Analysis of IgVH gene mutations in B cell chronic lymphocytic leukaemia according to antigen-driven selection identifies subgroups with different prognosis and usage of the canonical somatic hypermutation machinery

Analysis of IgVH gene mutations in B cell chronic lymphocytic leukaemia according to antigen-driven selection identifies subgroups with different prognosis and usage of the canonical somatic hypermutation machinery
of 14
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Analysis of IgV H  gene mutations in B cell chronic lymphocyticleukaemia according to antigen-driven selection identifiessubgroups with different prognosis and usage of the canonicalsomatic hypermutation machinery B-cell chronic lymphocytic leukaemia (B-CLL) cells weresrcinally considered to be the neoplastic counterpart of Blymphocytes unable to undergo somatic hypermutation (SHM;Hamblin, 2002; Chiorazzi & Ferrarini, 2003). The demonstra-tion that approximately 50% of B-CLL cases express a mutatedconfiguration of immunoglobulin (Ig)V H  genes (Schroeder & Dighiero, 1994; Oscier  et al  , 1997; Fais  et al  , 1998) changedthis view suggesting the existence of at least two distinct B-CLLsubsets that probably srcinated from pre- or post-germinalcentre (GC) B cells (Naylor & Capra, 1999). This differenthistogenesis has important prognostic implications, as B-CLLcases with a mutated (M) configuration of IgV H  genes have abetter prognosis than those with unmutated (UM) B-CLL(Damle  et al  , 1999; Hamblin  et al  , 1999; Maloum  et al  , 2000;Krober  et al  , 2002; Lin  et al  , 2002). More recently, theexpression of a homogeneous phenotype and gene profilerelated to experienced B cells virtually in all B-CLL cases (Klein et al  , 2001; Rosenwald  et al  , 2001; Damle  et al  , 2002) havestrongly questioned such a pathogenetic model, rather favour-ing a common transformation mechanism for both M and UMB-CLLs.The SHM process, as it occurs during a T-cell-dependentGC differentiation of normal B cells, introduces mutations in anon-random manner generating several biased nucleotidesubstitutions. Among them, the skewing of replacement (R)mutations from framework (FR) to complementarity-deter-mining regions (CDR) of IgV H  genes is the result of a positiveselection operated by the antigen, aimed at preserving the Massimo Degan, 1 Riccardo Bomben, 1 Michele Dal Bo, 1 Antonella Zucchetto, 1 Paola Nanni, 1 Maurizio Rupolo, 2 Agostino Steffan, 3 Vincenza Attadia, 1 Pier Ferruccio Ballerini, 4 DanielaDamiani, 5 Carlo Pucillo, 6 Giovanni DelPoeta, 7 Alfonso Colombatti 8 and ValterGattei 1 1 Clinical and Experimental Haematology Research Unit,  2  Medical Oncology,  3 Blood Bank,IRCCS, Aviano (PN),  4 Division of Internal  Medicine, De Gironcoli Hospital, Conegliano(TV),  5 Division of Hematology, and   6  Laboratory of Immunology, University of Udine, Udine, 7  S. Eugenio Hospital, University of Tor Vergata,Rome, and   8 Experimental Oncology 2 and Centrodi Riferimento Oncologico, IRCCS, Aviano (PN),Italy  Received 23 January 2004; accepted forpublication 22 March 2004Correspondence: Valter Gattei MD, Clinical andExperimental Hematology Research Unit,Centro di Riferimento Oncologico, IRCCS, ViaPedemontana Occidentale, 12, Aviano (PN),Italy. E-mail: vgattei@cro.it Summary Cases of B-cell chronic lymphocytic leukaemia (B-CLL) with mutated (M)IgV H  genes have a better prognosis than unmutated (UM) cases. We analysedthe IgV H  mutational status of B-CLL according to the features of a canonicalsomatic hypermutation (SHM) process, correlating this data with survival. Ina series of 141 B-CLLs, 124 cases were examined for IgV H  gene per centmutations and skewing of replacement/silent mutations in the framework/complementarity-determining regions as evidence of antigen-drivenselection; this identified three B-CLL subsets: significantly mutated (sM),with evidence of antigen-driven selection, not significantly mutated (nsM)and UM, without such evidence and IgV H  gene per cent mutations above orbelow the 2% cut-off. sM B-CLL patients had longer survival within the goodprognosis subgroup that had more than 2% mutations of IgV H  genes. sM,nsM and UM B-CLL were also characterized for the biased usage of IgV H families, intraclonal IgV H  gene diversification, preference of mutations totarget-specific nucleotides or hotspots, and for the expression of enzymesinvolved in SHM (translesion DNA polymerase  f  and  g  and activation-induced cytidine deaminase). These findings indicate the activation of acanonical SHM process in nsM and sM B-CLLs and underscore the role of the antigen in defining the specific clinical and biological features of B-CLL. Keywords :  chronic lymphocytic leukaemia, somatic hypermutation, antigenselection, translesion DNA polymerases, prognosis. research paper ª  2004 Blackwell Publishing Ltd,  British Journal of Haematology  ,  126 , 29–42 doi:10.1111/j.1365-2141.2004.04985.x  structural integrity of Igs by concomitantly providing a highmutational rate in antigen-binding regions (Chang & Casali,1994; Dorner  et al  , 1998a; Lossos  et al  , 2000). Moreover, theexcess of transitions over transversions, as well as of mutationstargeting specific nucleotides or nucleotide motifs, may beexplained by the activity of a specific mutational machinery (Smith  et al  , 1996; Dorner  et al  , 1998b; Neuberger  et al  , 1998).This includes the DNA polymerases (pol)  f  and  g , which areefficient translesion and error-prone DNA polymerases, i.e.capable of carrying out lesion bypassing DNA synthesis by copying DNA with a high error rate (Zan  et al  , 2001; Zeng et al  , 2001; Storb & Stavnezer, 2002). Moreover, the activation-induced cytidine deaminase (AID), a member of the cytidinedeaminase family exclusively expressed by GC B cells, isanother essential enzyme for both SHM and class switchrecombination of normal B cells (Muramatsu  et al  , 1999, 2000;Cerutti  et al  , 2002; Storb & Stavnezer, 2002).While some reports indicate a broad expression of AID inB-CLL (Albesiano  et al  , 2003; Oppezzo  et al  , 2003), with thehighest levels being confined to UM CD38 high subsets(McCarthy   et al  , 2003; Degan  et al  , 2004), expression of translesion DNA polymerases has not been yet investigated inthis disease. However, the finding of an intraclonal V H DJ H gene diversification in most B-CLL cases (Gurrieri  et al  , 2002)suggests the presence of mutational machinery that is stillfunctionally active following neoplastic transformation. Finally,evidence of antigen-driven selection in B-CLL, as evaluated by computing the unbalanced distribution of R/silent (S) muta-tions in CDR/FR, although reported by some studies (Hamblin et al  , 1999; Gurrieri  et al  , 2002; Messmer  et al  , 2004), hasnever been correlated with survival.In the present study, the IgV H  mutational status wasinvestigated in a large cohort of B-CLLs by using a strategy of cloning and sequencing of at least 5–10 transcripts for eachpatient. By extending some of our preliminary observations(Gattei  et al  , 2003), we demonstrated that the application of statistical models aimed at evaluating the impact of positiveselection operated by the antigen (Chang & Casali, 1994;Lossos  et al  , 2000) enabled the identification of three B-CLLsubsets, each with a different prognosis and usage of thecanonical SHM machinery. Patients, materials and methods B-CLL patients and cell samples and immunophenotypical analysis The study was performed on peripheral blood (PB) samples,collected after informed consent for routine diagnostic andfollow-up procedures, from 141 patients affected by typical B-CLL. Diagnosis was made according to standard criteria, inclu-ding expression of CD19, CD5, CD23 and low levels of clonally restricted surface Igs (Matutes  et al  , 1994). The median age atdiagnosiswas64 years(range32–97 years)andthemaletofemaleratiowas1 Æ 1(75males,66females).Theclinicalstageatdiagnosisaccording to Rai  et al   (1975) was available for 127 of 141 patientsas follows: stage 0, 33% ( n  ¼  42); stage I, 39 Æ 4% ( n  ¼  50);stage II, 23 Æ 6% ( n  ¼  30); stage III, 1 Æ 6% ( n  ¼  2); stage IV, 2 Æ 4%( n  ¼  3). Mononuclear cells were isolated by centrifugation onFicoll–Hypaque (Pharmacia, Uppsala, Sweden) gradient andeither used directly or cryopreserved in liquid nitrogen until use.Expression of phenotypic markers, including CD38, was inves-tigated by flow cytometry, as previously described (Gattei  et al  ,1997; Damle  et al  , 1999; Del Poeta  et al  , 2001). Cases in whichCD38 was detected in <30% of cells were identified as CD38 low  ,while cases with  ‡ 30% CD38 B-CLL cells were consideredCD38 high (Damle  et al  , 1999; Del Poeta  et al  , 2001).  Amplification, cloning and sequencing of V  H  DJ  H  transcripts Total RNA was extracted and reverse-transcribed as described(Gattei  et al  , 1997). The resulting cDNAs, checked for first-strand synthesis (Degan  et al  , 2000), were amplified using amixture of sense primers annealing either to the V H 1 throughV H 6 leader sequences or to the 5 ¢ -end of V H 1–V H 6 FR1utilized in conjunction with a mixture of antisense primerscomplementary to the germ line J H  regions, as reported (Fais et al  , 1998; Damle  et al  , 1999; Hamblin  et al  , 1999; Gurrieri et al  , 2002; Krober  et al  , 2002). The purified amplifiedproducts, inserted into the PCR2.1-TOPO vector (InvitrogenS.R.L., San Giuliano Milanese, Milan, Italy), were expanded inTOP10 One Shot TM competent cells (Invitrogen) and cloned.Plasmid DNAs were isolated from overnight cultures of randomly selected colonies and sequenced by using anautomatic DNA sequencer (ABI PRISM 3100, AppliedBiosystem, Foster City, CA, USA; Beckman CEQ2000XL,Fullerton, CA, USA). When tested for the DNA polymeraseerror rate according to Pasqualucci  et al   (1998), this strategy  yielded an incorporating error rate of 2 Æ 8  ·  10 ) 4 /bp. Inadditional experiments, plasmid DNA from two distinctcolonies from two different B-CLL patients (B176 and B211)were amplified and cloned by exactly reproducing the proce-dures described above. Among 20 sequenced colonies(6732 bp), a single mutation was detected (DNA polymeraseerror rate  ¼  1 Æ 5  ·  10 ) 4 /bp). Comparisons between the ob-tained sequences and those of the various germ line IgV H DJ H genes were performed with the Ig basic local alignment searchtool (IgBLAST) directory (http://www.ncbi.nlm.nih.gov/ig-blast) utilizing the MacVector 7.1 sequence analysis software(Accelrys; Symantec Co., Accelrys European Headquarters,Cambridge, UK). According to this algorithm, a stretch of seven or more nucleotides was required for D and J H assignment. Only when the same rearrangement was identifiedin at least 5–10 clones (three or four clones in 11 cases), a givenIgV H  sequence was further analysed. As described (Gurrieri et al  , 2002), alignment of the IgV H  sequences available for eachpatient revealed, along with mutations shared by all thetranscripts analysed, a number of partially shared or uniquemutations. For this reason, all mutational analyses (see below) M. Degan  et al  30  ª  2004 Blackwell Publishing Ltd,  British Journal of Haematology  ,  126 , 29–42  were carried out in each IgV H  transcript separately, and the percent mutation assigned to a given B-CLL was the mean value of the per cent mutations found in each individual transcript. IgV  H   mutation analysis A census of somatic point mutations was performed by solely considering the IgV H  sequence, i.e. from FR1 to FR3. Asevidence of antigen-driven selection (Chang & Casali, 1994;Dorner  et al  , 1998a; Lossos  et al  , 2000; Gurrieri  et al  , 2002),the probability that an excess or scarcity of R mutationsoccurred by chance in IgV H  CDR or FR, respectively, wascomputed by applying the available binomial and multinomialdistribution models (Chang & Casali, 1994; Lossos  et al  , 2000).Both these models calculate the expected frequency of mutations by taking into account the base composition of agiven germ line IgV H  sequence (Chang & Casali, 1994; Lossos et al  , 2000). In the absence of a positive selective pressureoperated by the antigen, nucleotide changes yielding R orS mutation will be distributed throughout the coding sequencewith a pattern superimposible to that of randomly occurringmutations ( P   > 0 Æ 05). Conversely, a skewing of R mutationfrom FR to CDR and/or of S mutations from CDR to FR willindicate a positive selection operated by the antigen; in thiscase, the probability that these mutations are due by chancewill be <0 Æ 05 (Chang & Casali, 1994; Lossos  et al  , 2000). In thepresent study,  P  -values were computed by applying themultinomial distribution model, taking advantage of theready-to-use computer program at the http://www-stat.stan-ford.edu/immunoglobulin website (Lossos  et al  , 2000). Due tosome flaws of the available JAVA applet, the multinomialmodel was not appropriate when analysing specific IgV H  (3–53and 4–39, overall accounting for nine cases of 77 M B-CLL inour series). In these cases, calculations of biased R/S mutationswere made according to the binomial distribution modelutilizing the reported formula (Chang & Casali, 1994). Theexpected values of transitions or transversions, of mutationstargeting specific nucleotides as well as the RGYW (i.e. purine/G/pyrimidine/A or T) or its inverse (WRCY) motifs werecomputed as reported (Smith  et al  , 1996; Dorner  et al  , 1998b;Neuberger  et al  , 1998). Expression of translesion DNA polymerases  g  and   f  and  AID Expression of DNA pol  g  and  f  was evaluated by quantitativereal-time polymerase chain reaction (QRT-PCR), while AIDexpression was investigated by conventional reverse transcrip-tion-PCR (RT-PCR) on a selected cohort of B-CLL cases (52cases). To avoid interference with normal B and T cells thatallegedly express these enzymes, only cases with more than 95%of neoplastic cells in PB samples were analysed. Briefly, cDNAsfrom B-CLL cells were amplified with primer pairs specific forpol  g , pol  f  and the housekeeping gene  b 2 -microglobulin( b 2 M), as described (Zan  et al  , 2001; Spessotto  et al  , 2002). Theincorporation of the SYBR   Green dye (Applied Biosystems)into the PCR products was monitored in real-time with asequence detection system (ABI PRISM 7700), resulting in acalculation of threshold cycles (Ct value), i.e. the PCR cyclenumber corresponding to the beginning of the exponentialgrowth of PCR products. Ct values were converted intoattomoles by means of QRT-PCR experiments carried outwith serial dilution of known concentrations of pol  g , pol  f  and b 2 M-specific amplicons. Results were reported as the ratio of the relative expression levels of pol  g / b 2 M or pol  f / b 2 M.Activation-induced cytidine deaminase expression waschecked in the same cohort of cases essentially as described(Albesiano  et al  , 2003; McCarthy   et al  , 2003; Oppezzo  et al  ,2003). Briefly, cDNAs were amplified with primers chosen toanneal to sequences corresponding to the first (sense, position45–64) and the fifth (antisense, position 675–653) exons of AID-specific cDNA (Albesiano  et al  , 2003; McCarthy   et al  ,2003; Oppezzo  et al  , 2003). Amplified products (10  l l) wererun on 1 Æ 5% ethidium bromide-stained agarose gels andanalysed under ultra-violet light. Statistical analysis Complete survival data were available for 141 patients. Themean follow-up was 67 months (range 8–220 months). Patientsurvivals were analysed using the Kaplan–Meier survival curvesand log-rank test (Armitage & Berry, 1987). Correlationbetween IgV H  mutational status and CD38 expression wasperformed by the chi-square test (Armitage & Berry, 1987).Differences between expected and observed values of transi-tions and transversions, mutations targeting specific nucleo-tides, as well as RGYW/WRCY motifs in IgV H  sequences wereassessed by the Student’s  t  -test for paired samples, by separately evaluating each transcript (Armitage & Berry,1987). Differences between the relative expression levels of pol  f / g  ratios were assessed by the Student’s  t  -test for unpairedsamples. Statistical significance was always accepted when P   < 0 Æ 05. Results IgV  H   per cent mutations, biased IgV  H   usage, intraclonal IgV  H   diversification, CD38 expression and survivals The mutational status of IgV H  genes was investigated in 124 of 141 B-CLLs by sequencing at least 5–10 transcripts (three tofour in 11 cases) for each patient. Values of IgV H  genemutations (Fig 1), represent the mean values of per centmutations due to a certain degree of intraclonal IgV H diversification (see below). According to this analysis, 47B-CLL cases (38%) had neoplastic cells bearing less than 2%average mutations when compared with the nearest germ linesequence (UM B-CLL), while 77 cases (62%) displayed  ‡ 2%average mutations and were classified as M B-CLLs (Damle et al  , 1999; Hamblin  et al  , 1999; Maloum  et al  , 2000; Krober IgV H  Gene Mutations and Prognosis in B-CLL ª  2004 Blackwell Publishing Ltd,  British Journal of Haematology  ,  126 , 29–42  31  et al  , 2002). Tables I and II summarize V H DJ H  assignmentsand the means of per cent mutations for UM and M B-CLLsrespectively.In agreement with previous reports (Keating  et al  , 2003;Messmer  et al  , 2004), a higher number of B-CLLs expressingthe V H 1 family was found in the UM, when compared with theM, B-CLL subgroups (37 Æ 5% vs. 10 Æ 1%; Table III). Conversely,M B-CLLs preferentially expressed V H  genes belonging to theV H 3 family (58 Æ 2% vs. 31 Æ 3%; Table III). No differences wereobserved between UM and M B-CLLs regarding the expressionof other IgV H  families, including V H 4 (about 20% in bothsubgroups) and other rarely expressed IgV H  genes, such asV H 2, V H 5 and V H 6 (Table III).Table II also reports a detailed analysis of IgV H  pointmutations for the M B-CLL group. According to Gurrieri  et al  (2002), they were divided into ‘shared’ (i.e. mutations found inall the IgV H  transcripts analysed), ‘partially shared’ (i.e.mutations present in a part of the analysed IgV H  transcripts)and ‘unique’ (i.e. mutations found only in distinct IgV H transcripts) mutations (Tables II and IV). Among 79 IgV H (449 IgV H  transcripts), 63 (about 80%) exhibited either nointraclonal IgV H  diversification (i.e. shared mutations only, 13IgV H , 16 Æ 5%) or a low number of unique and/or partially shared mutations, always in the presence of a great excess of shared mutations (Tables II and IV). Conversely, in 16 of 79IgV H , the sum of unique and partially shared mutationsexceeded the number of shared mutations (Tables II and IV),thus revealing a higher degree of intraclonal IgV H  diversifica-tion.Data on CD38 expression was available for 140 of 141patients, of which 51 cases (36 Æ 4%) were CD38 high B-CLL, andthe remaining 89 cases (63 Æ 6%) were classified as CD38 low  B-CLL (Damle  et al  , 1999; Del Poeta  et al  , 2001). In agreementwith some of the current literature (Damle  et al  , 1999), wefound a statistically significant correlation between CD38 high and low mutations of IgV H  genes by applying a chi-square test( P   < 0 Æ 01; Fig 1).A comparison of patient survivals with CD38 expression(140 cases) or IgV H  mutational status (124 cases) alsodemonstrated the positive impact of CD38 low  and IgV H mutations above the standard cut-off of 2% on prognosis(Fig 2) (Damle  et al  , 1999; Del Poeta  et al  , 2001; Krober  et al  ,2002; Lin  et al  , 2002).Altogether, IgV H  per cent mutations and usage, intraclonalIgV H  diversification, CD38 expression and survivals, as foundon our cohort of B-CLL patients, were in keeping with theabove reports. Evidence of antigen-driven selection identified three B-CLLsubsets The hypothesis that M B-CLL cases may derive as a transfor-mation of a competent B cell (Naylor & Capra, 1999), led us toinvestigate whether the IgV H  gene mutational status found in Fig 1.  CD38 expression and IgV H  mutational status in B-cell chronic lymphocytic leukaemia (B-CLL). B-CLL cases were arranged according to theIgV H  mutational status, expressed as per cent mutation (upper panel); the lower panel indicates, for each B-CLL case, the corresponding expressionlevel for CD38, reported as percentage of positive B cells. Cases for which IgV H  mutational status and CD38 expression data were not simultaneously available are reported as open histograms. Dotted lines refer to the established cut-offs of 2% mutation (upper panel) and 30% CD38 + cells (lowerpanel). Asterisks indicate three cases (B241, B235, B260) with biclonal disease, whose per cent mutations were separately reported. A chi-squarecorrelation between IgV H  mutational status (cut-off: 2% mutations) and CD38 expression (cut-off: 30% of positive cells) is also reported ( P   < 0 Æ 01).n.a., not available. M. Degan  et al  32  ª  2004 Blackwell Publishing Ltd,  British Journal of Haematology  ,  126 , 29–42  B-CLL cells was consistent with that of B cells positively selected by the antigen (Chang & Casali, 1994; Dorner  et al  ,1998a; Lossos  et al  , 2000). Within IgV H  genes, sequencesencoding for the antigen-contacting CDRs are inherently moreprone to R mutations than those encoding for the respectiveFRs (Chang & Casali, 1994), this susceptibility being furtherstressed by antigen selection (Chang & Casali, 1994; Dorner et al  , 1998a; Lossos  et al  , 2000). In our study, among 77 MB-CLLs, 449 IgV H  transcripts were analysed by applying themultinomial distribution model to assess the probability thatbiased R/S mutations in IgV H  gene CDR/FR arise randomly (Lossos  et al  , 2000). When the multinomial model gave aresult that was not appropriate (52/449 IgV H  transcripts,11 Æ 5%), we took advantage of the binomial distribution modelaccording to Chang and Casali (1994). Overall, a given IgV H transcript was defined as having evidence of antigen-drivenselection if the applied statistical model yielded a significantresult ( P   < 0 Æ 05) either in CDRs or FRs (Dorner  et al  , 1998a;Sakai  et al  , 2000).By using such an approach, we identified 52 of 77 M B-CLLsfor which evidence of antigen-driven selection was found in allor in part of the IgV H  transcripts analysed (Table II). TheseB-CLL cases were classified as ‘significantly mutated’ (sMB-CLL).Table IIsummarizesthecharacteristicsofthesMB-CLLcases, including V H DJ H  gene assignment, per cent mutations,number of IgV H  transcripts analysed and number of IgV H transcripts showing evidence of antigen-driven selection inCDR and/or FR. Among 289 analysed IgV H  transcripts(average 5 Æ 7, range 3–10), 247 showed a significant excess of R mutations in CDR (25 IgV H  transcripts) or a scarcity of R mutations in FR (114 IgV H  transcripts) or both (108 IgV H transcripts), while the remaining 42 IgV H  transcripts did notdisplay evidence of antigen-driven selection. Taken together,36 of 52 sM B-CLLs (69 Æ 2%) showed evidence of antigen-driven selection in all of the transcripts analysed, while in 16cases this evidence was documented only in some IgV H transcripts (Table II). The remaining B-CLL cases wereclassified as ‘not significantly mutated’ (nsM, 25 cases) or‘unmutated’ (UM, 47 cases) when the per cent mismatch wasabove or below the 2% cut-off respectively. For all these cases,the average number of IgV H  transcripts analysed for eachB-CLL patients was 6 Æ 1 (range 5–10) and none showedevidence of antigen-driven selection using the two availablestatistical models. Characterization of sM and nsM B-CLLs: biased IgV  H  usage and intraclonal IgV  H   diversification The sM subgroup of B-CLL cases displayed a pattern of IgV H usage similar to that reported for M B-CLLs (Table III).However, when compared with M B-CLL, a more evidentskewing towards the preferential usage of IgV H 3 family wasdocumented in sM B-CLL. In fact, almost 2/3 of sM B-CLLsrearranged IgV H  genes of the V H 3 family, whereas nsM B-CLLs Table I.  Mutation characteristics of IgV H DJ H  genes in the unmutatedB-cell chronic lymphocytic leukaemia (B-CLL) subgroup.CLL caseV H DJ H  gene% mutation*V H  D J H B 48 1-18 3-10 4 0 Æ 0B 201 3-15 3-3 4 0 Æ 0B 206 3-11 6-6 6 0 Æ 0B 265 3-73 3-22 3 0 Æ 0B 210 3-53 3-3 4 0 Æ 1B 16 3-13 3-3 4 0 Æ 2B 209 1-2 6-19 4 0 Æ 2B 232 4-39 6-13 5 0 Æ 2B 169 1-2 1-26 4 0 Æ 3B 205 3-21 5-24 6 0 Æ 3B 47 1-18 3-10 4 0 Æ 4B 149 4-59 3-22 6 0 Æ 4B 42 4-39 6-13 5 0 Æ 5B 255 5-51 3-22 4 0 Æ 5B 249 5-51 3-22 4 0 Æ 5B 202 1-69 3-16 3 0 Æ 6B 179 4-39 6-13 4 0 Æ 7B 191 3-11 3-9 6 0 Æ 7B 241 2-26 6-13 4 0 Æ 72-5 1-7 4 1 Æ 9B 116 1-69 2-2 2 0 Æ 8B 173 3-30 7-21 1 0 Æ 8B 220 1-69 3-22 6 0 Æ 8B 128 3-33 3-3 3 0 Æ 9B 159 6-1 2-2 4 0 Æ 9B 21 3-23 5-10 4 0 Æ 9B 196 4-34 6-13 6 1 Æ 0B 223 1-69 n.a. 3 1 Æ 0B 118 1-2 2-21 6 1 Æ 1B 18 1-2 1-26 6 1 Æ 1B 176 1-69 6-13 3 1 Æ 1B 83 3-23 2-8 6 1 Æ 1B 119 4-31 3-3 5 1 Æ 1B 166 4-39 4-17 5 1 Æ 1B 158 1-69 7-27 1 1 Æ 1B 244 3-30 n.a. 5 1 Æ 1B 161 4-59 3-22 4 1 Æ 2B 183 1-69 3-16 4 1 Æ 3B 245 3-33 7-27 4 1 Æ 3B 174 1-69 3-3 6 1 Æ 3B 17 1-46 3-10 6 1 Æ 4B 14 4-34 3-3 6 1 Æ 6B 110 1-18 6-19 4 1 Æ 6B 15 3-21 5-18 6 1 Æ 6B 184 3-48 2-8 4 1 Æ 7B 165 1-2 3-10 5 1 Æ 7B 39 1-2 6-13 4 1 Æ 7B 75 4-39 2-8 4 1 Æ 7n.a., not assigned.*Values represent the mean value of 5–10 transcripts analysed for eachB-CLL case. IgV H  Gene Mutations and Prognosis in B-CLL ª  2004 Blackwell Publishing Ltd,  British Journal of Haematology  ,  126 , 29–42  33
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks