HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Macintyre, E. Articles by Morris, S. W. Search for Related Content PubMed PubMed Citation Articles by Macintyre, E. Articles by Morris, S. W. Social Bookmarking                   What's this?

Non-Hodgkin's Lymphoma: Molecular Features of B Cell Lymphoma

Abstract

The rapid increase in the incidence of the B cell non-Hodgkin's lymphomas (NHL) and improved understanding of the mechanisms involved in their development renders timely a review of the theoretical and practical aspects of molecular abnormalities in B cell NHL.

In Section I, Dr. Macintyre addresses the practical aspects of the use of molecular techniques for the diagnosis and therapeutic management of patients with B cell NHL. While detection of clonal Ig rearrangements is widely used to distinguish reactive from malignant lymphoproliferative disorders, molecular informativity is variable. The relative roles of cytogenetic, molecular and immunological techniques in the detection of genetic abnormalities and their protein products varies with the clinical situation. Consequently, the role of molecular analysis relative to morphological classification is evolving. Integrated diagnostic services are best equipped to cope with these changes. Recent evidence that large scale gene expression profiling allows improved prognostic stratification of diffuse large cell lymphoma suggests that the choice of diagnostic techniques will continue to change significantly and rapidly.

In Section II, Dr. Willerford reviews current understanding of the mechanisms involved in immunoglobulin (Ig) gene rearrangement during B lymphoid development and the way in which these processes may contribute to Ig-locus chromosome translocations in lymphoma. Recent insights into the regulation of Ig gene diversification indicate that genetic plasticity in B lymphocytes is much greater than previously suspected. Physiological genomic instability, which may include isotype switching, recombination revision and somatic mutation, occurs in germinal centers in the context of immune responses and may explain longstanding clinical observations that link immunity and lymphoid neoplasia. Data from murine models and human disorders predisposing to NHL have been used to illustrate these issues.

In Section III, Dr. Morris reviews the characteristics and consequences of deregulation of novel "proto-oncogenes" involved in B cell NHL, including PAX5 (chromosome 9p 13), BCL8 (15q11-q13), BCL9, MUC1, FcRIIB and other 1q21-q22 genes and BCL10 (1p22). The AP12-MLT/MALT1 [t(11;18)(q21;q21)] fusion transcript is also described.

I. Clinical Impact of Molecular Diagnostics in the Management of Non-Hodgkin's Lymphomas

Elizabeth Macintyre, M.D.^*

Molecular techniques are of increasing practical importance in the analysis of non-Hodgkin's lymphomas (NHL), both for diagnostic purposes and in order to accurately evaluate prognosis. In addition, clonal or disease-specific abnormalities can provide markers for the detection of minimal residual disease. An increasing number of largely complementary techniques, including Southern blotting, PCR amplification of DNA or RNA and fluorescent in-situ hybridization (FISH), are responsible for these advances. In view of the number of investigational techniques available it has become necessary to evaluate their relative roles and to assess their contribution with respect to classical techniques, including morphology, immunology and karyotype analysis. This presentation will be limited to evaluation of cellular events that directly or indirectly reflect the oncogenic process; molecular detection of viral agents that play a role in lymphomagenesis will not be discussed. The former can be divided into oncogenic events, as detailed in Section III, and physiological Ig and TCR rearrangements that provide useful clonal markers but are not directly involved in lymphoma development, other than as described in Section II.

Clonal Markers
The vast majority of NHL have undergone physiological, clonal immunoglobulin (Ig) or T Cell Receptor (TCR) rearrangements (Section II). The identification of a clonal Ig/TCR rearrangement (lymphoid clonality) is therefore widely used at diagnosis and increasingly for follow-up. It probably represents the most frequently performed diagnostic molecular analysis in NHL.

Technical Aspects
Until recently, detection of lymphoid clonality was largely based on Southern blot analysis of predominantly IgH and TCRß genes. The technique is based on detection of aberrantly sized DNA fragments following restriction endonuclease digestion and hybidization with labeled probes. Both partial D-J and complete V(D)J rearrangements are detected and somatic mutation does not prevent hybridization. Chromosomal translocations involving Ig/TCR are also detected. The limited repertoire of TCR and TCR genes means that it is possible to determine the V and J segments involved in a given clonal rearrangement based on the size of DNA fragments obtained but that it is difficult to distinguish minor clonal populations from polyclonal TCR populations using a single V-J combination, since Southern analysis will not distinguish junctional diversity. The Southern technique is, however, long, technically demanding, requires a considerable amount of material (approximately 20µg) and, although its optimal sensitivity is approximately 5%, it is often less sensitive. It is therefore increasingly replaced by PCR analysis. Our current practice for detection of lymphoid clonality is to use Southern blotting to further investigate discordant or suspected false-negative IgH results and to evaluate potential IgH oligoclonality, which is poorly identified by consensus PCR strategies.

PCR detection of Ig/TCR V(D)J gene rearrangements exploits the fact that the CDR3 or N regions are clone specific and highly variable, both with regard to length and content. By using V- and J-specific or consensus primers and separation of PCR products on the basis of size it is possible to distinguish clonal from reactive populations. Amplification of a clonal lymphoid population generates a discrete band on electrophoresis whereas polyclonal rearrangements generate a smear of variably sized products. Amplification is performed from DNA, although RT-PCR amplification of Ig/TCR transcripts using V and C region primers, often with subsequent J segment fluorescent "run-off" analysis, is possible and has been widely exploited for functional repertoire analysis of mature lymphoid populations (Immunoscope).¹ DNA is preferred for clonality studies since it will also detect rearrangements in immature populations that may not be transcribed and is more representative of the genomic rather than the functional repertoire. RNA analysis, however, renders accessible complex loci such as TCR and TCRß. Detection of partial IgH D_H-J_H rearrangements requires the use of D_H consensus primers.²

A variety of detection methods for Ig/TCR PCR products exist, with, in general, the sensitivity being proportional to the complexity. Detection is classically performed on non-denaturing polyacrylamide gel electrophoresis (PAGE) with ethidium bromide staining, often with prior heating of PCR products to encourage heteroduplex formation, with a sensitivity of approximately 10-¹ to 10^-2 (1-10%). This is increasingly being replaced by analysis of fluorescent PCR products on DNA fragment analysers (automated sequencers), allowing increased resolution and an approximately 1 log increase in sensitivity (10^-2 to 10^-3; 0.1-1%).³ It is also more user friendly than its radioactive predecessor (DNA fingerprinting).⁴ Family-specific amplification is more sensitive than consensus strategies since there is less competition for reagents from normal, polyclonal lymphocytes, but is less well adapted to initial diagnostic screening, unless incorporated in a multiplex format.⁵ Alternative strategies that allow separation of PCR products on the basis of nucleotide content as well as size, such as denaturing-gradient gel electrophoresis (DGGE),⁶ temperature-gradient gel electrophoresis (TGGE)⁷ or single-stranded conformational polymorphism (SSCP),⁸ have been described, particularly for analysis of TCR. The most sensitive and specific detection techniques involve sequencing of the clonal V-J junction and synthesis of an anti-junctional (AJO) or allele-specific oligonucleotide (ASO), which is then used a hybridization probe or clone specific primer.^9,10,11 The sensitivity of such strategies is approximately 10^-5, although this varies with the junctional sequence. While these strategies represent optimal specificity and sensitivity, they are labor intensive and do not permit detection of clonal evolution. It is therefore probable that they will be replaced by simpler strategies that provide an intermediate level of detection and that may be more easily applicable to large scale clinical follow-up. The arrival of real-time PCR quantification has already had a major impact on the detection and follow-up of oncogenic fusion transcripts. Adapting this type of technology to the detection of clonal V(D)J rearrangements is less simple, but possible.^12,13

IgH represents the most useful gene target for detecting B cell clonality since it rearranges early during B lymphoid development and demonstrates extensive junctional diversity. Because of the marked combinatorial diversity it is necessary to use consensus or family-specific primers. Consensus V_H (FR1, FR2 and FR3) and J_H (FR4) primers have been described, and family-specific primers are used either individually or in a multiplex format. The main limitation of IgH PCR is false-negative results, the incidence of which varies with the IgH PCR strategy and the pathological subtype as a function of the degree of receptor editing (somatic mutation, intraclonal variability, etc.; see Section II and ¹⁴).

TCR represents the most useful marker for T cell clonality since it is rearranged at an early stage of both TCR and TCR ß lineage cells and is easily accessible by PCR. Due to the relatively limited combinatorial diversity it is possible to detect all functional V-J rearrangements with a small panel of V- and J-specific primers, commonly in 1 to 3 multiplex reactions.⁵ False negative results due to failure to recognise the V and J segments involved are very rare if all V and J combinations are included. Junctional diversity is, however, limited, with an approximate CDR3 length of 20 nucleotides compared to 60 nucleotides for IgH. A further difference in the gaussian distribution of polyclonal TCR versus IgH PCR products results from the fact that TCR rearrangements only undergo selection of in frame rearrangements in TCR -expressing cells, which represent a minority of circulating lymphocytes, and the PCR peaks generated therefore differ by only 1 bp (compared to 3 bp for IgH). A further problem in TCR PCR is the presence of normal conserved rearrangements with little junctional diversity ("canonical" TCR rearrangements that do not demonstrate N nucleotide additions) in DNA from circulating TCR -expressing cells, which can mistakenly be interpreted as clonal bands in polyclonal populations.¹⁵ For all these reasons, the main limitation of detection of TCR rearrangements by PCR is the risk of false-positive results. This appears to be particularly problematic in the analysis of peripheral blood samples from older individuals. It is therefore extremely important to analyze TCR PCR products using either high resolution electrophoretic techniques or electrophoretic systems that allow separation on the basis of nucleotide content as well as size, and to be aware of the size and nature of canonical rearrangements, particularly those involving V9-JP (also known as V2-J1.2).¹⁵ It is not possible to distinguish polyclonal and clonal TCR rearrangements on classical agarose gel electrophoresis.

Clinical Applications at Diagnosis
Analysis of lymphoid clonality is frequently useful in the distinction of malignant vs. reactive lymphadenopathy in cases for which morphological and immunological analysis is inconclusive. Detection of apparently clonal proliferations in clinically benign conditions such as large granular lymphocytosis, lymphomatoid papulosis or Sjögren's syndrome illustrate, however, that clonality is not synonymous with malignancy,^{16, 17} and it is increasingly clear that dys/autoimmune disorders represent one end of the spectrum of lymphoproliferative disorders. Interpretation of molecular results, particularly for TCR analyses (see above), should therefore be undertaken in close collaboration with histopathologists and should take into account clinical, immunological and (cyto)genetic features.

"Illegitimate" rearrangements, by which TCR rearrangements occur in B lineage cases and vice versa, are frequent in ALL, probably reflecting the fact that the blasts are arrested at a "recombinase competent" stage of maturation arrest. In contrast, such rearrangements are rare in NHL. This may be due to the fact that oncogenic conversion occurs in a recombinase-incompetent cell or, more likely, that the chromosomal configuration of illegitimate targets are not in an accessible configuration. For these reasons, analysis of Ig and TCR genes can aid determination of T- or B-cell lineage, although this distinction is predominantly based on immunophenotyping. Co-existent clonal Ig and TCR rearrangements are, however, recognized in disorders such as angioimmunoblastic lymphadenopathy with dysplasia (AILD).^{18, 19} It is probable that they originate from different cell populations, as a consequence of the immune deregulation.

The incidence of informativity of IgH PCR in B-cell tumors varies with the stage of maturation, mainly as a function of their pre- or post-germinal center phenotype.^14,20 It is lowest for follicular lymphoma (FL) and HIV-associated NHL and highest in CD5-expressing proliferations such as mantle cell lymphoma (MCL).^21,22,23 For the reasons stated above, molecular informativity is rarely a problem with TCR PCR analysis. Since histological distinction of clonal and reactive T lymphocytes is often problematic, particularly in extranodal tissues such as skin or gut, the latter represent an increasingly common source of material transmitted to laboratories performing lymphoid clonality analyses.^24,25

Molecular Genetic Abnormalities
Technical aspects
Only general aspects pertaining to diagnostic use of molecular genetic markers will be discussed here. Experience with their practical use is largely restricted to markers that are sufficiently common, indicative of a particular diagnostic or therapeutic subgroup or useful for disease follow-up.

In molecular terms, chromosomal abnormalities or their sub-microscopic equivalents are of two general types: those in which the breakpoints occur within the involved genes, leading to the production of a fusion RNA transcript and a chimeric protein (Section III), and those that represent Ig/TCR rearrangement errors (Section II). The former lead to a qualitative change, insofar as the fusion protein is more or less tumor specific. They are relatively rare in B lineage NHL, compared to their acute leukemic counterparts, and Ig errors predominate. It is not clear whether these differences are due to a bias of identification or a real difference in oncogenic mechanisms. Qualitative abnormalities are usually detected by reverse transcriptase (RT) PCR from RNA, although DNA-based detection is possible for some cases. The best characterized example in NHL is the NPM-ALK fusion transcript,^26,27 but the recently described AP12-MLT/MALT1 also falls into this category (see Section III).

Ig gene-associated translocations in B-NHL predominantly involve the IgH chain locus at 14q32 and, more rarely, Ig and Ig light chain loci at chromosomes 2p12 and 22q11, respectively. Errors of the recombination process lead to juxtapositioning of a proto-oncogene, with its consequent transcriptional deregulation and a quantitative change in its expression. The breakpoints occur outside the coding sequences, often at a distance of several kbp from the deregulated proto-oncogene. Genomic clustering facilitates detection of the translocation breakpoints by PCR from DNA, at least for those involving D-J breakpoints. Since the proto-oncogene-(D)-J junctional sequence demonstrates features identical to those seen in normal V(D)J rearrangement, these junctional sequences generate variably sized PCR products that can be used to distinguish similar translocations in different individuals and to identify contaminants.

Comparison of cytogenetic, molecular genetic and immunological analysis
Only karyotype analysis by classical morphological analysis of metaphase spreads or by multifluorescent SKY or M-FISH techniques^28,29 provides an overall evaluation of the whole genome and directs further analysis. It is clear that the latter allow a substantial increase in resolution and capacity to identify abnormal chromosomes, with a consequent increase in the detection of abnormal karyotypes. In contrast, molecular techniques allow detection of specific abnormalities in situations where karyotype analysis is difficult, such as evaluation of archival tissue, when insufficient metaphases have been obtained, or for the detection of intrachromosomal abnormalities. Furthermore, it is increasingly recognized that molecular-positive, karyotype-negative cases are not uncommon, particularly if molecular cytogenetic techniques are not employed. They include cases with complex abnormalities, those with mitoses that are not representative of the tumor clone and those with sub-microscopic or "perfectly equilibrated" abnormalities. Notable examples of the latter category include TEL-AML1 [t(12;21)(p13;q22) translocation] in B cell lineage ALL^30,31 and FGFR3-IgH [t(4;14)(p16;q32) translocation] in myeloma.^32,33

Use of whole chromosome FISH paints is appropriate for detection of numerical abnormalities, although these rarely have a practical impact in lymphoma management. Locus or rearrangement-specific probes can be used to aid diagnosis, especially for the detection of heterogeneous abnormalities such as MYC-IgH or BCL6 (3q27)^34,35 rearrangements, which are difficult to exclude by Southern and even more difficult by PCR. The telomeric location of the IgH locus renders particularly difficult the detection of the partner chromosome, thus explaining the common 14q+ category of abnormalities. FISH analysis with IgH probes can help to characterize these cases.³⁶ Specificity of detection can be improved by the use of dual color FISH.³⁷ Probes that allow identification of abnormalities in interphase nuclei are preferable for screening purposes in order to allow analysis of patients with normal or failed karyotypes.³⁸ Such studies have an important role in early follow-up since they have the advantage of being quantifiable. The use of FISH for the detection of rare events is, however, limited by the background noise and by the number of nuclei that can realistically be examined. The increasing use of interphase FISH should be subjected to the same quality control as PCR analysis. In general, close collaboration between those undertaking PCR and FISH analyses is the best method of ensuring that the most appropriate technique is used in each clinical context.

Since Ig/TCR-associated translocations lead to abnormal expression of an essentially normal proto-oncogene, despite potential somatic mutations in the coding sequences, its detection by immunological techniques represents an alternative form of diagnostic analysis. This has the following advantages: it is independent of genetic heterogeneity; it represents the functional oncogenic endpoint and therefore potentially correlates more closely with prognosis; and it allows in situ immunohistological analysis. Its disadvantages include the inability to distinguish physiological from pathological expression and the absence of specific monoclonal antibodies that will work reliably on fixed material. Distinction of normal from pathological expressions of some oncogenes may not be a problem, if the wild type protein is not normally expressed in lymphoid tissue or is expressed with an easily identified architectural pattern in diagnostic material. For example, immunological detection of ALK expression in anaplastic lymphoma (see below) or, theoretically, BCL8 (Section III) in DLCL is useful since neither protein is expressed in normal hematopoietic tissues. Immunological detection of BCL2 is useful at diagnosis of follicular lymphoma, since the pattern of expression in the germinal centers allows distinction from reactive nodes. Use of BCL2 antibodies cannot, however, be used for staging or follow-up. In contrast, antibodies to BCL6 have little place in diagnostic assessment since this protein is expressed in normal germinal center B cells, in lymphomas with BCL6 rearrangements and/or mutations and in those with a germline configuration.³⁹ Similarly, detection of PAX5 protein expression is unlikely to have diagnostic significance, since it is normally expressed in the B lymphoid lineage (Section III).

Clinical applications at diagnosis
Oncogenic markers are likely to play an increasingly important role in the classification of NHL, but it is not yet clear which techniques(s) will be most appropriate for their detection in a routine diagnostic setting. Several abnormalities often co-exist, particularly at later stages of the disease. Since cytogenetic abnormalities are often complex and the molecular breakpoints heterogeneous, identification of a given genetic marker has diagnostic/prognostic significance but failure to do so does not.

For example, detection of BCL2-IgH [t(14;18)] in follicular lymphoma by karyotype analysis at diagnosis is approximately 75% for involved lymph nodes and 50-70% by PCR using MBR and MCR primers, the discrepancy being largely due to variant breakpoints.⁴⁰ Aberrant follicular expression of the BCL2 protein is found in approximately 90% of cases. Consequently, absence of BCL2 protein is therefore a more appropriate basis for calling into question a diagnosis of follicular NHL than absence of BCL2 DNA rearrangement. Further analysis of BCL2 negative cases will determine whether they represent functionally equivalent variants expressing, for example, other BCL2 family members or NHL with distinct pathogenic mechanisms.

Genomic breakpoints in MYC-IgH in Burkitt's lymphoma with t(8;14)(q24;q32) are much more heterogeneous, extending from over 300 kbp either side of the c-MYC coding region and frequently involving the IgH switch regions. They are therefore not readily detectable by PCR, and Southern blot analysis is only informative if positive. c-MYC-specific antibodies have limited practical application, and FISH analysis currently represents the most rapid and appropriate method for diagnosis.³⁴

The t(11;14) represents an intermediate situation, with approximately 40% of cases of MCL demonstrating breakpoints in the major translocation cluster (MTC) of the BCL1 region situated approximately 100 kb upstream to CCND1. As such they are easy to detect by PCR.⁴¹ The majority of cases in MCL, and virtually all those in B lymphoproliferative disorders other than MCL (E.A. Macintyre, unpublished observations and ⁴²), however, do not involve this. MTC and their breakpoints are best detected by FISH, preferably by strategies that also work on interphase cells.⁴³ For diagnostic purposes, it is clear that detection of CCND1 expression, the functional endpoint of these genetic abnormalities, is the most appropriate tool. This can either be performed by analysis of RNA expression^44,45 or by immunohistochemistry.⁴⁶ Interpretation of the diagnostic significance of such expression cannot, however, be undertaken in isolation, since CCND1 is seen in a wide variety of B lymphoproliferative disorders with different prognoses and therapeutic requirements, including MCL, multiple myeloma, hairy cell leukemia, prolymphocytic leukemia and splenic lymphoma with villous lymphocytes.^42,47,48 In practice, we restrict BCL1-IgH PCR to analysis of CCND1 RT-PCR-positive cases and refer CCND1-positive, BCL1-IgH-negative cases to further molecular cytogenetic analysis by fiber FISH⁴⁹ in an attempt to determine whether other genetic clusters exist and whether these vary with diagnostic subtype.

The NPM-ALK fusion transcript in anaplastic large cell lymphoma (ALCL) illustrates nicely the way in which diagnostic practice evolves rapidly. The demonstration that certain CD30+ ALCL are associated with a t(2;5)(p23;q35) that involves NPM and ALK allowed diagnosis by RT-PCR⁵⁰ and, in some cases, by DNA PCR.⁵¹ Limited availability of material suitable for RNA analysis restricted application of the former. With the development of ALK-specific monoclonal antibodies,⁵² it became clear that not only did the vast majority of NPM-ALK-positive cases aberrantly express ALK, but that ALK-positive, NPM-ALK-negative cases exist.⁵³ Furthermore, the localization (nuclear and/or cytoplasmic) of ALK allowed distinction of certain variants.⁵⁴ At the present time, ALK immunophenotypic analysis probably represents the most appropriate technique at diagnosis, with RT-PCR being reserved for confirmation, partner identification and initial characterization if follow-up is planned.

Prognostic evaluation requires assessment of a large number of genetic targets. As detailed in Section III, the number of recognized oncogenes involved in B cell lymphomas is increasing rapidly. Deregulation of distinct oncogenes, such as BCL2 and BCL6 in follicular lymphoma, can coexist at diagnosis and is even more common at relapse. At least two IgH switch translocations in myeloma have been shown to simultaneously deregulate expression of two independent oncogenes from both the chromosome 14 and partner derivative chromosomes by juxtaposing them to the E and Eµ IgH enhancers, respectively: FGFR3 and MMSET/WHSC1 in the t(4;14)(p16;q32)^55,56 and myeov and CCND1 in the t(11;14)(q13;q32).⁵⁷ It is probable that this will also be the case for NHL switch translocations. Prognostic assessment also depends on evaluation of the direct or indirect consequences of genetic abnormalities, which are unlikely to be the same for all patients in a given diagnostic category. For all these reasons, specific detection of individual abnormalities is increasingly unrealistic, and it is likely that global assessment of gene expression in lymphoma by microarray technology^58,59 will replace detection of specific abnormalities (see below).

Practical Aspects
Improvements in our understanding of the mechanisms and consequences of genetic abnormalities in NHL will only improve clinical management if appropriate pathological material is available for the majority of patients. Certain practical aspects of analysis of lymphoma biopsies render the generation of high-quality nucleic acid preparations more problematic than for their leukemic counterparts, particularly for RNA and high molecular weight DNA. Biopsies are performed, frequently at a distance from the pathology laboratory, by surgical personnel who are often less conscious than their hematologyoncology colleagues of the importance of appropriate conservation conditions. While it is possible to obtain PCR-grade DNA suitable for the amplification of small fragments (up to approximately 500 bp in length) from appropriately fixed material, only fresh or cryopreserved material is suitable for cytogenetic, RNA-based PCR and most Southern blot analyses. Furthermore, variations in fixation conditions (type of fixative, use of buffered solutions, duration of fixation) have a significant effect on amplification of even small DNA fragments.^60,61 Given the increasing number of (cyto)genetic analyses that can only be performed from fresh or cryopreserved material, it is our opinion that the priority for molecular diagnostic laboratories should be to encourage the establishment of surgical and laboratory procedures that allow conservation of unfixed material in all cases. This implies centralizing lymphoma biopsies in locations where fresh material can be rapidly transferred to a laboratory, where it can be sub-divided into fractions destined for cyto/histological, cytogenetic, immunological and molecular analysis, based on locally defined priorities for cases with limited material. Alternatively, the creation of specialized hemtological malignancy centers that regroup the aforementioned expertise will optimize rational use of material, in addition to allowing an intergrated approach to patient management.⁶² While it is obviously desirable to be able to analyze fixed material, attempts to adapt the technology to its systematic use will retard the evolution of the place of molecular analyses in the diagnosis and classification of lymphoid disorders.

The heterogeneous nature of lymphomatous infiltration can complicate interpretation of results. Whenever possible, it is preferable to conserve sections that correspond to the material undergoing molecular analysis for histological assessment in case of discrepant results. The interpretation of molecular analyses of tissues with minimal morphological involvement is always more reliable, and often only possible, when baseline analysis of pathological diagnostic material is available. This is particularly true when Ig/TCR clonal markers are used.

PCR-based techniques also suffer from their sensitivity. Increased recognition of this problem has led to a diminution in the incidence of contamination due to the establishment of appropriate laboratory structure and techniques and to the inclusion of control samples. Despite this, a recent international assessment of PCR detection of BCL2-IgH rearrangements showed a 28% false-positive rate among laboratories with records of publication in molecular diagnostics.⁶³ The risk of contamination obviously correlates with the effort expended to detect rare events, and all attempts to minimize the use of nested PCR and the manipulation of PCR products, such as with the use of quantitative real time PCR (RQ-PCR) analyzers, will decrease this risk. In general, if a single round PCR reaction is insufficient for detection of a given target at diagnosis in fresh or cryopreserved material in which the malignant population represents at least 10%, either it is present in a minor sub-population and may not be representative of the tumor in general or the PCR should be optimized. This should not represent an indication for nested PCR. The identification of very low levels of an increasing number of molecular oncogenic markers such as BCL2-IgH^64,65 in normal individuals also complicates interpretation of low level positivity. It has been our practice to limit nested analysis to follow-up of known molecular markers and to avoid making molecular diagnoses on uninvolved material destined more appropriately for staging/follow-up analysis. The increasing use of RQ-PCR is obviously reducing the indications for nested strategies.

Detection of contamination is more difficult for fusion transcripts since the PCR products from all individuals are identical, whereas V(D)J rearrangements are clone specific and consequently highly variable. Conversely, since the vast majority of contaminants are PCR products, repeated analysis of RNA in the absence of reverse transcriptase will permit their detection, whereas this is not possible for DNA PCR. Hybridization of transferred PCR products increases specificity but will not distinguish false from true positive results. An alternative strategy involves performing a confirmatory, "shifted" one-round PCR from RNA or DNA using a second pair of oligonucleotides, one of which is external to the oligonucleotides used for initial screening (and used only for this purpose in order to limit the risk of contamination).⁶⁶ If appropriately sized products are obtained in both PCR, the specificity is provided by four hybridizing sequences, rather than three with hybridization strategies, and potential contaminants are unlikely to be positive with both PCR.

Despite these limitations, molecular analytical techniques have several advantages, including rapidity, simplicity, and economy. Retrospective analysis of appropriately archived material is possible. The small quantities of DNA/RNA required for PCR analysis (0.1-1 µg) makes analysis of material such as endoscopic or skin biopsies and cerebrospinal fluid feasible.^24,25 With increasing experience and technological standardization, molecular diagnostics are likely to find an increasingly important place in lymphoma diagnostics. An ongoing BIOMED 2 European Concerted Action program, "PCR-based clonality studies for early diagnosis of lymphoproliferative disorders" coordinated by J.J.M. van Dongen, represents an attempt to assist this process.

Detection of Minimal Residual Disease
Space constraints prevent a detailed review of techniques for detection of minimal residual disease and only general comments will be made here. In the early days following the application of PCR analysis to detection of rare events in hematological malignancies it was thought that maximal sensitivity would be optimal, that it would allow analysis of the "black hole" represented by complete morphological remission and that it would translate into direct clinical benefit. With 15 years of experience, it is clear that not all PCR are equal and that the sensitivity required depends on the clinical situation. The role of molecular follow-up in individual patient management should also take into account that the majority of therapeutic decisions are taken at a relatively early stage and that identification of patients at high risk of relapse at this stage is of more potential benefit than the early prediction of relapse. Identification of high-risk patients is based on either the intrinsic characteristics of the tumor at diagnosis or the kinetics of tumor response to treatment. Molecular assessment of the latter has been shown to represent a powerful, independent prognostic marker in childhood ALL, based on Ig/TCR clonality analysis.^10,11

The situation is, however, more complicated in NHL. Heterogeneous involvement of blood or bone marrow in all but leukemic forms of NHL leads to sampling inaccuracies. Molecular genetic follow-up of fusion transcripts has the advantage of being comparable between patients as a result of target homogeneity. In contrast, the variable nature of Ig containing translocations renders some targets more readily amplifiable than others. This is also the case for lymphoid clonality. These strategies are obviously applicable only to marker-positive patients, but molecular follow-up can have significant clinical impact only if applicable to the majority of patients in a therapeutic protocol. At the moment, this is the case only for BCL2-IgH in FL, NPM-ALK in ALCL and IgH in somatically unmutated NHL, such as MCL. Pooling of results obtained with different targets was not possible with qualitative PCR strategies and was questionable with semi-quantitative, competitive strategies. It is to be hoped that RQ-PCR techniques will allow the generation of quantitative follow-up irrespective of the marker employed, but this remains to be demonstrated in NHL.

Evidence that molecular techniques can be used for practical patient management is largely based on BCL2-IgH and, to a lesser extent BCL1-IgH and lymphoid clonality.^{67,68,69,70,71}

Future Perspectives and Conclusions
The imminent achievement of the human genome sequencing projects and the arrival of microarray technologies⁵⁸ are revolutionizing biological evaluation of NHL. Gene expression profiling has been shown to allow stratification of DLBCL patients into those with a germinal center-like profile and a relatively good prognosis when compared with those demonstrating an activated B lymphocyte-like profile.⁵⁹ Our current collective priority is to find the best way to rapidly and efficiently evaluate the use of these techniques both to direct day to day patient management and to improve our understanding of lymphomagenesis. We are increasingly being faced with an excess of genetic data. The role of biologists interested in lymphoma management will be to order this information with respect to current biological practise. Cooperation at a national or international level is an inherent part of this process, as is the exploitation of high quality, well characterised tissue banks. This in turn raises economic, structural, educational, ethical and legal issues. National scientific and health care structures are responding to this "sea change" in biological practice in different ways.

The rapid increase in the incidence of NHL, in the rate at which novel genetic markers are being identified and in our understanding of the mechanisms leading to these abnormalities makes for an exciting period in biological lymphoma management. It is to be hoped that these developments will provide the basis for novel therapeutic strategies That will translate into improved clinical outcome.

II. Physiological Genomic Instability and Lymphoid Neoplasia

Dennis Willerford, M.D.^*

Chromosomal translocations involving antigen receptor genes are a central feature of many lymphoid neoplasms, including the majority of NHLs as well as some acute leukemias and plasma cell neoplasms.^1,2,3,4 These translocations characteristically juxtapose a cellular gene regulating cell growth, survival, or differentiation with transcriptional enhancer elements for antigen receptor genes, leading to deregulated oncogene expression. While other genetic changes are required for malignant transformation, in many cases the translocation is sufficient to initiate the tumor pathway, suggesting that it represents an early and essential step in lymphoid oncogenesis. Two components can therefore be considered in the pathogenesis of lymphoid malignancies: 1) generation of potentially oncogenic translocations; and 2) evolution of clinical malignancy as a result of growth or survival advantages and further genetic changes. The latter component is likely to be polymorphous, reflecting individual pathways for different oncogenes—corresponding in turn to distinct clinical entities.⁴ In contrast, the recurrent involvement of antigen receptor genes, particularly the immunoglobulin (Ig) heavy-chain locus, suggests a restricted set of mechanisms for creation of antigen receptor translocations. As a consequence, it could be that a large fraction of lymphoid malignancies is accounted for by a limited set of genetic, environmental and immunological circumstances contributing to translocation events.

Lymphoid cells utilize several unique forms of physiological genomic instability in order to create the vast diversity of Ig and T cell receptor (TCR) genes which recognize antigens. V(D)J recombination assembles Ig and TCR variable region gene segments during lymphoid development and during immune responses. In addition, immunoglobulin genes undergo isotype switch recombination and somatic hypermutation in antigen-stimulated B cells. The correlation between these unique physiological forms of genomic instability and the recurrent involvement of antigen receptor genes in oncogenic translocations in lymphoid malignancies suggests that lymphoid cancers may arise from errors in physiologic recombination events. Although direct evidence for this hypothesis and mechanistic insights remain limited, recent advances in three areas provide new opportunities for dissecting the underlying causes of lymphoid cancers: identification of genes involved in physiologic antigen receptor recombination and progress in understanding both the mechanisms of DNA cleavage and rejoining; improved understanding of how genomic instability is regulated in antigen-stimulated B cells; and molecular and genetic studies of human disorders and gene-deficient mice characterized by susceptibility to antigen receptor translocations and defective cellular responses to DNA damage.

Somatic Rearrangement and Mutation of Antigen Receptor Genes
V(D)J recombination
Antigen receptor diversity is generated by DNA rearrangement of germline variable (V), diversity (D) and joining (J) segments (or V and J segments) to form a contiguous variable region exon (reviewed in ^5,6). Combinatorial diversity created by selection among multiple germline segments, along with extreme sequence diversity created at the point of assembly, leads to a vast repertoire of antigen specificities and forms the basis for adaptive immunity.^4,5 V(D)J recombination begins with site-specific double-stranded DNA scission that is targeted by characteristic recognition signal sequences (RSS) flanking the borders of recombining variable region gene segments. DNA cleavage is coordinated between two rearranging segments and is catalyzed by the Rag-1/Rag-2 endonuclease complex, leaving a blunt RSS (signal) end and a sealed hairpin structure at the coding end.^7,8 Absence of Rag-1 or Rag-2 blocks lymphoid development beyond an early progenitor stage in mice and causes a subset of autosomally inherited severe combined immunodeficiency in humans.^4,9

The distinct structures of signal and coding ends created by Rag-mediated DNA cleavage are processed by different pathways.^4,6 Coding end processing injects a measure of randomness into the primary sequence at the resulting joint. The sealed hairpin structure is opened, often asymmetrically, and frequently loses bases to nuclease activity. Additionally, expression of terminal deoxynuclotidyl transferase (TdT) in developing lymphocytes adds bases (N-nucleotides) in a template-independent fashion to the free ends prior to ligation. The extreme sequence diversity at the site of V(D)J coding joints corresponds to the antigen contact residues of TCR and Ig molecules and contributes to the diversity of the immune repertoire. It is this junctional diversity that is exploited for PCR detection of lymphoid clonality (Section I). Rejoining of coding ends requires the DNA-dependent protein kinase (DNA-PK), which is mutated in the mouse Scid strain, along with components of the nonhomologous end joining (NHEJ) pathway for DNA repair. The latter includes the Ku70 and Ku86 nuclear antigens, along with XRCC4 and mammalian DNA ligase IV.⁴ In contrast, signal ends are usually religated without gain or loss of germline sequence, leading to an excision circle or a chromosomal inversion event, depending on the orientation of the recombining segments. Resolution of signal ends requires the NHEJ components, but is accurate in the absence of DNA-PKcs. A summary of the genes involved in V(D)J recombination is provided in Table 1. The V(D)J recombination mechanism is shared among seven rearranging antigen receptor genes in T and B cells. Inasmuch as these genes rearrange differently in each lineage and at specific stages, additional regulatory mechanisms are required to target the correct loci for rearrangement. This involves remodeling of chromatin structure, a process that is linked to transcription of the unrearranged antigen receptor locus, which invariably precedes rearrangement.¹⁰

A role for V(D)J recombinase in creating oncogenic chromosome translocations has been proposed, based on comparison of translocation breakpoint sequences with normal V(D)J joints (reviewed in ^1,4). Bearing among the strongest similarities to normal V(D)J joins are t(14;18) in follicular B cell lymphomas, where chromosome 14 breakpoints frequently occur at or near the RSS bordering DH or JH segments.^1,11 Sequences resembling RSS have also been reported in 18q21 and other partner chromosomes. However, in many instances these bear poor homology to physiologic RSS or are located too far away from the breakpoint to play a role in Rag-mediated DNA cleavage.^1,4 A novel translocation mechanism has been proposed based on the observation that the Rag proteins exhibit transposase activity in vitro; however, the extent to which this activity operates in vivo is unknown.^12,13 Taken together, the available breakpoint sequence data suggest that most antigen receptor chromosome translocations do not represent the products of a normal V(D)J reaction. Nevertheless, it remains likely that unresolved DNA breaks in antigen receptor loci created during V(DJ) recombination are a factor in the translocation event.

Class switch recombination
Isotype class switching occurs in B cells stimulated by antigen and is accomplished by a deletional recombination event, replacing the Igµ/Ig constant region gene cluster downstream of the functionally rearranged Ig variable region with a constant region encoding one of the Ig, Ig or Ig isotypes. This leads to antibody molecules with identical antigen specificity but with specialized effector functions determined by the incoming constant region (reviewed in¹⁴). Recombination is targeted by repetitive switch region sequences located 5' of Cµ and each of the constant region clusters, excepting C. Class switching is regulated by T cell signals including cytokines as well as CD40 ligand/CD40 binding, which act in part by regulating germline transcription from noncoding exons located 5' of the constant region clusters.^14,15 Germline transcripts in switch recombination appear to play a direct role in initiating DNA breaks, through interaction with switch regions to produce DNA/RNA hybrid structures, which attract nuclease activity.¹⁶ Rejoining of DNA ends in switch recombination occurs in a reciprocal fashion¹⁷ and requires DNA-PKcs as well as the downstream components of the NHEJ pathway.^18,19,20 Thus, V(D)J and class switch recombination have distinct mechanisms for creating double-strand DNA cleavage but appear to share a common pathway for resolution of these breaks. Ig class switch recombination is implicated in t(8;14) in some cases of sporadic Burkitt's lymphoma, where breakpoints occur within the switch regions upstream of Cµ, C, or C,^21,22 as well as in murine plasmacytomas, where c-myc is juxtaposed with the switch a region on mouse chromosome 12.²³ Recently, a novel oncogenic mechanism based on switch recombination has been identified in a multiple myeloma cell line. An excised switch region product containing C and the 3' IgH enhancer were found inserted into chromosome 11 in close proximity to the cyclin D1 proto-oncogene, associated with cyclin D1 overexpression.²⁴

Somatic hypermutation
During immune responses, germinal center B cells undergo further diversification of Ig specificities by the induction of somatic hypermutation in the V regions of IgH and IgL genes (reviewed in²⁵). Subsequent selection processes favor survival and expansion of cells with increased affinity for antigen. Mutations occur at a rate of approximately 1/1000 bp per generation, and are confined to a region within about 1.5 kb downstream of the promoter. Extensive studies using transgenic mice indicate that transcription and somatic hypermutation are intimately linked and likely involve DNA repair processes, although the precise mechanism is not understood^25,26). Somatic hypermutation most commonly involves base substitutions; however, recent studies have indicated that insertions or deletions may also occur with significant frequency, implying that the mechanism may involve DNA strand breaks.^27,28,29 There is evidence that some Ig locus translocations may arise during the hypermutation process. For example, in some t(8;14) in Burkitt's lymphomas, the translocation involves a V(D)J rearranged IgH allele, and there are several examples of breakpoints occurring within the V region, correlating with the physiologic target of somatic hypermutation.^27,30 Somatic hypermutation may also introduce genetic alterations outside of the Ig loci. When placed near Ig loci by a translocation, the c-myc and bcl-6 genes frequently develop point mutations characteristic of somatic hypermutation, potentially contributing to the additional genetic changes required for progression of the malignant phenotype.^31,32 Recently, it was shown that somatic hypermutation also affects the germline bcl-6 gene in normal germinal center B cells, indicating that this process is not completely restricted to the Ig loci under physiologic circumstances.^33,34

Genomic Instability in Germinal Center B Cells
Antigen receptor diversification in lymphocytes occurs not only during development in the primary lymphoid organs but also as part of immune responses. This is particularly the case with B cells, which undergo a second phase of development in germinal centers. Germinal centers are specialized lymphoid organs that arise from a limited number of antigen-activated B cells, which migrate to primary lymphoid follicles and interact with antigen-bearing follicular dendritic cells (FDC).^35,36 There is intense proliferation of the oligoclonal B cell population, where cycling centroblasts segregate from quiescent centrocytes in discrete dark and light zones, respectively. In the dark zone, proliferating B cells undergo somatic hypermutation. B cell progeny bearing mutated Ig genes may exit the cell cycle and migrate to the light zone, where they interact with FDC bearing antigen and antigen-specific T cells. These interactions promote the survival of cells with high-affinity antigen receptors, leading to the rapid evolution of humoral immunity toward high-specificity effector mechanisms. Interactions between centrocytes and T cells also facilitate Ig class switching.³⁷ Centrocytes selected by antigen may re-enter the dark zone and undergo further clonal expansion and somatic hypermutation, or may exit the germinal center to differentiate into memory B cells or plasma cells.

V(D)J recombination of antigen receptor genes has until recently been regarded as an ordered differentiation process. In recent years it has become increasingly apparent that productively rearranged antigen receptor genes may undergo additional rounds of V(D)J rearrangement, a process termed receptor revision.³⁸ In the Ig light chain genes, this may involve joining of previously unrearranged V and J elements, while in the IgH locus, V-region replacement may occur via a cryptic RSS element present near the 5' end of the coding regions in the majority of V segments. In immature IgM+ bone marrow B cells, receptor revision occurs in response to Ig receptor signals, leading to a change in specificity, which may assist in preventing the emergence of autoreactive Ig molecules.^38,39 Rag-1 and Rag-2 are also expressed in peripheral B cells; such cells are markedly increased after antigenic stimulation, suggesting that receptor revision occurs during immune responses.^40,41,42,43 The developmental stage of Rag-expressing peripheral B cells and the role of receptor revision in immunity are the subjects of intense interest at present. Somatic hypermutation and receptor revision by V(D)J recombination have also been identified in peripheral T cells; however, their role in regulating the peripheral T cell repertoire is not yet known.^25,44

Given the intense genomic instability that accompanies the germinal center reaction, it is plausible to hypothesize that some or all of the critical genetic alterations leading to malignant transformation may arise during immune responses. Nodal B cell NHL subtypes resemble normal components of the germinal center, and successive pathological classification schemes have built on improving knowledge of the presumed normal cellular counterparts.⁴⁵ This connection is supported by the finding that many of these tumors harbor somatic mutations of Ig V regions (reviewed in ^3,27,30), suggesting that the target of critical transforming events was the product of a germinal center reaction. Several clinical observations also support a link between immunity and lymphoma. Helicobacter pylori infection is associated with gastric MALT lymphomas.⁴⁶ The malignant proliferation often retains an antigen-driven component, such that remissions may be obtained after eradication of the organism with antibiotics. A subset of the AIDS-related lymphomas is not associated with profound immunodeficiency or EBV-mediated lymphoproliferation, and appears to correlate with chronic immune stimulation. Lymphoma Ig specificities for HIV proteins or autoantigens have been documented in some cases.^47,48 Lymphomas may also occur in association with chronic hepatitis C, sarcoidosis and autoimmune diseases. The common thread here would appear to be chronic immune stimulation, perhaps in the setting of disordered immune responses.

Genetic Susceptibility to Antigen Receptor Locus Translocations
Given the vigor with which Ig genes are modified both during B cell development and immune responses, it is perhaps a wonder that lymphoid malignancies are not more common. It seems likely that tumor suppressor mechanisms are active in lymphocytes to ensure that physiologic DNA breaks are not rejoined in an inappropriate fashion. Clues to these functions may be obtained through study of humans and mice with genetic alterations wherein the frequency of oncogenic antigen receptor translocations is increased (Table 2).

Human genetic syndromes
Among inherited syndromes in humans, lymphoid malignancies are increased in two types of disorders: immune defects that impair host responses to EBV, and a subset of autosomal recessive disorders affecting DNA damage responses. The latter includes ataxia-telangiectasia (A-T), Nijmegen breakage syndrome (NBS), and Bloom's syndrome. A-T is characterized by progressive cerebellar ataxia, telangiectasias on sun-exposed surfaces, hypersensitivity to ionizing radiation, and cellular and humoral immunodeficiency (reviewed in ⁴⁹). The gene mutated in A-T, ATM, encodes a member of the PI-3 kinase gene family that includes the DNA-PKcs.⁵⁰ A-T fibroblasts exhibit general chromosomal instability, and nonmalignant translocations involving TCR genes are frequently found in blood lymphocytes.^51,52 Cancer occurs in approximately 38% of A-T patients,^49,52,53 of which 85% are lymphoid leukemias and lymphomas. The majority of these are of T cell origin and exhibit recurrent TCR locus translocations, often involving the TCL1 oncogene on chromosome 14q32.⁵⁴ NBS shares several features in common with A-T, including hypersensitivity to radiation, the characteristic chromosomal rearrangements in lymphocytes involving antigen receptor loci, and the predisposition toward malignancies, particularly those of lymphoid origin (reviewed in ⁵⁵). In contrast to the spectrum of malignancy in A-T, the most common lymphoid tumors in NBS patients are NHL of B cell origin. The NBS1 gene encodes nibrin, which forms a multi-protein complex that includes Rad50 and Mre11.^56,57 Following exposure to radiation this complex forms nuclear foci, a function that is defective in nibrin-deficient cells.⁵⁶ Bloom's syndrome is an autosomal recessive disorder characterized by defective growth, defects in humoral immunity, and an extraordinary incidence of cancer. Bloom's syndrome cells have a mutator phenotype and a distinct pattern of chromosomal instability, including recombinational exchanges between homologous chromosomes.^4,58 Mutations are found in the BLM gene, which is homologous to ReqQ DNA helicases in bacteria. Cancer afflicts the majority of Bloom's syndrome patients, usually at a young age; the most common tumors are lymphomas and lymphoid leukemias.

At present, there is only a superficial understanding of how the ATM, nibrin and BLM proteins function in cells. Both A-T and NBS patients and respective cell lines exhibit heightened sensitivity to agents inducing double-strand DNA breaks, such as ionizing radiation or bleomycin, as well as abnormal cell-cycle regulation in response to DNA damage. Neither protein appears to be directly involved in repairing DNA damage.⁵⁹ Similarly, a direct role for ATM, nibrin or BLM in V(D)J recombination has been discounted based on normal rearrangement of plasmid recombination substrates in patient fibroblasts,⁶⁰ (C. Yeo, P. Concannon, and D.M. Willerford, manuscript in preparation). ATM has protein serine-threonine kinase activity, and a number of substrates have been identified in vitro and in vivo, many of which tend to play roles in either cell-cycle checkpoint control or apoptotic pathways.^61,62,63 In particular, ATM phosphorylates p53 in response to DNA breaks, and accumulation of p53 in response to DNA damage is delayed or absent in A-T cells. Recently, it has been shown that ATM also phosphorylates nibrin.^64,65 The emerging picture of ATM function is that it plays a signaling role, receiving information from molecules that detect DNA damage and relaying that information to appropriate downstream effector molecules that trigger responses ranging from cell cycle arrest to apoptosis. In this context, ATM and its downstream signaling partners could respond to unresolved DNA breaks during V(D)J recombination, either suppressing indiscriminate repair pathways, or perhaps limiting the survival of cells at increased risk for abnormal rejoining events.

Susceptibility to lymphoma in gene-deficient mice
Lymphomas arising spontaneously in genetically mutant strains of mice represent potentially important models for understanding the pathogenesis of lymphomas in humans. Recently, mouse strains exhibiting susceptibility to translocation-associated lymphoid malignancies have been identified, providing potential clues to the translocation process. In mice with combined Scid and p53-null mutations there is a high incidence of B lineage tumors occurring at an early age, compared with tumors exhibiting longer latency and predominantly thymic origin in the parental p53-/- strain.^{66,67,68,69,70} The majority of these tumors exhibit translocations involving the IgH locus on chromosome 12. Translocation partners include regions near c-myc on chromosome 15 and n-myc on chromosome 12, frequently accompanied by amplification of the translocated segment and associated with high expression of myc proteins.⁷⁰ The tumors appear to arise at the pro-B cell stage, coincident with physiologic IgH rearrangement, suggesting that the translocations arise during attempted IgH locus rearrangement in Scid pro-B cells. This conclusion is supported by the finding that a Rag-2-null mutation blocked the development of t(12;15) pro-B cell lymphomas when introduced into the Scid p53-/- strain, demonstrating that initiation of V(D)J recombination was a required element in the oncogenic pathway.⁷⁰ These studies suggest that physiologic suppression of oncogenic DNA rearrangements during V(D)J recombination has two important elements: efficient rejoining of DNA ends created by Rag-mediated DNA scission, and an intact cellular response to DNA damage. This model is supported by recent observations showing that mutations in other components in the V(D)J rejoining pathway, including Ku86 and XRCC4, lead to a similar phenotype of pro-B cell tumors bearing t(12;15) when bred onto the p53-deficient background.^71,72,73 In the absence of effective DNA damage responses, accumulation of unresolved V(D)J ends may permit repair by alternative DNA mechanisms that are indiscriminate with regard to rejoining.

Conclusions
By pointing to oncogenes, chromosome translocations involving antigen receptor loci have provided critical insights into the process of cellular transformation in lymphoid malignancies. By comparison, little is known regarding how and why these translocation events arise. Direct evidence is accumulating that physiologic genomic instability aimed at diversifying antigen receptor genes predisposes lymphoid cells to this lineage-specific oncogenic mechanism, and it may soon be possible to achieve a general understanding of the causes of these tumors. This will require a more detailed understanding of the mechanisms of DNA breakage and rejoining in lymphoid cells, as well as of the genetic safety measures which suppress oncogenic translocations. Disruption of these mechanisms, whether by genetic variation, environmental exposures, or disordered immunity, may increase the likelihood of lymphomagenic translocations. The means to test these hypotheses directly is emerging, and may lead to strategies to manage individual and environmental risks.

III. Oncogene Deregulation in B Cell Non-Hodgkin's Lymphoma: Recent Advances

Stephan W. Morris, M.D.^*

Significant progress has been achieved over the past 20 years in the molecular characterization of various genetic abnormalities of pathogenic importance in NHL, spurred by the realization that characteristic chromosomal translocations, or rearrangements, tend to occur specifically in a given type of NHL, consistently altering the regulation of a particular gene. In the sections that follow, oncogenes involved by several of these recurrent chromosomal rearrangements in NHL are described. Due to space limitations, emphasis is placed on the more recently identified lymphomagenic genes; the interested reader will find a number of excellent reviews describing those genes identified in earlier studies to be important in the development of NHL (Table 3), such as the G₁ cell cycle control gene cyclin D1 (CCND1, BCL1), the antiapopotosis gene BCL2, the bHLH/leucine zipper transcription factor cMYC, the IkappaB(BCL3)/REL/NFkappaB transcriptional pathway genes, and the POZ/zinc finger transcription factor gene BCL6 that are altered predominately in mantle cell, follicular, small noncleaved cell (Burkitt's), extranodal diffuse large-cell, or nodal diffuse large-cell NHL, respectively.^1,2,3,4,5 Likewise, the role of aberrant anaplastic lymphoma kinase (ALK) gene expression in the genesis of NHL ("ALKomas") will not be detailed here, given that B cell NHLs appear to be only rarely ALK-positive (almost all ALK lymphomas being T-lineage or null).^6,7 The major characteristics of oncogenes deregulated in B cell NHL are summarized in Table 3.

PAX5
Approximately 50% of small lymphocytic lymphomas with plasmacytoid differentiation, so-called lymphoplasmacytoid lymphoma (LPL), contain a t(9;14)(p13;q32). The t(9;14) illustrates the involvement in NHL of the PAX (for paired homeobox) transcription factor family that normally controls embryonic development and organogenesis.^8,9 Members of this family contain two discrete DNA-binding domains—the paired box and the paired-type homeodomain—that display coordinate DNA binding specificity. Lymphomas containing t(9;14) possess a plasma cell-like phenotype with serum paraprotein production, and follow an indolent course followed by large-cell transformation. The t(9;14) results in juxtaposition of the PAX5 gene with the Ig heavy-chain gene on chromosome 14.¹⁰ PAX5 is normally expressed in fetal brain and liver during development but becomes restricted to B cells and testis after birth.¹¹ PAX5 is transcribed throughout B cell ontogeny, but undergoes downregulation during plasma cell differentiation. Knockout mice experiments have demonstrated that Pax5 is important for midbrain development and that loss of Pax5 function results in maturation arrest of lymphocytes at the pro-B cell stage.¹² Pax5 overexpression, by contrast, results in splenic B cell proliferation.¹³ Several genes important for B cell development have been proposed to be PAX5 targets including CD19, B cell receptor component Ig alpha (mb-1), transcription factors N-MYC and LEF-1 (positively regulated by PAX5), and the cell surface protein PD-1 and the p53 tumor suppressor (which are downregulated).¹⁴ Also, B cell SRC-family tyrosine kinase BLK, which transforms lymphoid progenitors in an activated form but is dispensable for B cell development and activation, is upregulated by PAX5.^15,16,17 Further, PAX5 appears to be required for normal IgH VDJ recombination, given that V-to-DJ recombination is reduced 50-fold in Pax5-deficient pre-B cells.¹⁴

Dysregulation of PAX5 transcription by a translocated IgH promoter has also been described rarely in NHL subtypes other than LPL and in myeloma.¹⁸ Interestingly, other PAX genes play a role in oncogenesis; for example, PAX3 and PAX7 fuse to the forkhead domain transcription factor (FKHR) in t(2;13) (q35;q14) and t(1;13)(p36;q14), respectively, in the skeletal muscle tumor rhabdomyosarcoma, and PAX5 itself is aberrantly expressed in some medulloblastomas and glioblastomas through an unknown mechanism.¹⁹

BCL8
Translocations affecting chromosome 15q11-q13 and various partners occur in about 4% of diffuse large-cell lymphoma (DLCL). Recently, Dyomin et al identified the chromosome 15 target gene, BCL8, which is expressed as a major transcript of 2.6 kb and a less prominent 4.5-kb message due to differential polyadenylation.²⁰ BCL8 expression is normally restricted primarily to the testis and prostate, with no transcripts found in hematopoietic tissues such as spleen, thymus, or blood leukocytes. RTPCR analysis of RNA from DLCL samples and lymphoma cell lines identified BCL8 expression in all cases having abnormalities of chromosome 15q11-q13, as well as in four of nine randomly selected DLCL cases and six of 15 DLCL cell lines (but none of three hyperplastic lymph nodes). The mechanism of aberrant BCL8 expression in lymphomas lacking 15q11-q13 rearrangements has not been defined.

The exact manner by which BCL8, which shares no homology with previously described genes, contributes to lymphomagenesis remains to be addressed. Likewise, the exact frequency of BCL8 abnormalities in NHL, whether aberrant BCL8 expression delineates a homogeneous clinical-pathologic entity, and the effect of BCL8 expression on the prognosis of NHL patients have yet to be investigated.

BCL9, MUC1, FcRIIB and Other 1q21-q22 Genes
Abnormalities of the chromosome 1 long arm, in particular 1q21-q22, occur in a variety of preneoplastic and neoplastic lesions, in both solid and hematopoietic malignancies.²¹ The third most common rearrangement site in B cell NHL is band 1q21, after 14q32 and 18q21, the locations of IgH and BCL2, respectively. Abnormalities of 1q21-q22 occur in B cell malignancies arrested at all differentiation stages, and no clear association with a particular NHL subtype has been reported. Aberrations of this chromosomal region are usually secondary and occur in about 10-15% of B cell NHL, being associated with a poor prognosis, especially in DLCL.²² One of these 1q aberrations, the duplication dup(1)(q21q32), is the second most common cytogenetic abnormality in Burkitt's lymphomas (after cMYC translocation), occurring in more than 30% of cases. Breakpoints in 1q21-q22 show surprising heterogeneity and involve several target genes; in addition, there exists considerable promiscuity in the partners with which these 1q regions rearrange. It is important to point out that a number of 1q21 abnormalities result in an unbalanced chromosome 1 translocation, suggesting that in addition to activation of specific target genes, haploid loss of the chromosome 1 long arm may also contribute to oncogenesis in some tumors. As described below, progress has been made recently in the characterization of various 1q21-q22 target genes.

In 1998, Willis and colleagues reported the cloning of a t(1;14)(q21;q32) in a pre-B acute lymphoblastic leukemia cell line (CEMO-1), identifying the novel BCL9 gene as the 1q21 target.²³ BCL9 encodes a 1,394-amino acid (aa) protein that contains several pentapeptide repeats, a potential nuclear localization signal, and a 30-aa region that shares 90% homology to a Drosophila embryo EST clone, but possesses no other recognizable domains. BCL9 is expressed ubiquitously as two low-level transcripts, a major 6.3-kb and a less prominent 4.2-kb message, while a 1.6-kb transcript is present only in spleen, thymus, and small intestine. The sequence of the 4.2- and 1.6-kb transcripts has not been reported.

The t(1;14) in CEMO-1 results in juxtaposition of IgH sequences adjacent to BCL9, the 1q21 break occurring in the 3' untranslated region (UTR) of BCL9. Transcript levels of BCL9 in CEMO-1 are 50-fold higher than in EBV-transformed normal B cells. Southern hybridization screening of a panel of 3' B cell malignancies containing 1q21 abnormalities showed only one additional case (an MCL) with a 30 UTR BCL9 breakpoint. FISH analysis using a BCL9-containing YAC identified one additional case (a follicular lymphoma) in which the gene was juxtaposed adjacent to the Igl locus from 22q11. The other 1q21 breaks in these 39 cases were heterogeneous, centromeric to BCL9 in half and telomeric in the remaining half. Thus, it was clear from this study that genes other than BCL9 are involved in many of the 1q21 rearrangements in B cell tumors.

Earlier this year, two reports described the characterization of a t(1;14)(q21;q32) in a large-cell lymphoma with an extranodal presentation (abdominal ascites), identifying dysregulation of the MUC1 gene, which is located 8 cM telomeric to BCL9.^24,25 This translocation was shown to link IgHG4 switch sequences to the portion of chromosome 1 immediately downstream of MUC1, leaving MUC1 intact, and MUC1 transcription and translation were shown to be dramatically upregulated. None of six other genes located in an 85-kb region immediately centromeric to MUC1 (CLK2, propin, COTE1, GBA, metaxin, or thrombospondin-3) were found to be overexpressed in this tumor. Southern blot analysis of 72 B cell NHLs of all histologic types containing a 1q21 rearrangement revealed MUC1 rearrangement in 4 (6%) of the tumors (one case each of diffuse small cleaved, follicular mixed, immunoblastic, and extranodal large-cell). In addition, increased copy number of the MUC1 locus was identified in 18 (10%) of 178 B cell NHLs of all histologies, with the amplification being low level (4-6 copies)^.24

MUC1 encodes a glycoprotein containing multiple copies of a tandemly repeated mucin-like domain. MUC1 mucin has been previously implicated in carcinogenesis. For example, adenocarcinomas with MUC1 mucin over-expression have high metastatic potential and poor prognosis; furthermore, Muc1null mice exhibit delayed progression of mammary tumors.²⁶ MUC1 expression is also associated with inhibition of T cell proliferation, resulting in an "anergic" state and inhibition of cell-cell and cell-matrix interactions that facilitate metastatic tumor expansion.²⁷ Expression of EMA (epithelial membrane antigen), which is equivalent to MUC1, occurs in essentially all ALCLs, and in lymphocyte-predominant Hodgkin's disease (75% of cases), plasmacytomas (75%), and T cell lymphomas (50%) due to unknown mechanisms other than 1q21 rearrangement. Thus, abnormal MUC1 expression may be involved in the progression of NHL, consistent with the involvement of 1q21 usually as a secondary abnormality in association with primary aberrations such as t(14;18)(BCL2) or t(8;14)(cMYC).

Yet additional 1q21-q22 gene loci undergo rearrangement in B cell NHL. Callanan et al recently reported analysis of a novel balanced t(1;22)(q22;q11) in three follicular lymphomas also containing t(14;18).²⁸ This characterization revealed FCGR2B, which encodes the immunoreceptor tyrosine-based inhibition motif (ITIM)-containing low affinity IgG Fc receptor FcRIIB, located at 1q22, as the target of this rearrangement. Three FcRII genes (FCGR2A,-B, and -C) and two FcRIIIB genes (FCGR3A and -B) are located in an 200-kb region in 1q22. The breakpoint characterized by Callanan et al interrupts the low affinity IgG Fc receptor locus just upstream of the FCGR3B promoter and 20 kb telomeric of the FCGR2B promoter, juxtaposing the 5' region of FCGR2B next to the Ig 3' enhancer. At least 30-fold overexpression of the two major 1.5- and 2.4-kb FCGR2B transcripts was found in a t(1;22)(q22;q11)-positive follicular lymphoma cell line compared to other B cell lines such as Raji Burkitt's lymphoma or EBV-transformed lymphoblastoid cells; by contrast, FCG3RB expression was absent. RT-PCR analysis of t(1;22)-containing cases revealed FCGR2A and -C (which encode immunoreceptor tyrosine-based activation motif [ITAM] Fc receptors) to be expressed at levels equivalent to EBV-transformed B cells and control lines such as myeloid U937 cells. Very high levels of the FcRIIB receptor alternative splice form FcRIIb2 were specifically expressed in t(1;22)-positive cases, whereas b1 isoform levels were not elevated above normal. Thus, t(1;22)(q22;q11) results in constitutive, high-level expression of an intact FcRIIB receptor (specifically, the b2 isoform).

FcRIIB is one of the ITIM-containing inhibitory coreceptors that effect negative regulation of immune responses mediated by ITAM receptors such as B cell antigen receptors (BCR).²⁹ For example, crosslinking of FcRIIB to a BCR expressed on the same cell by IgG/ antigen complexes downregulates BCR-induced activation signaling due to recruitment of the inositol phosphatase SHIP to the complex by interaction with the FcRIIB ITIM. It is presently unclear how high-level constitutive FcRIIb2 expression might contribute to B cell lymphomagenesis, but Fc receptors can definitely affect B cell growth; for example, activation of these receptors by antibody binding enhances the growth and differentiation of murine B lineage progenitors in vitro, and FcRII-deficient mice have a significantly increased B cell compartment.³⁰ In addition, overexpression of FcRIIB in nonlymphoid tumor cells can enhance tumorigenicity in vivo, perhaps by allowing the tumor cells to interfere with IgG activity, thereby diminishing antitumor antibody responses.³¹

Cloning of yet another 1q21 breakpoint has revealed two additional Fc receptor family genes to be involved in the pathogenesis of B-lineage malignancies. Hatzivassiliou et al reported in abstract form the cloning of a t(1;14)(q21;q32) in a plasmacytoma cell line, identifying two genes within 20 kb of one another in the region spanning the 1q21 break.³² One of these genes, provisionally named MUM-2 (multiple myeloma-2), is expressed in spleen and lymph nodes as three transcripts of 2.6, 2.8, and 3.4 kb generated by alternative polyadenylation signals. The MUM-2 ORF predicts a 515-aa cell surface glycoprotein containing four extracellular Ig-type domains, a transmembrane and a cytoplasmic domain, with 37% identity (51% homology) to FcRI over its first three extracellular domains. In the plasmacytoma from which it was cloned, the 1q21 breakpoint interrupts the MUM-2 coding region and juxtaposes it to IgH in the same transcriptional orientation, producing a fusion transcript between the first two MUM-2 exons and the transmembrane and cytoplasmic exons of C to encode a putative MUM2-C fusion protein. The second gene, MUM-3, is located 3' to MUM-2, expresses three transcripts of 3.0, 5.0, and 6.5 kb in lymph nodes and spleen, and encodes a protein with an extracellular portion containing six Ig-type domains homologous to members of the Fc and Ig-type adhesion receptor families. MUM-3 expression is elevated not only in myelomas but also in Burkitt's lymphoma cell lines containing 1q21 abnormalities. The exact role that these novel cell surface receptors play in normal lymphocyte development and function, and in B cell malignancies, awaits further study.

API2-MLT/MALT1
Marginal zone lymphomas have been separated into three distinct disorders by the REAL classification—primary nodal, primary splenic, and extranodal lymphoma of the mucosa-associated lymphoid tissue (MALT) type.³³ Although these disorders share a CD5- and CD10- negative B cell phenotype, they have subtle morphological differences and distinct clinical behaviors, suggesting that their pathogenesis may differ.^34,35 Marginal zone lymphomas of the MALT are the most common subtype of lymphoma arising in extranodal sites, and account for 5-10% of all NHLs.³⁶ These lymphomas have frequent multicentric extranodal involvement including the gastrointestinal tract, lung, thyroid, and mammary, salivary and lacrimal glands, and an indolent clinical course. Most MALT lymphomas originate in the setting of chronic inflammation triggered by infection or autoimmune disorders, including H. pylori gastritis, Sjögren's syndrome and Hashimoto's thyroiditis. Malignant B cell proliferation in gastric MALT tumors is dependent in part upon H. pylori-specific tumor-infiltrating T cells, and eradication of H. pylori by antibiotic treatment results in tumor regression in most patients.³⁷ The t(11;18)(q21;q21) is a recurrent abnormality in marginal zone lymphomas, having been reported in up to 50% of extranodal low-grade MALT tumors. In 1999, Dierlamm and coworkers showed t(11;18) to produce a fusion of API2 at 11q21 with MLT (for MALT lymphoma-associated translocation) at 18q21.³⁸ Simultaneously, two other groups also characterized the novel 18q21 gene, which both termed MALT1.^39,40

API2 belongs to a family of inhibitor of apoptosis proteins (IAPs) first identified in baculoviruses, in which they suppress host cell apoptotic responses to viral infection; five human IAPs have been identified—NIAP, API1 (also named cIAP1, HIAP2, MIHB), API2 (cIAP2, HIAP1, MIHC), XIAP-hILP, and survivin.⁴¹ IAPs contain a BIR (baculovirus inhibitor of apoptosis repeat) motif in one to three copies, a caspase recruitment domain (CARD), and a C-terminal zinc-binding RING finger domain (the latter being found in all IAPs except NIAP and survivin).^42,43 IAP-1 and -2 were originally identified as proteins recruited to the tumor necrosis factor receptor II (TNFRII) cytoplasmic domain via association with the TNFR-associated factors (TRAFs), TRAF-1 and -2. API2 is highly expressed in lymphoid cells in the spleen and thymus and suppresses apoptosis by binding to and inhibiting caspases-3 and -7, as well as the cytochrome-c mediated activation of caspase-9.⁴¹ Engineered mutant IAPs that contain only BIR domains can bind and inhibit caspa