of MM sequences are unrelated to known CLL subsets. Among the remaining, 3 MM sequences could be assigned to previously identified CLL subsets (namely n. 25, n. 37 and n. 71 according to Murray et al., Blood 2008) whereas 2 MM formed 2 novel provisional subsets with CLL. When HCDR3 MM were compared to the database of non MM/CLL HCDR3, similarities were found with clones from normal B cells, while similarities with autoreactive clones and other B cell malignancies were sporadic.

The analysis of the largest database of MM IgH sequences so far reported indicate the following: 1) Family usage in the MM IgH repertoire follows a nearly physiological distribution apart for modest skewing of a number of IGHV genes; 2) MM specific HCDR3 clusters do not occur to a frequency detectable with currently available databases; 3) The vast majority of MM sequences are not related to known CLL clusters; 4) MM IgH sequences share more similarities with normal IgH sequences compared to those derived from pathological B lymphoid cells. Thus to state of current knowledge there is little evidence in favor of an antigen driven pathogenesis for this neoplasm.

No relevant conflicts of interest to declare.