Analysis of the B-Cell Receptor B29 (CD79b) Gene in Familial Chronic Lymphocytic Leukemia

Total blood samples from 10 different Italian families, including 2 members affected with CLL, were stored at −20°C in the Clinica de la Sapienza (Rome, Italy). Families 1 through 8 have been studied for the VH gene expression of their Ig.21a Families 13 and 14 have been added in this study. All of the patients studied demonstrated typical CD5⁺ B-CLL. They comprised 9 men and 11 women aged from 35 to 95 years who were staged at diagnosis following Binet’ staging22 (stage A, 14 patients; stage B, 4; stage C, 1; in 1 case, stage could not be determined). Two patients progressed to stage B during evolution. At the time of the study, 5 patients had received previous treatment, whereas 15 patients did not receive any previous treatment.

Total DNA was purified from up to 10 mL of whole citrated blood with the QIAamp Blood Maxi kit (QIAGEN, Hilden, Germany). The purified DNA ranged in size up to 50 kb, with fragments of approximately 30 kb predominating. DNA yields ranged from 15 to 600 μg according to the cell densities.

PCR amplification was performed with genomic DNA using Taq DNA polymerase (GIBCO-BRL, Bethesda, MD) and 2 sets of oligonucleotides for the B29 gene. The first one, designed to amplify the Ig domain and the transmembrane region, consisted of the forward primer 5′GTAGTGCTTGTTCGCGGATC3′ (corresponding to nucleotides 2006-2025 in the DNA sequence23) and the reverse primer 5′CTTGTCCAGCAGCAGGAAGA3′ (corresponding to nucleotides 2824-2805). The second one, designed to amplify the intracytoplasmic domain, consisted of the forward primer 5′GATGACAGCAAGGCTGGCAT3′ (nucleotides 3122-3141) and the reverse primer 5′CTCCTGGCCTGGGTGCTCA3′ (3368-3350).

Thirty cycles of amplification were performed under the following conditions: 95°C for 30 seconds, 60°C for 30 seconds (for Ig domain and transmembrane segment) or 65°C for 30 seconds (for intracytoplasmic domain), and 72°C for 1 minute.

Amplification of V_H Ig genes was performed from genomic DNA using 5′ primers designed to anneal to the leader sequence (L_H) of V_H1 to V_H6 families, and a consensus J_H 3′ primer. The L_Hprimers consisted of the following: L_H1: 5′CCATGGACTGGACCTGG3′; L_H2: 5′ATGGACACACTTTGCTMCAC3′; L_H3: 5′CCATGGAGTTTGGGCTGAGC3′; L_H4: 5′ATGAAACACCTGTGGTTCTTCC3′; L_H5: 5′ATGGGGTCAACCGCCATCC3′; and L_H6: 5′ATGTCTGTCTCCTTCCTCATC3′. The consensus J_Hprimer consisted of 5′ACCTGAGGAGACGGTGACCAGGGT3′. In all cases, 30 cycles of amplification (95°C for 30 seconds, 65°C for 30 seconds, and 72°C for 1 minute) were performed.

PCR products were gel purified and subcloned into a TA cloning vector, followed by transformation into INVαF′ cells (Invitrogen, San Diego, CA). Recombinant plasmids were purified and sequence analysis was performed.

Nucleotide sequence data of V_H Ig genes were analyzed and compared with the GenBank database (Bethesda, MD). The complete sequences have been contributed to GenBank for families 1 to 8 (accessions nos. AF087415-AF087430) .

Ten families, each with 2 members affected with CLL, were studied. Clinical and biological data concerning these patients are reported in Table 1. Except for 2 patients (1A and 14B), they all had lymphocytosis, whether treated or not. It can be observed in the 6 families with vertical transmission of the disease that the mean age of the parental generation was 81 years as compared with a mean of 50 years for the second generation. This earlier appearance of CLL in the second generation suggested that genetic factors could influence the development of the disease.

DNA from 20 patients corresponding to the 10 families described above had the B29 gene sequenced. For each B-CLL sample analyzed, 2 classes of B29 DNA clones were identified in the PCR-generated clones corresponding to both B29 alleles. Mutations identified on the B29 DNA sequence are described in Tables 2 and3. In all patients (except patient 3B), 1 allele had the silent mutation TGT→TGC at amino acid position 122 (Cys) in the Ig domain (Table 2), corresponding to a polymorphism of the B29 sequence. In the Ig domain, only 1 other silent mutation was observed (patient 1A), whereas 7 replacement mutations producing an amino acid change were noted in 7 patients: patient 1A at position 56 (Lys → Arg), 13A at position 56 (Lys → Glu), 2A at position 61 (Val → Ala), 4B at position 66 (Tyr → His), 5B at position 72 (Gly → Ser), 3B at position 99 (Ser → Phe), and 2B at position 102 (Glu → Gly). These mutations do not affect known important sites of this domain, namely the 5 extracellular cysteine residues allowing for intrachain and interchain (with Igα) S-S bonds or the 4 glycosylation sites on the Asn residues. Concerning the transmembrane portion, 1 silent mutation was observed in patient 4A at amino acid position 145, whereas 4 replacement mutations at positions 145 (Phe → Ser), 167 (Leu → Pro), 169 (Ile → Leu), and 169 (Ile → Thr) were noted in patients 4B, 14A, 5B, and 7B, respectively. Mutations in the intracytoplasmic domain were detected in only 4 patients (Table 3): 2 silent mutations at amino acid position 192 in patient 6A and 193 in patient 14B and 2 replacement mutations at amino acid position 202 (Asp → Asn) and 223 (Gly →Ser) in patients 2A and 5A, respectively. The Tyr residues, substrates for phosphorylation, were not affected by the mutations. All of the replacement mutations of the B29 gene are described in Fig 1.

The only identical mutation observed between 2 members of the same family was the silent TGT → TGC (amino acid position 122) in the Ig domain. We never observed another identical silent or replacement mutation between both members from the 10 different families.

As we used whole blood for DNA extraction because blood samples were stored as pellets at −20°C and fresh cells were not available, the possibility exists that the amplified B29 gene could be derived from nontumoral cells such as T cells, neutrophils, and monocytes. However, the majority of patients included in this study displayed high numbers of monoclonal B cells. In these conditions, most patients had a remarkable excess of neoplastic DNA used for the PCR amplification and sequencing; thus, any important deletion or insertion affecting the B29 gene should not escape detection.

The composition of the H chain variable domain sequences obtained from the 20 patients with familial clustering of CLL is summarized in Table 4. The V_H distribution (V_H1, 10%; V_H2, 0%; V_H3, 4%; V_H4, 30%; V_H5, 10%; and V_H6, 5%) is close to that corresponding to the normal complexity of the system (V_H1, 22%; V_H2, 6%; V_H3, 43%; V_H4, 22%; V_H5, 6%; and V_H6, 2%), except for V_H2, which is not represented in our familial cases. Among the 18 patients whose V_H genes have been sequenced, 14 cases expressed less than 98% similarity with the germinal counterpart, and 11 of 18 had less than 95% similarity. Among these 11 cases, 8 presented a CDR/FR rate of mutation greater than 0.3, which could suggest, at least for some cases, the existence of an antigen-driven selection process.21a 

The B-CLL lymphocyte differs from the normal CD5⁺ B cell by constant expression of low amounts of surface Ig and CD79b4,5; resistance to infection by the EBV virus, despite expressing the CD21 molecule10; inability to adequately respond when stimulated through the BCR pathway8,9; and overexpression of the bcl-2 protein.12 It remains unknown whether some of the particular characteristics of the B-CLL lymphocyte could be malignancy-related.

Because genetic or familial factors have been suggested in the etiology of CLL19,20 and genetic aberrations were reported to exist at the level of the B29 gene,7 we were interested in studying the B29 gene in several families who had 2 affected members each. The excess risk of CLL for individuals with affected first degree relatives suggests that familial CLL might constitute a useful model to study the pathogenesis of this disease, as has been the case in other neoplastic disorders. However, previous studies in familial CLL cases failed to show the presence of consistent chromosomal or proto-oncogene abnormalities.24,25 We could not perform such karyotypic studies on our patients, because the material we studied here consisted of blood samples frozen at −20°C. Furthermore, the number of cases reporting common HLA phenotypes among affected CLL siblings is small.19,26,27 The study of family histories or family trees from our series shows evidence favoring a vertical transmission, which may be consistent with the expression of an autosomal dominant gene. Anticipation as defined by worsening severity or earlier age at onset with each generation for an inherited disease has been reported to occur in familial cases of CLL. In a recent series of 9 parent-child pairs with CLL,28 8 showed younger ages at onset in the child and showed a mean decline in age at onset of 21 years, with very similar results found in a British study.29 This suggests that dynamic mutation of unstable DNA sequence repeats could be a common mechanism of inherited hematopoietic malignancy, with implications for the role of somatic mutation in the more frequent sporadic cases. In the 6 families from our series in which a vertical transmission was observed, the mean age of the parental generation was 81 years as compared with a mean age of 50 years for the second generation, which is lower than the mean age of 64 years found in French series.30 Because sampling bias is unlikely to explain these findings, these results could point to the existence among familial CLL patients of another genetic mechanism for developing disease.

In B cells, the CD79b protein is supposed to be essential in the transport of sIg to the membrane,16,17 initiation of signal transduction through the BCR,15 and apoptosis.18 Thus, a defect in expression of this protein may play a major role in the pathophysiology of the disease. Recent work indicating the presence of genetic aberrations leading to affected expression and/or function in B-CLL patients suggests that these alterations could have a genetic origin.7 To better define whether these aberrations could play a primary role in CLL leukemogenesis, we have sequenced the B29 gene at the genomic level in 10 different families including 2 affected members each.

Our results demonstrated that some mutations were present either in the Ig, transmembrane, or cytoplasmic domains of the CD79b protein, but we never observed any insertion or deletion in the B29 gene leading to a truncated protein. Indeed, we have not screened normal genomic DNA for mutations in B29 gene, but the few published sequences derived from normal subjects showed the presence of some polymorphism such as the TGT or TGC in AA position 12231 that we frequently found in our series. Thus, because the few mutations described in the present work have not been reported before in normal B29 gene sequence, it is difficult to conclude whether they correspond to new polymorphisms or whether they are CLL related. However, they are not genetically determined, because they never occured at the same position among 2 members of the same family.

Our results are in accordance with those reported by Rassenti et al,32 who studied unrelated patients affected with CLL and also failed to observe deletion or insertion in Ig, transmembrane, or cytoplasmic domains, but only base substitutions. Thus, our familial CLL did not differ from these common CLL cases. However, both our series and that from Rassenti et al32 differ with the results reported by Thompson et al,7 who demonstrated by sequencing cDNA clones generated from B-CLL cell B29 mRNA, that point mutations, insertions, or deletions, largely located in the B29 transmembrane and cytoplasmic domains, affected B29 expression and/or function. This discrepancy with our data could result from a different origin of the 10 families studied here, but the present findings exclude, at least in our ethnic group, the presence of causal mutations. Both alleles were sequenced and genetic alterations were not detected in paternal and maternal alleles. The few mutations observed here did not induce structural abnormalities of the protein, which could explain its low expression in CLL; furthermore, these mutations are not coding for amino acids involved in the intrachain or interchain bonds between the CD79b and CD79a proteins, namely the Cys of the Ig domain, glycosylation sites, or intracytoplasmic sites for phosphorylation. Thus, the function of the CD79b protein would not be predicted to be affected by these mutations.

Analysis of the V_H gene usage among these 10 families affected with CLL showed a higher level of mutation than that described in previous studies concerning unrelated patients,33,34 and the vast majority of the CLL variable domains contained a high degree of somatic mutation and exhibited an excess of replacement mutations in the CDR intervals. These findings suggest that familial CLL cases may preferentially derive from B-cell progenitors that have responded to antigen.21a 

Because our results appear to exclude the possibility that a genetic defect could account for the low expression of the CD79b molecule in B-CLL, it would be interesting to study the possibility that the low expression of the CD79b protein on the B-CLL could result from a mechanism of alternative splicing. This event could result in overexpression of a truncated CD79 transcript form, which lacks exon 3, encoding for the extracellular domain of this molecule, and low if any expression of the complete CD79b transcript. Previous work has shown that both transcripts exist in B cells in normal conditions,35 although activated B cells inverse relative expression of the truncated form.36 Thus, the possibility exists that low expression of CD79b in CLL could result from preferential transcription of the truncated CD79b form. Although this possibility needs to be tested, results from Thompson el al7 indicate that transcripts coding for the complete protein are found in consistent amounts in at least half of the CLL patients they have studied. Most of these patients were expressing low to undetectable amounts of the CD79b membrane protein, indicating that CD79b expression may not be exclusively explained by the absence of complete transcripts.

In these conditions, the possibility that a defect at the posttranscriptional level accounting for the discrepancy between the presence of adequate transcripts and very low expression of CD79b and sIg needs to be considered. This defect could occur at the level of intracellular synthesis, could consist of accelerated proteolysis of CD79b and sIg proteins, or in inadequate transport of the BCR to the cell membrane as described on tolerant B lymphocytes.37