In this study, we analyzed the targeting of the somatic hypermutation (SHM) mechanism at specific hotspot sequence motifs in the VH and Vκ genes of 10 follicular lymphoma (FL) cases and the Vκ and Vλ genes of 11 κ- and six λ-light chain expressing multiple myeloma (MM) cases. These sequences were analyzed for targeting of specific motifs, ie certain highly mutable trinucleotides (3-NTPs), the tetranucleotide (4-NTP) RGYW and its complementary, WRCY (where R = purine, Y = pyrimidine and W = A or T). Comparisons were carried out between mutation frequencies in RGYW vs WRCY and the incidence of mutations in complementarity determining region (CDR)-1 vs CDR2 vs CDR3. Statistically significant differences were obtained when comparing: (1) the ratio of mutations in 4-NTPs (RGYW, WRCY, RGYW+WRCY)/mutations in the whole V sequence in MM-Vκ vs MM-Vλ; (2) the total number of mutated 4-NTPs in MM-Vκ vsFL-Vκ; (3) the number of mutated RGYW 4-NTPs in MM-Vκ vsFL-Vκ and FL-VH vs FL-Vκ; (4) the number of mutated WRCY 4-NTPs in MM-Vκ vs FL-Vκ (P = 0.006) and FL-VH vs FL-Vκ; (5) the targeting of RGYW vs WRCY in the CDRs of FL-VH genes. Similar results (regarding statistical significance) were obtained when undertaking intergroup comparisons for 3-NTPs. These findings conform well with relevant data derived from normal peripheral B cells. The differences observed in favor of 4-NTP (RGYW and WRCY) targeting in FL-VH vsFL-Vκ and MM-Vκ vs FL-Vκ may implicate differences in the evolution of SHM coupled with selection in different stages of B cell ontogeny. Several explanations can be offered for the fact that hotspot sequences were not always targeted by SHM in FL and MM: (1) other unrecognized motifs may be targets of SHM; (2) ‘inappropriately’ introduced mutations were fixed and propagated by the neoplastic process; (3) certain FL and MM cases might have lost their ability to correct mutations introduced in classic hotspots due to deficient mismatch-repair (MMR) mechanisms; conversely, in other cases with intact MMR function, the hotspot to non-hotspot targeting of somatic hypermutation is balanced.
The antigen-independent phase of B cell ontogeny takes place in the bone marrow and is characterized by ordered immunoglobulin (lg) gene rearrangements leading to the assembly of distinct variable (V), diversity (D) (for heavy chains only) and joining (J) gene segments into a V(D)J gene complex, a process known as V(D)J recombination.1 Successful rearrangement of heavy chains (HC) lg genes and subsequently that of light chain (LC) lg genes (κ or λ) will enable the developing B cell to later express on its surface a fully functional lg receptor with unmutated V region sequences.2
The second, antigen-dependent, phase will start when the ‘naïve’ B cell exiting the bone marrow enters into the follicles of the secondary lymphoid organs where contact with and selection by antigen takes place.3 The molecular hallmark of this phase is the introduction of mutations within rearranged IgV genes, at a rate much higher than usual, a phenomenon described as somatic hypermutation.4
Antigen selection in B cell ontogeny is evidenced by non-random distribution of somatic mutations in lg HC and LC V genes, a feature providing useful information concerning the ontogenetic assignment of B cell neoplastic disorders.5 In this context, an increased ratio of replacement (R) to silent (S) mutations in the complementarity determining regions (CDRs) of IgV genes, has been considered as the most reliable surrogate marker of selection by antigen for higher avidity.6 However, irrespective of subsequent selection, somatic hypermutation is primarily targeted at specific hotspots within V genes, ie tri- or tetra-nucleotide sequence motifs, such as the RGYW motif (R = purine, Y = pyrimidine, W = A or T) and its complementary, WRCY.7
While the exact mechanism of somatic hypermutation remains elusive, this phenomenon is characterized by certain unique features.8 The nature of mutations indicates a preference for transitions over transversions with purines being more frequently targeted than pyrimidines, suggesting strand bias; mutations are concentrated mainly in the CDRs and most often are single nucleotide substitutions rather than deletions or insertions; certain codons are targeted more often by the mutational process, while others are less likely to tolerate changes; finally, a striking bias exists for G and C over A and T nucleotide mutations.9
In the present study, we analyzed the distribution of somatic hypermutation and its targeting at specific mutational hotspots in the clonotypic VH and Vκ genes of follicular lymphoma (FL) as well as the Vκ and Vλ genes of multiple myeloma (MM), tumors corresponding to antigen-selected intra-germinal center (GC) and post-GC stages of B cell ontogeny.
Materials and methods
Included in the present study were the clonotypic VH and Vκ gene sequences of 10 FL cases10 as well as the clonotypic Vκ and Vλ gene sequences of 11κ- and six λ-light chain expressing MM cases analyzed previously by our group.11,12 The sequences have been submitted to the EMBL database (http://www.ebi.ac.uk/embl/index.html) with the following accession numbers: for FL-VH, AJ410896 to AJ410905; for FL-Vκ, AJ410886 to AJ410895; for MM-Vκ AJ410906 to AJ410916; and for MM-Vλ, AJ410917 to AJ410922.
The analysis aimed at determining whether specific nucleotide motifs, ie the tetranucleotide RGYW and its complementary WRCY were targeted by the somatic hypermutation machinery. A further objective was to identify whether the incidence of mutations was dependent on the actual sequence and location of the nucleotide motifs in the CDRs and framework regions (FWRs). To this purpose, we examined eight highly mutable trinucleotides (AGC, GCT, ATT, AAT, TAC, GTA, AGT and AGA),13 most of which are part of the RGYW/WRCY motif.
Statistical analyses of the distribution of mutations in tri- and tetra-nucleotides were carried-out by Student's t-test and paired χ2. We compared the incidence of mutations in VH vs Vκ sequences in FL, Vκ vs Vλ sequences in MM, and MM-Vκ vs FL-Vκ. Furthermore, we compared the incidence of mutated tetranucleotides and trinucleotides in the same subgroups of sequences. We then examined possible differences in the frequency with which different tetranucleotides (RGYW or WRCY) were targeted by the somatic hypermutation mechanism.
The absolute numbers of mutations in tetra- and tri-nucleotides and their distributions in each CDR, FWR, entire V sequence, RGYW/WRCY motifs, as well as the numbers of mutated tetra- and tri-nucleotides and their distributions in each CDR, FWR, entire V sequence, RGYW/WRCY motifs are demonstrated in Table 1. In general, no statistically significant differences were found when comparing the number of tetranucleotides that might serve as mutational hotspots in any kind of sequence examined (FL-VH, FL-Vκ, MM-Vκ, and MM-Vλ). However, in FL-VH genes the number of RGYW tetranucleotides was significantly higher than the number of WRCY tetranucleotides (P < 0.001). Regarding ‘mutation load’, MM-Vκ sequences were more heavily mutated than FL-Vκ sequences; a similar finding was noted between FL-VH and FL-Vκ sequences.
Statistically significant differences were obtained when comparing tetranucleotide and trinucleotide hotspots, as shown below.
(1) The ratio of mutations in tetranucleotides (RGYW, WRCY, RGYW+WRCY)/mutations in the whole V sequence in MM-Vκ vs MM-Vλ (P = 0.01) (Table 2); (2) the total number of mutated tetranucleotides in MM-Vκ vs FL-Vκ (P < 0.01) (Table 3); (3) the number of mutated RGYW tetranucleotides in MM-Vκ vs FL-Vκ (P < 0.001) and FL-VH vs FL-Vκ (P < 0.001) (Table 3 and Figure 1); (4) the number of mutated WRCY tetranucleotides in MM-Vκ vs FL-Vκ (P = 0.006) and FL-VH vs FL-Vκ (P < 0.001) (Table 3 and Figure 1); and (5) the targeting of RGYW was significantly higher when compared to WRCY tetranucleotides in the CDRs of FL-VH genes (P = 0.004) (not shown).
It has been demonstrated that in Vκ genes from normal peripheral B cells carrying non-productive Vκ-Jκ rearrangements, each RGYW tetranucleotide and its corresponding WRCY inverse repeat contained mutations at comparable frequencies, with mutations in G and C being significantly more prevalent. Moreover, mutations in codons contained within the RGYW/WRCY motifs were significantly more frequent in the CDRs than in the FWRs of productive vs non-productive rearrangements.13 These results were interpreted as strong indicators that the hypermutation mechanism targets the over-represented RGYW motifs in Vκ genes on both DNA strands and that the resulting replacement mutations are preferentially selected in the productive repertoire.
In the present study, we examined the targeting of somatic hypermutation in the clonogenic Ig-V sequences of FL and MM, B cell tumors corresponding to advanced (post-immune) differentiation stages in B cell ontogeny. FL is a germinal center (GC) B cell malignancy.14 The neoplastic cells can be considered as counterparts to intra-GC B cells undergoing active selection by antigen; they exhibit ongoing Ig gene hypermutation and bear the t(14;18) chromosomal translocation, which leads to the formation of the hybrid bcl-2/IgH gene.15 Sequence analysis of rearranged VH and Vκ genes in the FL cases under study has demonstrated that VH genes were mostly hypermutated, whereas Vκ genes markedly differed regarding both the mutational load and the distribution of mutations.10 It appears that the potential contribution of FL-Vκ genes in antigen selection of the clonogenic B cells is less important than that of VH genes.12,14 Similar conclusions have been reached from single-cell studies in the normal peripheral B cell repertoire, indicating a more limited mutational load both in the expressed as well as non-functional Vκ genes compared to their partner VH genes in IgM+/CD5− B cells.16 The fact that clonogenic Vκ genes are frequently unmutated indicates that the somatic hypermutation machinery might have ceased to operate in the Vκ locus at the time when neoplastic transformation had occurred. In the majority of FL cases included in our study (7/10; 70%), significant clustering of mutations was observed in the CDRs of either VH or Vκ genes;10 thus, it is reasonable to argue that a complementarity imprint of antigen selection witnessed by the clonogenic VH and VL sequences might constitute an important event in the pathogenesis of FL.14 No physiological analog to this phenomenon has been observed; however, similar observations were made in an analysis of clonogenic VH and VL genes in MM.17 Moreover, as mentioned earlier, FL are characterized by ongoing somatic mutations of their Ig V genes;5 however, as the present analysis was restricted to diagnostic samples no information can be gathered regarding ongoing intraclonal diversification of FL Ig VH and Vκ genes and whether these ongoing mutations are targeted at specific hotspots. It would be rather interesting to examine this issue in future studies.
Multiple myeloma represents a malignancy of the immune system characterized by the presence of a continuously differentiating population of mainly late stage B cells giving rise to plasma cells.18 The analysis of variable heavy chain (VH) region gene rearrangements in MM indicates that, before transformation, the malignant stem cell (whose exact origin remains elusive) has already undergone antigen selection with consistent lack of intraclonal diversification.19,20 In our series, analysis of LC V region genes has revealed somatic hypermutation of almost the same magnitude as that reported by others for VH genes.11,17,21 This finding is in contrast to our observations in FL Vκ genes and offers more direct evidence that MM originates from transformation of late post-GC B cell clones.22 Furthermore, it indicates that hypermutation of VL genes might serve as a surrogate marker of assigning discrete developmental stages regarding GC and post-GC B cell development.23 Our findings regarding the distribution and possible targeting of somatic mutations in rearranged IgV genes of FL and MM conform well with the aforementioned data derived from normal peripheral B cells; for example, the corresponding malignant B cells carrying productive Vκ-Jκ rearrangements and having experienced antigen selection generally exhibit an increased incidence of RGYW targeting in CDRs vs FWRs.13 The differences observed in favor of tetranucleotide (RGYW and WRCY) targeting in FL-VH vs FL-Vκ and MM-Vκ vs FL-Vκ may implicate differences in the evolution of somatic hypermutation coupled with selection in memory (post-GC) vs GC B cells.24,25,26,27,28,29,30,31 Interestingly, this pattern of somatic hypermutation targeting to tetranucletides (H > κ > λ) as well as the trend for strand-biased targeting only in the CDRs of FL-VH genes (RGYW being significantly more mutated than WRCY) and not in LC V genes recapitulates the temporal pattern of Ig gene rearrangements in early B cell ontogeny, where Ig HC genes rearrange first, to be followed by κ and then λ LC genes;32 however, exceptions to this generally occurring pattern of ordered light chain gene rearrangements are known to occur (instead of κ preceding λ, in a minority of cases the λ light chain locus is targeted first by the ‘recombinase’ machinery).33
In the present analysis, hotspot sequences were not always targeted by somatic hypermutation in FL and MM. Several explanations can be offered for this observation: (1) other unrecognized motifs may be targets of somatic hypermutation; (2) despite the fact that the cell of origin in FL and MM has been assigned at an ontogenetic stage postulated to be subject to antigen selection in GCs, it could be argued that many mutations introduced in Vκ genes would not favor selection by antigen under normal conditions;34 neoplastic transformation at this stage might have over-ruled this requirement for cell survival by rescuing these cells from apoptosis, similar to what has been reported for Hodgkin's disease.35 Therefore, these ‘inappropriately’ introduced mutations were fixed and propagated by the neoplastic process; (3) mismatch-repair (MMR) mechanisms may play an important role in the distribution of somatic mutations in GC B cells;36 thus, one may speculate that at an early, MMR-independent phase of somatic hypermutation, mutations are introduced at specific hotspots (G/C biased); later, during a second MMR-dependent stage, mutations are introduced in the target sequence with an A/T bias through the action of an error-prone polymerase. This leads to a redistribution of mutations away from hotspots, thus equilibrating hotspot vs non-hotspot targeting of somatic hypermutation. Therefore, in the absence of MMR genes, diminished mutation accumulation and increased hotspot focusing would be anticipated. MMR gene inactivation and deficiency, leading to the replication-error phenotype, is a common event in B cell malignancies. With this in mind, it would be reasonable to speculate that in some cases of FL and MM the neoplastic B cells might have lost their ability to correct mutations introduced in classic hotspots; conversely, in other cases with intact MMR function, the hotspot to non-hotspot targeting of somatic hypermutation is balanced.
Vλ genes in MM appear to be targets of somatic hypermutation, albeit at sequences outside the recognized hotspots for VH and Vκ genes. This might implicate that Vλ genes are targeted by the somatic hypermutation mechanism at later developmental stages of B cell ontogeny, where the aforementioned hotspots are no longer important in directing mutation targeting.37
The CDR3 of IgH, ie the IgH-V subregion mainly responsible for antigen binding is formed by the junctions of the rearranged VH–D–JH gene segments. In normal B cell development, the amino acid composition of the IgH CDR3 is the main determinant of positive selection.38 Formation of CDR3 region is accomplished during rearrangement of V–(D)–J genes with the insertion of N-nucleotides by terminal deoxynucleotide transferase (TdT), while random deletion and insertion of bases is effected at the borders of the rearranging genes (V/J genes for light chains and V/D/J genes for heavy chains). The randomly selected D gene segment of a rearranged VDJH complex can potentially be found in all three possible reading frames (RFs).39 However, as we have shown, particular D gene RFs can be encountered preferentially in certain B cell lymphoproliferative disorders as a result of the maturation status of the corresponding transformed B cell and selection by antigen.40 The similarly restricted pattern of D gene RF usage (predominantly RF2 and RF3) in both functional IgH junction sequences in FL and MM and non-functional bcl-2/DJH junction sequences in FL suggests that selection forces might affect Ig genes while still in the process of active recombination (DJH complexes); in this context, it has been proposed that pre-B cell clones are submitted to an antigen- independent initial selective pressure, whereby Dμ chains would be merely selected for their ability to interact with surrogate light chains.41 This observation becomes important in indicating that even at early stages of B cell ontogeny, where somatic hypermutation is not operating, selective forces might affect the conformation of V region and prime certain B cell clones to fine-tune their specificity for antigen later, in the germinal center, under the influence of somatic hypermutation.
Tonegawa S . Somatic generation of antibody diversity Nature 1983 302: 575–581
Grawunder U, West RB, Lieber MR . Antigen receptor gene rearrangement Curr Opin Immunol 1998 10: 172–180
Rajewsky K . Clonal selection and learning in the antibody system Nature 1996 381: 751–758
Storb U . Progress in understanding the mechanism and consequences of somatic hypermutation Immunol Rev 1998 162: 5–11
Stevenson F, Sahota S, Zhu D, Ottensmeir C, Chapman C, Oscier D, Hamblin T . Insight into the origin and clonal history of B-cell tumors as revealed by analysis of immunoglobulin variable region genes Immunol Rev 1998 162: 247–259
Chang B, Casali P . The CDR1 sequences of a major proportion of human germline IgVH genes are inherently susceptible to amino acid replacement Immunol Today 1994 15: 367–373
Betz AG, Rada C, Pannell R, Milstein C, Neuberger MS . Passenger transgenes reveal intrinsic specificity of the antibody hypermutation mechanism: clustering, polarity, and specific hot spots Proc Natl Acad Sci USA 1993 90: 2385–2388
Wagner SD, Milstein C, Neuberger MS . Codon bias targets mutation Nature 1995 376: 732
Neuberger NS, Ehrenstein MR, Kllx N, Jolly CJ, Yélamos J, Rada C, Milstein C . Monitoring and interpreting the intrinsic features of somatic hypermutation Immunol Rev 1998 162: 107–116
Stamatopoulos K, Kosmas C, Papadaki T, Pouliou E, Belessi C, Afendaki S, Anagnostou D, Loukopoulos D . Follicular lymphoma immunoglobulin κ light chains are affected by the antigen selection process, but to a lesser degree than their partner heavy chains Br J Haematol 1997 96: 132–146
Kosmas C, Viniou NA, Stamatopoulos K, Courtenay-Luck NS, Papadaki T, Kollia P, Paterakis G, Anagnostou D, Yataganas X, Loukopoulos D . Analysis of κ light chain variable region in multiple myeloma Br J Haematol 1996 94: 306–317
Kosmas C, Stamatopoulos K, Papadaki T, Belessi C, Yataganas X, Anagnostou D, Loukopoulos D . Somatic hypermutation of immunoglobulin genes: focus on follicular lymphoma and multiple myeloma Immunol Rev 1998 162: 281–292
Dörner T, Foster SJ, Brezinschek H-P, Lipsky PE . Analysis of the targeting of the hypermutational machinery and the impact of subsequent selection on the distribution of nucleotide changes in human VHDJH rearrangements Immunol Rev 1998 162: 161–171
Stamatopoulos K, Kosmas C, Belessi C, Kyriazopoulos P, Papadaki T . Molecular insights to the immunopathogenesis of follicular lymphoma Immunol Today 2000 21: 298–305
Stamatopoulos K, Kosmas C, Belessi C, Papadaki T, Afentaki S, Anagnostou D, Loukopoulos D . t(14;18) chromosomal translocation in follicular lymphoma: an event occurring with almost equal frequency both at the D to JH and at later stages in the rearrangement process of the immunoglobulin heavy chain gene locus Br J Haematol 1997 99: 866–872
Klein U, Goossens T, Fischer M, Kanzler H, Braeuninger A, Rajewsky K, Küppers R . Somatic hypermutation in normal and transformed human B cells Immunol Rev 1998 162: 261–280
Sahota SS, Leo R, Hamblin TJ, Stevenson FK . Myeloma VL and VH sequences reveal a complementary imprint of antigen selection in tumor cells Blood 1997 89: 219–226
Kosmas C, Stamatopoulos K, Stavroyianni N, Zoi K, Belessi C, Viniou N, Kollia P, Yataganas X . Origin and diversification of the clonogenic cell in multiple myeloma: lessons from the immunoglobulin repertoire Leukemia 2000 14: 1718–1726
Bakkus MHC, Van Riet I, Van Camp B, Thielemans K . Evidence that the clonogenic cell in multiple myeloma originates from a pre-switched but somatically mutated B cell Br J Haematol 1994 87: 68–74
Vescio RA, Cao J, Hong CH, Lee JC, Wu CH, Der-Danielian M, Wu V, Newman R, Lichtenstein AK, Berenson JR . Myeloma Ig heavy chain V region sequences reveal prior antigenic selection and marked somatic mutation but no intraclonal diversity J Immunol 1995 155: 2487–2497
Wagner SD, Martinelli V, Luzzatto L . Similar patterns of Vκ gene usage but different degrees of somatic mutation in hairy cell leukemia, prolymphocytic leukemia, Waldenström's macroglobulinemia, and myeloma Blood 1994 83: 3647–3653
Kosmas C, Stamatopoulos K, Stavroyianni N, Belessi C, Viniou N, Yataganas X . Molecular analysis of immunoglobulin genes in multiple myeloma Leuk Lymphoma 1999 33: 253–263
Küppers R, Klein U, Hansmann M-L, Rajewsky K . Cellular origin of human B-cell lymphomas N Engl J Med 1999 341: 1520–1529
Kelsoe G . V(D)J hypermutation and receptor revision: coloring outside the lines Curr Opin Immunol 1999 11: 70–75
Milstein C, Neuberger MS, Staden R . Both DNA strands of antibody genes are hypermutation targets Proc Natl Acad Sci USA 1998 95: 8791–8794
Nakamura N, Kuze T, Hashimoto Y . Hara V, Hoshi S, Sasaki Y, Shirakawa A, Seto M, Abel M. Analysis of the immunoglobulin heavy chain gene variable region of CD5-positive and -negative diffuse large B cell lymphoma Leukemia 2001 16: 452–457
Capello D, Fais F, Vivenza D, Migliaretti G, Chiorazzi N, Gaidano G, Ferrarini M . Identification of three subgroups of B cell chronic lymphocytic leukemia based upon mutations of BCL-6 and IgV genes Leukemia 2000 14: 811–815
Fais F, Gaidano G, Capello D, Gloghini A, Ghiotto F, Roncella S, Carbone A, Chiorazzi N, Ferrarini M . Immunoglobulin V region gene use and structure suggest antigen selection in AIDS-related primary effusion lymphomas Leukemia 1999 13: 1093–1099
Driessen A, Tierens A, Ectors N, Stul M, Pittaluga S, Geboes K, Delabie J, De Wolf-Peeters C . Primary diffuse large B cell lymphoma of the stomach: analysis of somatic mutations in the rearranged immunoglobulin heavy chain variable genes indicates antigen selection Leukemia 1999 13: 1085–1092
Kosmas C, Stamatopoulos K . Immunoglobulin light chain variable region genes in multiple myeloma Leukemia 1999 13: 827–830
Kon S, Sasamori T, Kasai K, Yamano H, Endo T, Kon H, Kikuchi K . Ongoing somatic mutations of the immunoglobulin gene in MALT lymphoma with widespread MLP type polypoid lesions Leukemia 1998 12: 1495–1497
Korsmeyer SJ, Hieter PA, Ravetch JV, Poplack DG, Waldmann TA, Leder P . Developmental hierarchy of immunoglobulin gene rearrangements in human leukemic pre-B-cells Proc Natl Acad Sci USA 1981 78: 7096–7100
Pauza ME, Rehmann JA, LeBien TW . Unusual patterns of immunoglobulin gene rearrangement and expression during human B-cell ontogeny: human B-cells can simultaneously express cell surface κ and λ light chains J Exp Med 1993 178: 139–149
Nossal GJV . Negative selection of lymphocytes Cell 1994 76: 229–239
Kanzler H, Küppers R, Hansmann M-L, Rajewsky K . Hodgkin's and Reed–Sternberg cells in Hodgkin's disease represent the outgrowth of a dominant tumor clone derived from (crippled) germinal center B cells J Exp Med 1996 184: 1495–1505
Rada C, Ehrenstein MR, Neuberger MS, Milstein C . Hot spot focusing of somatic hypermutation in MSH2-deficient mice suggests two stages of mutational targeting Immunity 1998 9: 135–141
Farner NL, Dörner T, Lipsky PE . Molecular mechanisms and selection influence the generation of the human Vλ–Jλ gene repertoire J Immunol 1999 162: 2137–2145
Padlan EA . On the nature of antibody combining sites: unusual structural featutres that may confer on these sites an enhanced capacity for binding ligands Proteins 1990 7: 112–124
Raaphorst FM, Raman CS, Nall BT, Teale JM . Molecular mechanisms governing reading frame choice of immunoglobulin diversity genes Immunol Today 1997 18: 37–43
Stamatopoulos K, Kosmas C, Stavioyianni N, Belessi C, Papadaki T . Selection of immunoglobulin diversity gene reading frames in B-cell lymphoproliferative disorders Leukemia 1999 13: 601–604
Millili M, Schiff C, Fougereau M, Tonnelle C . The VDJ repertoire expressed in human pre B cells reflects the selection of bona fide heavy chains Eur J Immunol 1996 26: 63–69
About this article
Ig gene diversification and selection in follicular lymphoma, diffuse large B cell lymphoma and primary central nervous system lymphoma revealed by lineage tree and mutation analyses
International Immunology (2010)
Blood Cells, Molecules, and Diseases (2008)
Characterization of monoclonal anti-β2-glycoprotein-I and anti-prothrombin antibody fragments generated by phage display from a patient with primary antiphospholipid syndrome
Journal of Autoimmunity (2006)
Molecular Medicine (2004)
American Journal of Clinical Pathology (2003)