Leukemia (1998) 12, 601–609 1998 Stockton Press All rights reserved 0887-6924/98 $12.00 http://www.stockton-press.co.uk/leu Characterization of the immunoglobulin light chain variable region gene expressed in multiple myeloma H Kiyoi1, K Naito2, R Ohno3, H Saito1 and T Naoe1
1Department of Infectious Diseases, Nagoya University School of Medicine, Nagoya; 2Department of Medicine, Komaki City Hospital, Komaki; and 3Department of Medicine III, Hamamatsu University School of Medicine, Hamamatsu, Japan
We studied the organization, diversification and clinical sig- the usage of VH, D, and JH segments, the degree of N nucleo- nificance of the immunoglobulin light chain (IgL) variable tide addition, and the rate of somatic mutations.7–13 However, region genes expressed in 17 -chain and 16 -chain producing little is known of the organization and diversification of the IgL multiple myeloma (MM) samples. The V genes from 31 MM samples had over 84.9% homology to the known germline V/ variable region gene on -chain. On the other hand, multiple genes, whereas one V and one V gene had only 75.5% and myeloma (MM) has a variety of clinical features relating to 65.9% homology, respectively. While all five J segments were clinical stage, prognosis and complication.14 Because renal equally used, only J-1 or J-2/3 was used among seven J dysfunction in MM reportedly depends on the physico- segments. N nucleotide addition was found at two V -J and chemical character of the Bence–Jones protein (BJP),15 the IgL five V-J junctions. The -chain complementarity determining region (CDR)-3 was longer and more variable than the -chain sequences expressed in MM may have clinical significance in CDR-3 mainly due to junctional flexibility of V and J seg- this disease. ments. Somatic mutations were more frequent in the J than In this study, we determined the IgL variable region the J segments, and were distributed in the CDR-3 as well sequences from 17 -chain and 16 -chain producing MM as the frame work region (FWR)-4. Those of the J segments, samples by means of the reverse-transcriptase polymerase however, were limited to FWR-4. In FWR-4, replacement chain reaction (RT-PCR) method, and analyzed their organiza- mutations were clustered at codon 106 of -chain and 103 of -chain. Thus nucleotide mutation or conservation was depen- tion, diversification and association with each respective IgH dent on position, indicating a structural necessity of IgL for the variable region sequence and clinical features. development of myeloma cells in addition to a non-random dis- tribution of mutations. There was no characteristic IgL sequence according to the isotype of M-protein, clinical stage Materials and methods or renal complication. Keywords: Ig ;Ig ; multiple myeloma; somatic mutation Patients and MM samples
The clinical features of MM patients are summarized in Introduction Table 1. Clinical staging was classified according to the cri- teria of Durie and Salmon.14 None of the patients had received Immunoglobulin (Ig) molecules are composed of a unit that any chemotherapy before sample collection. After obtaining consists of two identical heavy chains (IgH) and two identical informed consent, BM samples were collected from 33 MM light chains (IgL). Amino-terminal domains of the IgH and IgL patients; 17 -chain and 16 -chain type. Mononuclear cells are highly varied and pair to form the antigen-binding site. were separated by Ficoll–Hypaque density gradient centrifug- The enormous diversity of each chain is derived from the ation from the BM samples. assortment of a particular V, (D in the heavy chain), and J gene segment, junctional diversity, N nucleotide addition and 1–3 somatic mutations. Ig gene rearrangements occur in a non- Sequencing of the IgL variable region gene random fashion during early B cell development.4 Usually the IgH loci rearrange prior to the IgL loci. Successful VH-D-JH Total cellular RNA was prepared from the mononuclear cells gene rearrangement in either of the two IgH alleles, which using the guanidium isothiocyanate/cesium chloride method leads to expression of protein with surrogate light chain on as described.16 cDNA was synthesized from the total RNA the cell surface, suppresses further rearrangement in the IgH using oligo d(T) priming and Moloney murine leukemia virus locus (allelic exclusion), and initiates the rearrangement of the reverse transcriptase (SuperScript II; GIBCO BRL, Gaithers- 5 IgL loci. In mammalian cells, there are two isotypes of IgL burg, MD, USA) according to the manufacturer’s recommen- molecules: kappa ( ) and lambda ( ). A mature B cell dations. The -or-chain variable region genes were ampli- expresses either isotype (isotypic – exclusion). Previous fied by polymerase chain reaction (PCR) using consensus V studies have demonstrated that the -loci rearrange earlier and C primers, or consensus V and C primers. Each primer and/or more frequently than the -loci during B cell develop- was synthesized according to the consensus sequence at the ment. However, analysis of -locus-disrupted mice has shown 5Ј portion of the respective exon, which was determined from that rearrangements of the -loci are not essential for those of published germline sequences,17–37 and contained XbaI sites 6 the -loci. in the consensus V and V primers or HindIII sites in the Studies on IgH variable region gene sequences at various consensus C and C primers, to allow ligation of the ampli- stages of development have revealed stage-specific trends in fied fragments into recombinant vectors; consensus V:5Ј- GGTCTAGATGAC(C/G)CAG(T/A)CTCCA-3Ј, consensus V: 5Ј-AGTCTAGA(T/C)TGAC(T/G)CAG(G/C)(C/A)(G/C)CC(C/T)T- Correspondence: T Naoe, Department of Infectious Diseases, Nagoya Ј Ј Ј University School of Medicine, 65 Tsurumai-cho, Showa-ku, 3 ,C:5-GGAAGCTTAAGACAGATGGTGCA-3 , and con- Nagoya 466, Japan; Fax: 81 52 744 2801 sensus C:5Ј-CAAAGCTTGG(T/C)GGGAACAGAGTGA-3Ј. Received 29 April 1997; accepted 12 May 1997 The amplified fragments were separated on 8% polyacryl- IgL chain sequence of multiple myeloma H Kiyoi et al 602 Table 1 Clinical profiles of patients with MM and characterization of the IgL variable region genes
UPN Age Sex Iso- Stage BJP V /V V /V V /V J /J Mutation Deletion Addition Length of VH JH JH (years) type family germ- mutation gene of J/J of J/J of N CDR-3 family gene mutation line (%) (%) (nucleo- (nucleotide) (amino (%) tide) acid)
Kappa BK1 51 M BJP III A P I L5a 2.7 1 7.7 0 0 9 — — — BK2 67 F BJP III A P III L2 3.4 4 7.7 0 3 10 — — — BK3 43 M BJP III B P II A2 2.9 5 2.7 2 0 9 — — — AK1 57 F IgA II A N I L5 15.1 2 0 1 0 9 II 4 2.4 AK2 74 F IgA III A N I L4 11.5 4 2.7 2 0 9 III 4 2.7 AK3 55 F IgA II A P I L8 8.8 3 7.9 1 0 9 IV 5 7.1 AK4 60 M IgA I A N IV B3 9.7 3 7.7 0 0 9 I 5 6.3 AK5 67 M IgA II B P III A27 4.9 5 0 0 0 9 ND ND ND AK6 67 M IgA III A N V B2 24.5 1 0 1 0 9 III 5 5.6 DK1 74 F IgD III B P I L5 3.4 1 5.1 0 0 9 IV 4 6 GK1 69 M IgG I A N I L12 7.7 3 8.3 3 0 9 III 6 7.1 GK2 73 F IgG III A P III A27 9.5 1 5.3 1 0 9 I 4 2 GK3 57 F IgG II A N III L6 7.3 5 0 0 0 9 I 4 4.4 GK4 77 F IgG III A P I L8 12.6 1 2.7 2 0 9 IV 5 5.4 GK5 66 M IgG III B P I L12 4.2 2 2.6 0 0 9 III 4 7.1 GK6 42 F IgG III B P II A17 3.6 4 5.3 1 7 10 III 4 0 GK7 58 F IgG I A N II A3 4.3 5 0 2 0 9 ND ND ND mean ± s.d. 7.0 ± 3.9 3.8 ± 3.2 0.94 ± 0.97 9.1 ± 0.3 4.7 ± 2.4
Lambda BL1 69 M BJP II A P I 1-17a 8.0 2/3 2.9 2 0 11 — — — BL2 63 M BJP III A P II 1-7 7.7 1 2.8 0 0 10 — — — BL3 50 M BJP III A P I 1-13 9.5 2/3 2.9 2 6 13 — — — AL1 83 M IgA I A N VII 3-2 3.1 2/3 2.8 0 0 9 III 6 1.7 AL2 74 M IgA II A N III-3 2-13 4.8 1 3 3 0 9 V 3 0 AL3 72 M IgA II A N III-2 2-14 11.0 1 5.9 2 0 11 V 4 4.8 DL1 57 M IgD III B P III-1 2-7 34.1 2/3 16.7 0 0 8 III 6 1.8 DL3 62 F IgD III B P II 1-2 8.3 2/3 3.3 6 27 14 V 6 2.6 DL4 56 M IgD III A P II 1-4 10.6 2/3 10 4 2 10 ND ND ND GLA 65 F IgG III A P III-2 2-14 12.9 1 13.9 0 0 11 V 4 2.4 GL2 66 F IgG I A N II 1-4 4.6 2/3 11.1 0 0 10 III 6 3.9 GL3 69 M IgG I A N I 1-19 4.6 2/3 2.9 1 2 11 I 4 2.2 GL4 55 M IgG III A P I 1-19 4.6 2/3 14.7 2 0 11 ND ND ND GL5 71 F IgG III A P III-1 2-7 16.3 2/3 9.3 0 1 12 IV 6 1.9 GL6 63 M IgG I A N III-2 2-1 7.2 2/3 11.1 0 0 9 III 6 14.5 GL7 47 F IgG III A P II 1-2 3.1 1 5.6 0 0 10 ND ND ND mean ± s.d. 7.8 ± 3.8 7.5 ± 4.9 1.7 ± 1.8 10.6 ± 1.5 4.0 ± 4.0
Clinical stage was classified according to the criteria of Durie and Salmon.14 Three patients (BK2, BL4 and DL3) presented with plasma cell leukemia at diagnosis. Five patients (BK3, DK1, GK5, GK6 and DL1) underwent hemodialysis. aNomenclature of the germline V and V genes was according to Kabat et al17 and Kawasaki et al,39 respectively. Since the variable region sequences in AK6 and DL1 had less than 80% homology to the known germline counterparts, they were excluded from the analysis on mutation. P, positive; N, negative; ND, not done;
amide gels in TBE buffer (90 mM Tris-borate, 2 mM EDTA: J,V, and J gene segments were assigned according to the pH 8.0), and were removed from the gels by the crush and published germline sequence data.17–39 The germline counter- soak procedure. The isolated fragments were digested with parts of the V/ genes were determined using sequence XbaI and HindIII restriction endonucleases (Boehringer analysis software (DNASIS; Hitachi Software Engineering, Mannheim Yamanouchi, Tokyo, Japan), and ligated to the Yokohama, Japan) with the GenBank database. M13mp18/19 phage vector (Boehringer Mannheim Yamanouchi). The ligated materials were transfected into Escherichia coli strain JM 109. More than 10 recombinants per Results clone were picked up and cultured in 2 × TY medium (16 g/l Bacto-tryptone; Difco Laboratories, Detroit, MI, USA, 10 g/l We analyzed the IgL variable region gene sequences from 17 Bacto-yeast extract; Difco Laboratories, 5 g/l NaCl). Single- -chain type MM, isotypes of which were IgG in seven, IgA stranded phage DNA was prepared from these cultures, and in six, IgD in one and BJP in three, and 16 -chain type MM, was sequenced using the fluorescein-conjugated universal isotypes of which were IgG in seven, IgA in three, IgD in three primer on a DNA sequencer (373-A; Applied Biosystems, Fos- and BJP in three. Both primer pairs used in this study rep- ter City, CA, USA) according to the manufacturer’s instruc- resented a single amplified band on each sample. Mono- tions. The sequence data of each clone were confirmed if all clonality of each amplified product was confirmed by sequences from each recombinant plaque were identical. V, sequence analysis. All analyzed sequences had no intraclonal IgL chain sequence of multiple myeloma H Kiyoi et al 603 variations, and neither a stop codon nor frame-shift. Deduced Usage of the J and J gene segments amino acid sequences are shown in Figure 1. Usage of the J and J gene segments is presented in Table 1. Usage of the J gene segments was apparently random. The Usage of the V and V genes J-1 segment was necessarily rearranged to the proximal V genes, whereas half of the rearranged J-5 segments were All -chain variable region genes except in one patient (AK6) joined to the distal V genes. However, this had no statistical were homologous to one of the known V germline sequences significance probably because of the limited number of ana- with identity ranging from 97.3 to 84.9% (Table 1). The - lyzed samples. On the other hand, usage of the J segments chain variable region gene from patient AK6 had 75.5% hom- was highly biased to J-1, J-2 and J-3 segments. Since J-2 ology to the V germline sequence B2, suggesting a novel and J-3 segments have the same sequence, we designated gene of the V-V family, or extensive somatic mutations. Eight, these two together as J-2/3 gene segments. three, four and one rearranged V gene belonged to V-I, V- II, V-III and V-IV gene families, respectively (Table 1). Ten of the 17 assigned V genes were physically located in the Somatic mutation of the IgL variable region genes J-proximal portion, when plotted on a map of the germline V locus on chromosome 2 (data not shown). One V gene Somatic mutations of the V/ genes were determined by was located at the B cluster. Seven and three V genes were comparing with the most homologous germline sequences, located at the proximal L and A clusters, respectively. although germline V/ gene repertoires and their polymor- Fifteen of the 16 V genes exhibited homology to one of phism have yet to be completely characterized. In the the known germline V genes ranging from 96.9 to 83.7% rearranged V genes, 7.0 ± 3.9% of mutations were observed (Table 1). However, one V gene from DL1 had only 65.9% (Table 1). In FWR and CDR, the average mutation rates were homology even to the most homologous germline V gene, 4.1% and 8.7%, respectively. The replacement/silent mutation 2–7.39 This lower homology suggests this V gene is novel, (R/S) ratios of the V gene were 1.43, and 1.03 and 1.84, while it is also possible that this gene has extensive somatic in whole, FWR and CDR, respectively. In V gene segments, mutations and/or polymorphism. Four, five, five and one 7.8 ± 3.8% of mutations were found, and average mutation rearranged V genes belonged to the V-I, V-II, V-III, and rates in the FWR and CDR were 5.6% and 12.3%, respect- V-VII gene family, respectively (Table 1). ively. The R/S ratios of the V gene in whole, FWR and CDR
Figure 1 Amino acid sequences of the IgL variable regions expressed in 33 MM samples. Borders of FWRs and CDRs were defined according to Kabat et al.17 IgL chain sequence of multiple myeloma H Kiyoi et al 604 were 1.73, 1.52 and 1.96, respectively. The R/S ratio above gene (7.5 ± 4.9% vs 3.8 ± 3.2%, P = 0.003, modified Student’s 2.9 are thought to reflect antigen-selected substitutions, if the t-test). Figure 2 shows the relationship between the amino acid amino acids used by the germline is random.40 However, position of the - and -chains, and the incidence of mutation. since IgH CDRs contain codons which have a greater propen- Notably, the mutations of the J segments were distributed sity for causing a change in the replacement mutation, it has only in their FWR-4 portions, whereas those of the J seg- been suggested that the expected R-mutation in each germline ments were in both CDR-3 and FWR-4 portions (Figure 2). On should be calculated to examine the antigen selection.41 In close inspection, the incidence of the mutations in FWR-4 was the present study, we also calculated the expected R-mutation found to differ from codon to codon. There was no nucleotide within FWRs and CDRs on each rearranged V and V gene mutation at codons 98, 99 and 102 and the -chain, nor at (Table 2). In 14 of the 16 V genes and all of the V genes, codons 98, 105 and 106 of the -chain. Furthermore, replace-
VL genes had fewer than expected R-mutations in FWRs. The ment mutations were remarkably clustered at codon 106 of overall numbers of FWR R-mutations were 71 in V and 88 the -chain, and codon 103 of the -chain (Figure 2). Kabat in V genes, which are significantly less than the 121 and 127 et al17 have reported that codon 106 in the -chain and 103 expected mutations, respectively (P = 0.005 in V and in the -chain are the most variable in FWR-4, suggesting that P = 0.04 in V gene). In CDRs, seven of the 16 V genes and this finding is not specific for MM. eight of the 15 V genes had greater than expected R- mutations, while these are not significant. The overall num- bers of CDR R-mutations were 105 in V and 102 in V genes, Junctional sequence of V-J and V-J which are the same as expected. rearrangements Because the germline J and J gene segments have been completely sequenced, sequence differences reflect real All V-J and V-J junctional sequences were in-frame. mutations of the rearranged J and J segments. The mutation Nucleotide deletions at the 5Ј coding end of the J gene seg- rate of the J segments was significantly higher than that in J ments were observed in 10 of the 17 J and eight of the 16
Table 2 Observed and expected replacement mutations in V/ genes
UPN FWR CDR
Total Replace Expected Total Replace Expected mutation mutation R-mutation mutation mutation R-mutation
-chain BK1 5 3 4.3 2 2 1.4 BK2 5 2 4.3 4 3 2.8 BK3 4 2 3.5 4 3 2.7 AK1 15 7 13.0 23 13 16.1 AK2 10 3 8.6 22 16 15.4 AK3 11 5 9.5 12 6 8.5 AK4 16 9 13.8 11 7 7.8 AK5 7 6 6.1 6 4 4.2 DK1 6 3 5.0 3 2 2.1 GK1 12 5 10.5 11 9 7.7 GK2 9 3 7.8 17 9 12.0 GK3 10 4 8.7 9 6 6.3 GK4 16 7 14.0 19 12 13.3 GK5 4 4 3.6 7 3 4.9 GK6 6 4 5.2 4 3 2.7 GK7 4 4 3.8 8 7 5.6 total 140 71 121.4 162 105 113.5
-chain BL1 6 4 5.0 6 3 4.2 BL2 3 2 2.4 9 7 6.4 BL3 5 3 4.3 7 5 3.4 AL1 5 3 4.3 3 2 3.5 AL2 9 6 7.7 9 5 2.4 AL3 21 13 18.1 21 16 14.7 DL3 19 7 16.5 14 10 9.8 DL4 12 7 10.3 16 8 11.2 GL1 11 7 9.5 10 8 7.1 GL2 12 9 10.5 13 7 9.1 GL3 10 7 8.8 10 7 7.0 GL4 10 6 8.7 11 7 7.8 GL5 16 10 13.7 12 8 8.5 GL6 3 1 2.6 9 7 6.5 GL7 4 3 3.6 4 2 2.8 total 146 88 126.6 154 102 107 IgL chain sequence of multiple myeloma H Kiyoi et al 605 odialysis later (Table 1). There was no relationship between the clinical features and the organization and mutation of the IgL variable region gene. In the light chain deposition disease which reportedly causes renal dysfunction, it has been sug- gested that an unusual sequence and/or glycosylation of the -chain is related to the pathogenesis.43,44 However, in the patients with renal dysfunction, no common structure of IgL was observed. A potential N-glycosylation site, Asn-Phe-Thr at codon 70,44 was found in only one patient AK4 (Figure 1), whose renal function was normal. Furthermore, the V germ- line A27 gene was used in two patients (AK5 and GK2), but only the former had renal dysfunction.
Discussion
Usage of the V/ genes and J/ segments
The human Ig -chain locus is located on the short arm of chromosome 2, and consists of at least 76 mapped V genes, five J segments and a constant region gene.21,45 The V germ- line genes were divided into seven families (I–VII) by Figure 2 Somatic mutations in the rearranged J and J gene seg- sequence homology.17 As to the V family gene usage in ments. The cumulative number of V/ genes with a mutation at each human normal B cells, V -I and V -II families are most fre- codon encoded by the J and J gene segments is presented. Silent quently used, followed by V-III and V-IV, and the remaining and replacement mutations are indicated by open and closed 46,47 boxes, respectively. families are rarely used. In human lymphoproliferative dis- orders, including six acute lymphoblastic leukemia, 11 chronic lymphocytic leukemia, 23 malignant lymphoma and J segments (Table 1). The average number of deleted nucleo- three MM, V-I family was predominantly used (59%),and V- tides in the J and J segments was 0.94 ± 0.97, ranging from II (15%), V-III (13%), and V-IV (16%) families were used 0 to 3, and 1.5 ± 2.1, ranging from 0 to 6 nucleotides, respect- equally.46 The usage of the V gene families in normal and ively. Deletion of the 3Ј end of the V genes seemed to occur malignant B cells essentially reflects the size of each V family more frequently in -chain than in -chain, although the exact member. In this study, genes of the V-I family were predomi- 3Ј border of V/ germline genes was not necessarily deter- nantly used (50%), followed by those of the V-II (18.8%) and mined (data not shown). Possible N nucleotide addition was V-III (25.0%) family, as in previous reports except for under- found at two V-J (BK2 and GK6) and five V-J (BL3, DL3, utilization of the V-IV family genes. A previous study of DL4, GL3 and GL5) junctions (Figure 3). lymphoproliferative disorders showed that V-IV family genes Based on the terminology of Kabat et al,17 the CDR-3 of the were frequently observed in the rearranged -chain gene of - and -chains was defined as spanning codons 89 to 97. the -chain producing cells.48 In this study, rearrangements of The length of the -chain CDR-3 (10.6 ± 1.5 amino acids) was the -chain gene were analyzed in -chain producing MM, longer (P Ͻ 0.001, modified Student’s t-test) and more vari- and this may be associated with the under-utilization of V- able than that of the -chain CDR-3 (9.1 ± 0.3 amino acids). IV family. In chronic lymphocytic leukemia cells, HUMKV325 (A27) V germline gene, which has rheumatoid factor speci- ficity, has been preferentially found.49,50 In MM, no biased V Correlation of the variable region genes between IgL segment usage was detected. and the paired IgH Human V locus has a characteristic physical feature. Within 50 kb upstream of the J gene cluster, three germline We have previously sequenced the IgH variable region genes V genes (B1, B2 and B3), members of the V-IV, -V and -VII expressed in MM samples used in this study.42 Accordingly family respectively, are located, and this locus is called the B we analyzed the pairing pattern of IgH and IgL variable region region. In the locus further upstream, there are two duplicated genes from the view point of gene organization and mutation. megabase blocks (J-proximal and distal), each of which con- However, there was no preference in the pairing of IgH and tains about 35 V genes and is divided into three regions: L, IgL (Table 1). Mutation rates of the V / and VH genes did not A and O. The V genes in the distal block lie in inverted orien- correlate (data not shown), nor did those of the J / and JH tation with respect to the proximal block and J segments. segments (Table 1). Most of the rearranged V segments have been assigned to the J-proximal germline V segments,24 and the present study supports this assignment. Two likely reasons for the biased Relationship between clinical features and the IgL usage are: (1) most V segments in the proximal region sequences rearrange by a deletion mechanism, whereas those in the dis- tal region require an inversion mechanism, which might lead Three patients (BK2, BL4 and DL3) presented with plasma cell to an unsuccessful rearrangement. (2) The relative distances leukemia at diagnosis. Seven patients (BK3, AK5, DK1, GK5, of the proximal and distal regions from the J genes may GK6, DL1 and DL3) had renal dysfunction in which the serum influence the biased V segment usage. This mechanism, the creatinine level was over 2.0 mg/dl at diagnosis, and five so called ‘one-dimensional tracking model’, has been pro- (BK3, DK1, GK5, GK6, and DL1) of the seven underwent hem- posed in the rearrangement of the IgH variable region genes, IgL chain sequence of multiple myeloma H Kiyoi et al 606
Figure 3 V-J junctional sequences with N nucleotides. Germline counterparts of the V gene and J segments are shown in parenthesis. Identical and deleted nucleotides are indicated by dots and dashes, respectively. Since the V and J genes from patient DL3 had no homology at the 3Ј end and 5Ј end, respectively, 27 nucleotides at the V/J junction were tentatively assigned as an N region.
especially at the earlier stage of B cell ontogeny.51,52 The data C locus. The remaining five V genes from MM cells were that both J-distal V genes rearranged to the J-5 segment assigned to region III, and all of them rearranged with the J- supports the possibility of secondary rearrangements in the 2/3 gene. The relatively preferential usage of the V genes in locus, although the number of sequences analyzed was lim- region I might be also influenced by the physical distance ited. In the rearrangement of the IgH gene, the usage of JH from the J -C locus as in the IgH and Ig gene loci. Of note segments revealed a developmental stage-specific trend.7–13 is that all V genes in region III rearranged with the J-2/3 At the earliest stage of B cell development, the initial D-JH gene, and that the V -J junctions were well modified. This rearrangements predominantly occur between DHQ52 and JH- result suggested secondary V -J rearrangement. On the other 1 segments which are located close to each other.53 During hand, the bias of J segments to J-1 and J-2/3 might be B cell development, a second or a third D-JH joining might induced by the unique structure of germline J genes. Three Ј Ј occur, then the more 5 D segments and more 3 JH segments of the seven J segments (J -4, 5, and 6) are pseudogenes, and would join at a later stage.54 However, V genes locating in seven tandem J-C regions span approximately 30 kb.37 the distal region risk unsuccessful rearrangement because of the inverted orientation of the V genes. Thus, only a limited number of productive V-J rearrangements might result from Somatic mutations in the IgL gene the secondary rearrangement. The human -chain locus is located on the long arm of Somatic mutations contribute to the diversification of the chromosome 22. In contrast to the -chain locus, each J seg- rearranged Ig genes and are important for affinity maturation ment is associated with one respective constant gene.37 The during B cell development. The degree of mutation may V germline genes have been recently reclassified into 10 reflect the developmental stage of B cells, and analysis of families (I–X) on the basis of conserved key amino acids and somatic mutations in malignant cells could indicate the stage nucleotide sequences.36,55 More recently, the entire human at which malignant transformation took place. The V genes Ig gene locus has been sequenced, and 36 potentially active from chronic lymphocoytic leukemia, prolymphocytic leuke- V gene segments have been determined.38,39 However, in mia and hairy cell leukemia are rarely mutated or conserve this study, one V gene from DL1 had only 65.9% homology germline sequences.56 More highly mutated V genes have ¨ even to the most homologous germline gene, suggesting a been observed in Waldenstrom’s macroglobulinemia and novel V gene or extensive somatic mutations and/or poly- MM.56 In our studies, the V/ genes mutated at 7.0 ± 3.9% morphism. Usage of the V-I, -II, and -III families might also and 7.8 ± 3.8%, respectively, supporting the previous data. In reflect the genomic complexity of the V families. On the VH genes of MM clones, R-mutations in CDRs were greater physical map of the Ig locus, 10 V genes from MM cells than expected, while those in FWRs were significantly were assigned to region I, the region most proximal to the J- fewer.41 This suggested the structural necessity of FWRs for a IgL chain sequence of multiple myeloma H Kiyoi et al 607 functional antibody molecule. Fewer R-mutations than Comparison of the clinical features and IgL sequences expected in FWRs of VL genes were in agreement with VH genes, although R-mutations in CDRs were not greater than Renal dysfunction, one of the most critical complications in predicted by chance. This discrepancy might be due to the MM, is induced mainly by exhaustion of the tubular absorp- 42 lower R-mutations in the paired VH genes. It is possible that tion capacity of IgL molecules followed by deposition of IgL. this is due to the underestimation of somatic mutations, since Secondary amyloidosis is another cause of nephropathy, we did not determine the exact germline VL genes in each which is more common in -producing myeloma. In light patient. chain deposition disease, an MM-associated amino acid Somatic mutations were more frequently observed in the sequence, observed in a deposited -chain of abnormal size rearranged J than rearranged J segments, and distribution with glycosylation, has been speculated to play a role in the patterns differed between J and J segments. While mutations pathogenesis.43 It has been reported that somatic mutation of of the J segments occurred in the FWR-4 portions, those of the V-IV family gene, which was deposited in the kidney, the J segments occurred in the CDR-3 and FWR-4 portions. resulted in a potential N-glycosylation site.44 However, one Furthermore, R-mutations were clustered at codon 106 in the patient (UPN-AK4), whose V gene had such a potentially N- J segments, and at codon 103 in the J segments. Our study glycosylated site, did not have renal dysfunction. Thus, we supported the previous finding that -chains conserved Phe could not find a clear correlation between the primary and Thr at codons 98 and 102, respectively,57 suggesting that sequence of IgL and nephropathy. However, the nephropathy these conserved codons reflect the structural requirement for has been suggested to depend on the physicochemical -chain and/or the association with IgH during B cell matu- character of the IgL molecule itself.15 A number of factors, ration. In the JH gene segments of MM samples, frequent R- including the amount of BJP synthesized, secondary modifi- mutations have been observed at codon 102:58 in six of 28 cation of the deposited protein and serum calcium level, MM samples in our study.42 The significance of amino acid might be associated with the renal dysfunction in MM. Further replacement in FWR-4 has not been determined. The characterization of IgL sequences is necessary to clarify this extremely high R/S ratio on these codons suggests that amino issue. acid replacements even in FWR modulate the antigen-binding affinity and/or specificity. Acknowledgements
Junctional sequence of the V-J and V-J We thank Ms Chisato Kamiya for technical assistance and Ms rearrangements Chiaki Nagai for preparing the manuscript. This work was sup- ported, in part, by a Grant-in-Aid from the Japanese Ministry of Education, Science and Culture. Nucleotide deletions of the 5Ј coding end of the J and J gene segments were observed in about half the MM samples. So far it has been reported that no in-frame V-J junctions delete more than four nucleotides,59 and that nucleotides are References never deleted in in-frame V-J junctions.57 However, the present study showed that deletions of up to six nucleotides 1 Tonegawa S. Somatic generation of antibody diversity. Nature 1983; 302: 575–581. could occur in in-frame V -J junctions. These findings are 2 Honjo T. Immunoglobulin genes. Annu Rev Immunol 1983; 1: explained by a structural necessity for conserving codon 97, 499–528. Thr, in -chain and codon 98, Phe, in -chain, that would 3 Alt FW, Blackwell TK, DePinho RA, Reth MG, Yancopoulos GD. result in negative selection against further nucleotide deletion Regulation of genome rearrangement events during lymphocyte from the J and J segments. On the other hand, codon 96 differentiation. Immunol Rev 1986; 89: 5–30. which is most variable in -chain was thought to be generated 4 Alt FW, Yancopoulos GD, Blackwell TK, Wood C, Thomas E, Boss M, Coffman RR, Rosenberg N, Tonegawa S, Baltimore D. Ordered mainly by the junctional deletion and flexibility of the J arrangement of immunoglobulin heavy chain variable region seg- segments. ments. EMBO J 1984; 3: 1209–1219. N nucleotide addition considerably increases the diversity 5 Melchers F, Karasuyama H, Haasner D, Bauer S, Kudo A, Sakagu- of the IgH CDR-3. However, most of the V-J and V-J chi N, Jameson B, Rolink A. The surrogate light chain in B-cell junctions reportedly lack N nucleotides.48,56,57,59,60 We found development. Immunol Today 1993; 14: 60–68. possible N nucleotide additions in two V-J and five V-J 6 Takeda S, Zou Y, Bluethmann H, Kitamura D, Muller U, Rajewsky K. Deletion of the immunoglobulin chain intron enhancer abol- junctions, although the incidence of N addition was far less ishes chain gene rearrangement in cis but not chain gene than in the IgH genes. The lack of N nucleotides is explained rearrangement in trans. EMBO J 1993; 12: 2329–2336. by the fact that terminal deoxynucleotidyl transferase (TdT) is 7 Schroeder HW Jr, Wang JY. Preferential utilization of conserved no longer expressed at the stage of IgL gene rearrangement.61 immunoglobulin heavy chain variable gene segments during Recently, it was shown that in a small subpopulation, -chain human fetal life. Proc Natl Acad Sci USA 1990; 87: 6146–6150. loci have already rearranged at the early stage of B cell devel- 8 Yamada M, Wasserman R, Reichard BA, Shane S, Caton AJ, Rov- 61,62 era G. Preferential utilization of specific immunoglobulin heavy opment at which IgH gene rearrangement occurs. In chain diversity and joining segments in adult human peripheral addition, the V-J junctions in this population did not con- blood B lymphocytes. J Exp Med 1991; 173: 395–407. tain N nucleotides, indicating a limited number of N nucleo- 9 Sanz I. Multiple mechanisms participate in the generation of diver- tides in -chain even when TdT is present.61,62 The length of sity of human H chain CDR3 regions. J Immunol 1991; 147: -chain CDR-3 was almost fixed at nine amino acids, whereas 1720–1729. that of -chain was slightly longer and varied. Therefore, 10 Stewart AK, Huang C, Long AA, Stollar BD, Schwartz RS. VH-gene representation in autoantibodies reflects the normal human B-cell defects of N nucleotides in the IgL genes seem to be associa- repertoire. Immunol Rev 1992; 128: 101–122. ted not with B cell stage but the structural necessity of the 11 Kiyoi H, Naoe T, Horibe K, Ohno R. Characterization of the IgL molecule. immunoglobulin heavy chain complementarity determining IgL chain sequence of multiple myeloma H Kiyoi et al 608 region (CDR)-III sequences from human B cell precursor acute 32 Daley MD, Peng H, Misener V, Liu XY, Chen PP, Siminovitch KA. lymphoblastic leukemia cells. J Clin Invest 1992; 89: 739–746. Molecular analysis of human immunoglobulin V germline genes: 12 Wasserman R, Ito Y, Galili N, Yamada M, Reichard BA, Shane S, subgroups VIII and VIV. Mol Immunol 1992; 29: 1515–1518. Lange B, Rovera G. The pattern of joining (JH) gene usage in the 33 Combriato G, Klobeck HG. V and J-C gene segments of the human IgH chain is established predominantly at the B precursor human immunoglobulin light chain locus are separated by 14 kb cell stage. J Immunol 1992; 149: 511–516. and rearrange by a deletion mechanism. Eur J Immunol 1991; 21: 13 Steenbergen EJ, Verhagen OJHM, van Leeuwen EF, Behrendt H, 1513–1522. Merle PA, Wester MR, von dem Borne AEGKr, van der Schoot CE. 34 Winkler TH, Fehr H, Kalden JR. Analysis of immunoglobulin vari- B precursor acute lymphoblastic leukemia third complementarity- able region genes from human IgG anti-DNA hybridomas. Eur J determining regions predominantly represent an unbiased recom- Immunol 1992; 22: 1719–1728. bination repertoire: leukemic transformation frequently occurs in 35 Williams SC, Winter G. Cloning and sequencing of human fetal life. Eur J Immunol 1994; 24: 900–908. immunoglobulin V gene segments. Eur J Immunol 1993; 23: 14 Durie BGM, Salmon SE. A clinical staging system of multiple 1456–1461. myeloma. Cancer 1975; 36: 842–854. 36 Stiernholm NBJ, Kuzniar B, Berinstein NL. Identification of a new 15 Solomon A, Weiss DT, Kattine AA. Nephrotoxic potential of human V gene family-VX. J Immunol 1994; 152: 4969–4975. Bence–Jones proteins. New Engl J Med 1991; 324: 1845–1851. 37 Vasicek TJ, Leder P. Structure and expression of the human 16 Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Labora- immunoglobulin genes. J Exp Med 1990; 172: 609–620. tory Manual. 2nd edn. Cold Spring Harbor Laboratory Press: Cold 38 Williams SC, Frippiat J-P, Tomlinson IM, Ignatovich O, Lefranc Spring Harbor, NY, 1989. M-P, Winter G. Sequence and evolution of the human germline 17 Kabat EA, Wu TT, Reid-Muller M, Perry HM, Gottesman KS, V repertoire. J Mol Biol 1996; 264: 220–232. Foeller C. Sequences of Proteins of Immunological Interest, 5th 39 Kawasaki K, Minoshima S, Nakato E, Shibuya K, Shintani A, edn, National Institutes of Health Publication: Bethesda, MD, Schmeits JL, Wang J, Shimizu N. One-megabase sequence analy- 1991. sis of the human immunoglobulin gene locus. Genome Res 18 Pech M, Smola H, Pohlenz HD, Straubinger B, Gerl R, Zachau 1997; 7: 250–261. HG. A large selection of the gene locus encoding human immuno- 40 van Es JH, Gmelig Meyling FHJ, Logtenberg T. High frequency of globulin variable regions of the kappa type is duplicated. J Mol somatically mutated IgM molecules in the human adult blood B Biol 1985; 183: 291–299. cell repertoire. Eur J Immunol 1992; 22: 2761–2764. 19 Klobeck HG, Bornkamm GW, Combriato G, Mocikat R, Pohlenz 41 Vescio RA, Cao J, Hong CH, Lee JC, Wu CH, Der Danielian M, HD, Zachau HG. Subgroup IV of human light chains is encoded Wu V, Newman R, Lichtenstein AK, Berenson JR. Myeloma Ig by a single germline gene. Nucleic Acids Res 1985; 13: 6515– heavy chain V region sequences reveal prior antigenic selection 6529. and marked somatic mutation but no intraclonal diversity. J Immu- 20 Pohlenz HD, Straubinger B, Thiebe R, Peck M, Zimmerman FJ, nol 1995; 155: 2487–2497. Zachau HG. The human V locus. Characterization of extended 42 Kiyoi H, Naito K, Ohno R, Naoe T. Comparable gene structure of immunoglobulin gene regions by cosmid cloning. J Mol Biol 1987; the immunoglobulin heavy chain variable region between mul- 193: 241–253. tiple myeloma and normal bone marrow lymphocytes. Leukemia 21 Meindl A, Klobeck HG, Ohnheiser R, Zachau HG. The V reper- 1996; 10: 1804–1812. toire in the human germline. Eur J Immunol 1990; 20: 1855–1863. 43 Picken MM, Frangione B, Barlogie B, Luna M, Gallo G. Light 22 Lautner-Rieske A, Huber C, Meindl A, Pargent W, Schable KF, chain deposition disease derived the I light chain subgroup. Bio- Thiebe R, Zocher I, Zachau HG. The human immunoglobulin chemical characterization. Am J Pathol 1989; 134: 749–754. ´ locus. Characterization of the duplicated A regions. Eur J Immunol 44 Cogne M, Preud’homme JL, Bauwens M, Touchard G, Aucouturier 1992; 22: 1023–1029. P. Structure of a monoclonal kappa chain of the VIV subgroup 23 Schable KF, Klobeck HG, Ohnheiser R, Zachau HG. The variable in the kidney and plasma cells in light chain deposition disease. genes in the human immunoglobulin locus. Biol Chem Hoppe- J Clin Invest 1991; 87: 2186–2190. Seyler 1993; 374: 1001–1022. ¨ 45 Huber C, Huber E, Lautner-Rieske A, Schabl KF, Zachan HG. The 24 Cox JPL, Tomlinson IM, Winter G. A directory of human germ- human immunoglobulin locus. Characterization of the partially line V segments reveals a strong bias in their usage. Eur J Immunol duplicated L regions. Eur J Immunol 1993; 23: 2860–2967. 1994; 24: 827–836. 46 Cuisinier AM, Fumoux F, Moinier D, Boubli L, Guigou V, Milili 25 Anderson MLM, Szajnert MF, Kaplan JC, McColl L, Young BD. M, Schiff C, Fourgereau M, Tonnelle C. Rapid expansion of human The isolation of a human Ig V gene from a recombinant library immunoglobulin repertoire (VH, V,V) expressed in early fetal of chromosome 22 and estimation of its copy number. Nucleic bone marrow. New Biol 1990; 2: 689–699. Acids Res 1984; 12: 6647–6661. 47 Solomon A. Light chains of immunoglobulins: structural–genetic 26 Alexander D, Chuchana P, Brockly F, Blancher A, Lefranc G, correlates. Blood 1986; 68: 603–610. Lefranc M-P. First genomic sequence of a human Ig variable gene belonging to subgroup I. Functional genes, pseudogenes and 48 Cannell PK, Amlot P, Attard M, Hoffbrand AV, Foroni L. Variable vestigial sequences are interspersed in the IGLV locus. Nucleic kappa gene rearrangement in lymphoproliferative disorders: an Acids Res 1989; 17: 3975. analysis of V gene usage, VJ joining and somatic mutation. Leuke- 27 Brockly F, Alexander D, Chuchana P, Huck S, Lefranc G, Lefranc mia 1994; 8: 1139–1145. 49 Kipps TJ, Fong S, Tomhave E, Chen PP, Goldfein R, Carson F. M-P. First nucleotide sequence of a human immunoglobulin vari- able gene belonging to subgroup II. Nucleic Acids Res 1989; High frequency expression of a conserved light-chain D variable 17: 3976. region gene in chronic lymphocytic leukemia. Proc Natl Acad Sci 28 Siminovitch KA, Misener V, Kwong PC, Song Q-L, Chen PP. A USA 1987; 84: 2916–2920. natural autoantibody is encoded by germline heavy and lambda 50 Rassenti LZ, Pratt LF, Chen PP, Carson DA, Kipps TJ. Auto-anti- light chain variable region genes without somatic mutation. J Clin body encoding kappa L chain genes frequently rearranged in Invest 1989; 84: 1675–1678. lambda L chain expressing chronic lymphocytic leukemia. J 29 Frippiat J-P, Chuchana P, Bernard F, Buluwela L, Lefranc G, Lefr- Immunol 1991; 147: 1060–1066. anc M-P. First genomic sequence of a human Ig variable lambda 51 Yancopoulos GD, Desiderio SV, Paskind M, Kearney JF, Baltimore gene belonging to subgroup III. Nucleic Acids Res 1990; 18: D, Alt FW. Preferential utilization of the most JH-proximal VH gene 7134–7134. segments in pre-B-cell lines. Nature 1984; 311: 727–733. 30 Bernard F, Chuchana P, Frippiat J-P, Buluwela L, Lefranc M-P. 52 Wood C, Tonegawa S. Diversity and joining segments of mouse Genomic sequence of IGLV1S2, a human immunoglobulin vari- immunoglobulin heavy chain genes are closely linked and in the able lambda gene belonging to subgroup I. Nucleic Acids Res same orientation: implications for the joining mechanism. Proc 1990; 18: 7139. Natl Acad Sci USA 1983; 80: 3030–3034. 31 Daley MD, Olee T, Peng H, Soto-Gil RW, Chen PP, Siminovitch 53 Nickerson KG, Berman J, Glickman E, Chess L, Alt FW. Early KA. Molecular characterization of the human immunogloublin VI human IgH gene assembly in Epstein–Barr virus-transformed fetal germline gene repertoire. Mol Immunol 1992; 29: 1031–1042. B cell lines. Preferential utilization of the most JH-proximal D seg- IgL chain sequence of multiple myeloma H Kiyoi et al 609 ment (DQ52) and two unusual VH-related rearrangements. J Exp mutations in D and/or JH segments of Ig gene in Waldenstroms Med 1989; 169: 1391–1403. macroglobulinemia and chronic lymphocytic leukemia (CLL) with
54 Reth MG, Jackson S, Alt FW. VHDJH formation and DJH replace- Richter’s syndrome but no common ALL. Blood 1995; 85: ment during pre-B differentiation: non-random usage of gene seg- 1913–1919. ments. EMBO J 1986; 5: 2131–2138. 59 Ramsden DA, Paige CJ, Wu GE. K light chain rearrangement in 55 Chuchana P, Blancher A, Brockly F, Alexandre D, Lefranc G, mouse fetal liver. J Immunol 1994; 153: 1150–1160. Lefran M-P. Definition of the human immunoglobulin variable 60 Milstein C, Even J, Jaruis JM, Gonzalez-Fernandez A, Gherardi E. lambda (IGLV) gene subgroups. Eur J Immunol 1990; 20: 1317– Non-random features of the repertoire expressed by the members 1325. of one V gene family and of the V-J recombination. Eur J Immunol 56 Wagner SD, Martinelli V, Luzzatt L. Similar patterns of V gene 1992; 22: 1627–1634. usage but different degrees of somatic mutation in hairy cell leuke- 61 Ehlich A, Schaaf S, Gu H, Kitamura D, Muller W, Rajewsky K. ¨ mia, prolymphocytic leukemia, Waldenstrom’s macroglobuline- Immunoglobulin heavy and light chain genes rearrange indepen- mia and myeloma. Blood 1994; 83: 3647–3657. dently at early stages of B-cell development. Cell 1993; 72: 57 Stiernholm NBJ, Berinstein NL. A mutated promoter of a human 695–704. Ig V gene segment is associated with reduced germline transcrip- 62 Victor K, Vu DK, Feeney AJ. Limited junctional diversity in light tion and a low frequency of rearrangement. J Immunol 1995; 154: chains. Junctional sequences from CD43+B220+ early B cell pro- 1748–1761. genitors resemble those from peripheral B cells. J Immunol 1994; 58 Aoki H, Takishita M, Kosaka M, Saito S. Frequent somatic 152: 3467–3475.