Despite the extensive information elucidating various steps in lymphoid cell ontogeny and functional maturation, little is known of the stepwise genetic changes that lead to lymphomagenesis. Lymphocytes are distinctive in having directed gene recombination events as crucial elements of their normal development. It is also clear that a large number of errors in these recombination events have been documented in lymphomas, and the resulting chromosomal translocations have been implicated in the neoplastic transformation. Table 59.1-3 lists translocations associated with particular types of lymphoid malignancy. In some cases, however, the precise linkage between the chromosomal translocation and the transformation of the translocation-bearing cell remains hypothetical. For example, the t(14;18) translocation is characteristic of follicular lymphomas (see below). This translocation activates expression of bcl-2 in follicular center B cells that normally do not express it. However, the overexpression of bcl-2 in follicular center B cells does not lead to uncontrolled growth, 31 the translocation has been found in normal persons without lymphoma, 32 and some lymphomas that bear the characteristic translocation do not even express bcl-2 protein.33 Thus, if the translocation is involved in the neoplastic transformation at all, it is certainly not necessary to maintain the transformed phenotype in vivo. The translocation may be only one step of a multistep process that for most tumors is not understood.
Table 59.1-3: Chromosomal Translocations Associated with Lymphoid Malignancies |Favorite Table|Download (.pdf) Table 59.1-3: Chromosomal Translocations Associated with Lymphoid Malignancies
|Disease ||Translocation ||Genes Involved ||Mechanism ||Effects |
|Precursor B-cell ||t(9;22)(q34;q11.2) ||Abl 9q34; Bcr 22q11.2 ||Oncogene fusion ||Tyrosine kinase activation |
| ||t(v;11)(v;q23) ||Various partners; MLL 11q23 ||Oncogene fusion ||Drosophila trithorax transcription |
| ||t(12;21)(p13;q22) ||TEL 12p13; CBFA2 21q22 ||Oncogene fusion ||Ets-related transcription with core binding factor |
| ||t(1;19)(q23;p13.3) ||PBX 1q23; E2A 19p13.3 ||Oncogene fusion ||Homeodomain plus bHLH transcription factor |
|Precursor T-cell ||See Table 52·4 || || || |
|Follicular lymphoma ||t(14;18)(q32;q21) ||Igh 14q32; bcl2 18q21 Rare bcl 2 association with light-chain genes ||Transcriptional deregulation ||Prevention of apoptosis |
|Diffuse large B-cell ||t(3;v)(q27;v) ||bcl6 3q27 with various partners ||Transcriptional deregulation ||Interferes with transcriptional repression of bcl 6 |
|Burkitt lymphoma ||t(8;14)(q24;q32) ||Igh 14q32; Myc 8q24 Also associates with light-chain genes ||Transcriptional deregulation ||Promotes cell profliferation |
|MALT lymphoma ||t(11;18)(q21;q21) ||ApI1 11q21; MLT 18q21 ||Fusion protein ||Inhibits apoptosis |
| ||t(1;14)(p22;q32) ||bcl-10 1p22; Igh 14q32 ||Transcriptional deregulation ||Inhibits apoptosis |
|Mantle cell lymphoma ||t(11;14)(q13;q32) ||bcl-1 11q13; Igh 14q32 ||Transcriptional deregulation ||Overexpression of cyclin D1 |
|Lymphoplasmacytoid ||t(9;14)(p13;q32) ||Pax5 9p13; Igh 14q32 ||Transcriptional deregulation ||Expression of B cell transcription factor |
|Anaplastic large cell ||t(2;5)(p23;q35) ||ALK 2p23 NPM 5q35 ||Fusion protein ||Tyrosine kinase activation |
|T-prolymphocytic leukemia ||t(7;14)(q34;q32.1) ||TCRβ 7q34; TCL1 14q32.1 ||Transcriptional deregulation ||β Barrel protein transcription |
| ||t(14;14)(q11;q32.1) ||TCRα/δ 14q11; TCL1 14q32.1 ||Transcriptional deregulation ||β Barrel protein transcription |
| ||t(X;14)(q28;q11) ||MTCP1 Xq28; TCRα/δ 14q11 ||Transcriptional deregulation ||β Barrel protein transcription |
Chromosome translocations can result in altered cell biology in at least two major ways: (1) a novel chimeric gene product with new biologic activities can be created; or (2) a normally unexpressed or tightly regulated gene can be placed under the control of an expressed gene promoter (in lymphomas, usually the antigen receptor gene), resulting in abnormal or unregulated expression of the normally unexpressed or regulated gene. While distinctive chromosomal translocations are the dominant genetic abnormalities in most types of lymphoid malignancy, activating mutations in dominant oncogenes, inactivating mutations in tumor-suppressor genes, and viral gene expression also contribute to lymphoid neoplasia. The more common entities with defined genetic lesions are discussed.
One other central theme is prevalent throughout this discussion of the pathogenesis of the individual lymphoid malignancies—namely, activation of the NF-κB pathway. This plays a central role in diffuse large B cell lymphomas, MALT lymphoma, Hodgkin disease, multiple myeloma, and others (as discussed below). The regulation of NF-κB transcriptional activity, and the diverse genetic lesions that lead to its dysregulation, are the subject of an excellent review.187
Diffuse Large B Cell Lymphoma
Diffuse large B cell lymphoma (DLBCL) is the most common lymphoma, accounting for about 40 percent of cases in the United States, and is the lymphoma whose incidence is increasing at the fastest rate. This tumor is composed of diffuse infiltrates that are large neoplastic B cells, which can be diverse cytologically. Attempts to subclassify these heterogeneous tumors by microscopic features have been unsuccessful, and no constellation of immunophenotypic markers identifies distinct subsets. Patients present with a dominant rapidly growing mass, either in lymph nodes or in extranodal sites (40 percent of cases) such as the gastrointestinal tract, central nervous system (CNS), skin, testis, or bone. The cause is unknown. About half of patients are curable with combination chemotherapy. Clinical poor prognostic factors include age >60 years, high lactic acid dehydrogenase (LDH), poor performance status, advanced stage disease, and two or more extranodal sites of involvement.
The tumor cells express clonal immunoglobulin gene rearrangements and the V regions are mutated, suggesting that DLBCL is a tumor of germinal center or post–germinal center B cell origin. Recurring chromosomal translocations occur in approximately 50 percent of DLBCL, most often involving the bcl6, bcl2, and c-myc genes that are brought under transcriptional control of immunoglobulin regulatory elements.
About 30 to 35 percent of tumors contain a balanced translocation between chromosome 3q27 and one of the immunoglobulin genes, the heavy chain genes on 14q32, κ chain genes on 2p11, and λ chain genes on 22q11.34 Other partner genes have been identified on chromosomes 1q21, 2q23, 5q13, 5q31, 9p13, 11q13, 12p11, 12q11, and 12q23 in other tumors, and in some cases, a der(3) is identified without identification of the reciprocal partner. The gene on 3q27 involved in these cases is bcl6, which encodes a zinc finger transcriptional repressor important in the formation of germinal centers. Normally, bcl6 protein is detectable in germinal center B cells, but it is not normally expressed in pre– or post–germinal center cells.35 The translocations involving 3q27 usually truncate bcl6 in its 5′ flanking region or within the first exon or intron, and the coding sequence is usually intact; thus, the gene is translocated under the influence of a heterologous promoter, a mechanism called promoter substitution. In come cases, Bcl6 is upregulated through mutation of negative regulatory elements in the bcl6 gene that occur by somatic hypermutation mechanisms that normally target Ig variable gene segments.216 As a result of genetic alterations, Bcl6 is expressed constitutively and is unresponsive to signals that normally regulate differentiation of antigen-stimulated B cells into Ig-secreting plasma cells, including cytokines and the interactions with T cells and dendritic cells within the germinal center microenvironment. These processes lead to the inhibition of transcriptional repressive functions of Bcl6, allowing activation of the plasma cell transcriptional program under the control of Blimp-1 and XBP-1. Cell cycle arrest with loss of B cell phenotypic markers occurs, and the protein synthesis and processing machinery are dramatically increased. Germinal center B cells with dysregulated expression of Bcl6 suffer a differentiation arrest at a very dangerous point in their lifecycle, namely at a time of high proliferation rate that is accompanied by lineage-specific alterations of their primary DNA sequences that involve double-strand break intermediates (somatic hypermutation and class-switch recombination). This is thought to promote further chromosomal alterations; in addition, somatic hypermutation leads to the mistaken targeting of oncogenes, including c-myc, pim-1, and bcl6.216 Furthermore, in normal germinal center B cells, Bcl6 inhibits the expression of DNA damage checkpoint pathways (including p53 and ATR), thereby allowing Ig-recombination events to occur.220 However these pathways are constitutively repressed in DLBLC harboring a bcl6 translocation, promoting the survival of malignant clones that have accrued aberrant DNA damage. A second mechanism for inhibition of plasma cell differentiation in DLBCL appears to occur as the result of mutation and inactivation of both Blimp-1 alleles, which was detected in 8 of 34 cases (24%) of activated B cell–like DLBCL.217
Other more indolent B cell lymphomas can evolve to DLBCL. Nearly all follicular lymphomas at some time in their course ultimately acquire mutations in p53 and become more aggressive in pattern of growth and natural history.36 This progression from follicular lymphoma may explain why about 20 percent of patients with DLBCL have tumors containing the t(14;18) (see below). CLL converts to DLBCL in about 5 percent of cases. MALT lymphomas can undergo histologic progression to DLBCL. When DLBCL contains a translocation typical of another type of lymphoma, it often signals that the DLBCL evolved from additional genetic damage in a more indolent lymphoma.
The application of cDNA microarray technology to DLBCL has led to the definition of at least two subsets of disease with differences in their gene expression profiles. One group of patients (about 40 percent of the total) have tumors that express genes suggesting that the cells are of germinal center origin (referred to as GCB DLBCL), and the other group has tumors with a pattern of gene expression that resembles mitogen-activated peripheral blood B cells (called ABC DLBCL).37 Importantly, these molecular profiles seem to reflect distinct prognoses; the 5-year survival of patients with germinal center–like DLBCL was 75 percent compared with <25 percent for the activated B cell–like DLBCL group. Interestingly, constitutive activation of the NF-κB pathway appears to be required for survival of the ABC subgroup, 162 and although in most cases the mechanism of NF-κB activation is unclear, it has been demonstrated that 10% of ABC DLBCL contains mutations in the coiled-coil domain of the CARD11 (Carma-1) protein that is a key regulator of NF-κB activation by the B cell receptor.194 These mutations enhance the dimerization of CARD11, leading to constitutive NF-κB activation. ABC DLBCLs have mutated Ig genes but lack an ongoing somatic hypermutation process (suggesting a post–germinal center phenotype), whereas GCB DLBCLs appear to be arrested at the stage of B cell development with ongoing somatic mutation.138 A separate gene expression analysis of DLBCL defined three subsets as oxidative phosphorylation, B cell receptor/proliferation, and host response based on pathway analysis demonstrating increased expression of genes involved in mitochondrial function, cell cycle, or the host inflammatory response to the tumor, respectively.206 A heterogeneous group of tumors that do not fit either the GCB or ABC pattern constitutes 17 to 40 percent of DLBCL in different studies. 125,206 Thus, gene expression profiling has provided additional insights into the cell biology of lymphomas and may provide information that can be exploited clinically in their management.
Follicular lymphoma is the second most common lymphoma, accounting for about one-third of cases in the United States. It is characterized by a follicular pattern of growth of two morphologically distinct cell types: small cells with cleaved nuclear contour and large cells that are similar to DLBCL cells. In most cases, the fraction of the tumor consisting of large cells is 5 percent or less. As the fraction of large cells increases, the clinical aggressiveness of the tumor appears to increase and the natural history of follicular large cell lymphoma resembles DLBCL. The disease is predominantly lymph node–based, but a large fraction of patients have disease spread to the bone marrow. Few patients have involvement of other extranodal sites. The clinical prognostic factors that have been noted in DLBCL also apply to follicular lymphoma, but fewer patients have poor prognostic factors. Median survival for patients with follicular lymphoma is about 12 years. Many types of therapy cause temporary regression of the disease, but it often recurs. More aggressive therapy appears to be achieving longer remissions, but in contrast to the curable DLBCL, it is not yet clear that permanent remission can be achieved in follicular lymphoma. At autopsy, more than 90 percent of patients with follicular lymphoma have sites of disease that contain DLBCL.
The tumor cells contain functional, clonal immunoglobulin gene rearrangements and the V segments are heavily mutated. There is also evidence of ongoing V gene mutation and the accumulation of mutations in other genes as well.36 Thus, this tumor is thought to be derived from germinal center B cells.
Virtually all cases show cytogenetic abnormalities. The most common change (85 percent) is the t(14;18)(q32;q21) translocation that places the bcl2 gene on chromosome 18 under the influence of the immunoglobulin heavy chain gene promoter, usually joining to the JH region.38 Rare translocations juxtapose bcl2 with the κ or λ genes. The bcl2 protein is involved in the inhibition of programmed cell death, or apoptosis.39 Bcl2 is not normally expressed in the germinal center. Its overexpression might permit survival of a cell that was destined to die because it contains nonadaptive mutations that do not result in high-affinity binding to antigen being presented by the follicular dendritic cells.40 However, the protein does not promote growth or block maturation. The translocation apparently occurs early in B cell development at the time of the initial immunoglobulin gene rearrangements, yet it permits maturation to surface IgM- and IgD-expressing follicular center B cells.
In transgenic mice overexpressing bcl2, generalized polyclonal expansion of B cells is detected, but tumors do not occur unless another genetic lesion is introduced.41 Furthermore, some healthy individuals often have low frequencies of the translocation in their B cells but never develop lymphoma.226 Thus, although t(14;18) is important as an initiating genetic event, secondary genetic events are clearly required for lymphomagenesis. The identity of these additional genetic lesions is currently unknown, although recurrent chromosomal alterations include -1p32-36, -6q11-27, +7, +12, and +X. 180,204 Analysis using high-density single-nucleotide-polymorphism arrays has confirmed multiple chromosomal copy number alterations and loss of heterogosity, as well as common areas of uniparental disomy, including the mutated p53 locus on 17p. 171,225 Epigenetic inactivation of members of the Rb pathways (p15Ink4b, p16Ink4a, Rb1) has been noted in follicular lymphoma and may be important in transition to DLBCL.157
Bcl6 can be rearranged in 15 percent of cases or mutated in up to 40 percent of cases; trisomy 7 and trisomy 18 may be seen in 20 percent of cases. None of these appears to influence the natural history of disease. However, 6q23-26 lesions and 17p (p53) mutations alter the natural history and the prognosis.42 p16 (INK4a) inactivation by deletion, mutation, or methylation has been noted in histologic transformation to DLBCL.43
Gene expression profiling of lymph nodes from follicular lymphoma patients has identified two signatures that predict survival. These were termed immune-response-1 (>10-year survival) and immune-response-2 (3.9-year survival) signatures and reflected differences in host response to the tumor (i.e., in sorted CD19-negative cells) and not the follicular lymphoma B cells (CD19-positive). Immune-response-1 was characterized by T cells (enriched for FOX-P3 cells, as shown in later studies) and monocytes, while the immune-response-2 signature was characterized by genes expressed by macrophages and dendritic cells. 149,161
Extranodal marginal zone B cell lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma) occurs in extranodal sites and comprises small lymphocytes and monocytoid B cells that infiltrate the marginal zone of reactive B cell follicles and extend into the interfollicular region. In epithelial tissues, the infiltrate can replace the mucosa and extend deeper into the tissue. These lesions are usually (50 percent) gastric in origin, but lesions involving the thyroid (4 percent), salivary (14 percent), and lacrimal glands (12 percent), skin (11 percent), bronchial epithelium (10 percent), and small intestine (mainly in the Middle East) also occur. The underlying pathogenesis of the disease appears to be chronic infection or autoimmunity. In gastric lesions, H. pylori is detected in more than 90 percent of cases.44 Sjögren syndrome precedes salivary MALT lymphoma, 45 Hashimoto thyroiditis precedes thyroid MALT lymphoma, 46 Borrelia infection is associated with skin MALT lymphoma, 47 and intestinal parasites appear to predispose to small intestinal MALT [formerly called α heavy chain disease or immunoproliferative small intestinal disease (IPSID)].48 The tumors may show plasmacytic differentiation, but the presence of a serum paraprotein is unusual except in the case of IPSID. Patients present either with an extranodal mass of an involved organ or with stomach symptoms in the case of gastric MALT. The tumors are usually not disseminated, and a high proportion of patients are curable. The differential diagnosis is usually not difficult for an experienced hematopathologist. The tumors are monoclonal B cells that express CD20, CD79a, and CD21. They are distinguished from small lymphocytic lymphoma through the absence of CD5, from mantle cell lymphoma through the absence of cyclin D1 expression, and from follicular lymphoma through the absence of CD10 expression. The clonal immunoglobulin gene rearrangements in MALT lymphoma typically show mutated V regions, and the cell of origin is felt to be a post–germinal center marginal zone B cell.
Although early tumor growth is fueled by antigen and chronic inflammation, the tumor evolves such that in many cases it is not curable by eradicating the inflammatory stimulus. Most genetic data have been obtained in gastric MALT lymphoma in which CagA-positive strains of H. pylori induce a strong host inflammatory response that results in production of genotoxic reactive oxygen species. In about half the patients, the lesions regress completely when H. pylori is eradicated with antibiotics, and these tumors often do not have detectable genetic lesions. Of those tumors that do not regress, nearly all contain genetic lesions that upregulate expression of the MALT and Bcl10 oncogenes and constitutively activate the NF-κB signaling pathway. The t(11;18)(q21;q21) leads to a novel fusion protein (IAP2-MALT1) in which the amino-terminus of IAP2 (inhibitor of apoptosis 2, which binds and inactivates caspases 3, 7, and 9) is fused to the MALT1 protein. 165,195 The t(14:18)(q32;q21) and t(1;14)(p22;q32) bring the full-length MALT1 or Bcl10 genes under control of the heavy chain enhancer on chromosome 14.249 In response to external signals impinging on the lymphoid receptors, including the TCR, BCR, TLR, TNFR, Rank, CD30, or CD40 receptors, Bcl10 and MALT1 form a multimeric complex that includes TRAF6, which leads to the ubiquitinylation of the IKK-γ (NEMO) subunit of IKK and activation of the canonic pathway of NF-κB.200 The high-level expression of MALT or BCL10 by the IgH enhancer leads to activation of this pathway in a manner that is uncoupled from external, receptor-mediated stimuli. The IAP2-MALT1 protein appears to oligomerize and activate NEMO in a Bcl10- and TRAF6-dependent manner after recruitment of the complex to lipid rafts. 55,227 Thus, as with other B cell malignancies (e.g., large cell lymphoma and multiple myeloma), genetic lesions leading to constitutive activation of the NF-κB pathway are crucial to lymphomagenesis. A novel translocation has been identified, t(3;14)(p14;q32), in which the FOXP1 transcription factor is brought under control of the IgH enhancer.239 The mechanism of lymphomagenesis by this genetic rearrangement is not known.
Tumors bearing t(11;18) do not appear to progress to DLBCL, even though they are resistant to H. pylori eradication.52 By contrast, MALT lymphoma of the stomach or other organs can progress to DLBCL; when that occurs, the tumors have heterogeneous genetic lesions including amplification of rel or myc, deletions of 13q, and gains on chromosome 3, 7, 9, 12, and 18.53 Histologic progression is often associated with mutations in p53 and bcl6.54
Mantle cell lymphoma (MCL) is a diffusely growing neoplasm composed of small- to medium-sized lymphoid cells with slightly irregular nuclei. The infiltrates can appear vaguely nodular but they do not have the structure of follicles in follicular lymphoma. In most cases, nodal involvement is dominant but dissemination to bone marrow, peripheral blood, liver, spleen, and the gastrointestinal tract is common. Multiple lymphomatous polyposis can be an extreme manifestation of involvement of the gastrointestinal tract. The median survival is 3 to 5 years and the disease is relatively refractory to standard treatment approaches. Blastic and pleomorphic variants have been described with more rapid tumor growth and a considerably shorter survival (6 months).56 The malignant cell is a CD5-positive B cell with unmutated immunoglobulin genes. It is thought to be derived from mantle zone lymphocytes.
The typical genetic abnormality is t(11;14)(q13;q32), which activates the transcription of the cyclin D1 gene on chromosome 11. The overexpression of cyclin D1 (not normally expressed in the lymphoid lineage) is a reliable diagnostic marker for this disease, although it is lacking in <5% of cases; in some of these the cyclin D2 or D3 is translocated to Ig loci. Cyclin D1 normally regulates the activity of cyclin-dependent kinase 4/6, which promotes phosphorylation of the retinoblastoma protein. This leads to the release of E2F transcription factors and cell cycle progression from G1 into S phase. While it is clear that this translocation occurs as the result of aberrant VDJ recombination in early B cell precursors in the bone marrow, subsequent differentiation to a mature B cell occurs, and further oncogenic events are clearly required for lymphomagenesis. This is supported by studies suggesting that cyclin D1 is a weak oncogene in rat fibroblasts, 57 and that transgenic mice with cyclin D1 expressed from an Ig gene regulatory element did not develop lymphomas (but did cooperate with myc in double transgenic mice).58,199
Therefore, it is perhaps not surprising that despite the translocation upregulating cyclin D1 expression, MCL cells—especially those with an aggressive phenotype—exhibit a number of other genetic alterations that further abrogate Rb checkpoint functions. Some 5 to 10 percent of MCLs encode a cyclin D1 mRNA with a deletion of AUUUA sequences in the 3’ untranslated region, leading to further overexpression of cyclin D1 from this long-lived mRNA species.163 Homozygous deletion of p16Ink4a and overexpression of Bmi-1, a component of Polycomb repressor complex 1 that epigenetically silences the INK4A locus, are additional mechanisms that enhance CDK4/cyclin D activity. Some MCLs lack Rb expression due to intragenic deletions involving the Rb locus, suggesting that in these cases cyclin D1 may have other oncogenic targets. Additionally, high-level expression of CDK4/cyclin D1 complexes likely binds and titers p27KIP1 away from CDK2/cyclin E complexes, promoting the final phosphorylations of Rb required for S-phase entry.223
MCL has a high level of genomic instability, possibly related to homozygous inactivation of ATM in approximately 50% of cases. This, in conjunction with loss of the p53 pathway, may foster the high level of aneuploidy and tetraploidy observed in aggressive variants of MCL. 148,175,231
Burkitt lymphoma is a highly aggressive, rapidly growing malignancy with a growth fraction of nearly 100 percent. The histologic picture is a diffuse monotonous infiltrate of medium-sized cells. It exists in three clinical forms: endemic Burkitt lymphoma occurs in Africa, affecting children age 4 to 7 years; sporadic Burkitt lymphoma occurs throughout the world in children (accounting for about 50 percent of lymphoma in children) and young adults (adult median age, 30 years); and immunodeficiency-associated Burkitt lymphoma is associated with HIV infection and may be a presenting sign of AIDS or may occur years after the AIDS diagnosis.3 Extranodal sites are commonly involved: the jaw in endemic, the abdomen (masses often start in the ileocecal region) in sporadic, and the CNS in AIDS-associated Burkitt lymphoma. CNS spread is a risk for any of the forms. Sporadic Burkitt lymphoma can involve the lymph nodes or organs such as breasts, kidneys, and ovaries. Nodal presentations are more common in adults. When the bone marrow is involved, the malignant cells may circulate; Burkitt leukemia (FAB L3) is also known as B cell ALL and is the most aggressive form of acute lymphoid leukemia (see below).
Because of the high growth rate, tumors may reach massive proportions before diagnosis and treatment. The tumor cells are readily killed by chemotherapy, resulting in rapid tumor lysis syndrome, as the burden of nucleic acids places metabolic stress on the host (hyperkalemia, hyperuricemia, hyperphosphatemia, hypocalcemia, arrhythmia, renal failure).
The tumor cells express monoclonal surface IgM, CD19, CD20, CD79a, bcl6, and CD10 and are negative for CD5, CD23, terminal deoxynucleotide transferase, and bcl2. The immunoglobulin molecules contain mutations in the variable regions, and the tumor is thought to be derived from a germinal center B cell.
All cases have translocations that activate the expression of the myc oncogene on chromosome 8q24.59 In 80 percent of cases, the translocation is to the heavy chain gene complex on 14q32; in 15 percent of cases, the partner is the κ chain genes on chromosome 2; in 5 percent of cases, rearrangement involves the Λ chain genes on chromosome 22. The chromosome 8 breakpoints are 5′ of c-myc when they rearrange to chromosome 14 and 3′ when they partner with 2 and 22. Heterogeneity in breakpoints is also noted in different forms of Burkitt lymphoma. In endemic Burkitt, the chromosome 8 breakpoint is more than 1000 kilobases 5′ of c-myc and the chromosome 14 breakpoint is within the JH region. In sporadic Burkitt lymphoma, the chromosome 8 breakpoint is within 3 kilobases 5′ of c-myc and the chromosome 14 breakpoint is in the switch region.60 These translocations result in the continuous unregulated expression of c-myc by at least two mechanisms: promotion from the immunoglobulin gene enhancer, and disruption of the negative regulatory site that would normally control c-myc expression. The c-myc exon 1–intron 1 border, a site containing regulatory sequences for gene expression, is routinely mutated in translocated genes. Furthermore, mutations are also noted in c-myc exon 2 that interfere with the ability of the Rb-related regulatory protein p107 to suppress the transactivation domain of c-myc.61 Unlike so many other translocated genes in lymphoma, c-myc is an authentic transforming gene that induces telomerase expression, promotes growth, and is directly responsible for the neoplasia.
In addition to c-myc expression in all cases of Burkitt lymphoma, p53 is mutated and 6q is deleted in about 30 percent of both sporadic and endemic cases.62 Loss of p53 function can be viewed as a necessary second lesion since unregulated proliferation induced by translocated myc induced apoptosis in most primary cells. Interestingly, the Myc transactivation domain is often mutated in Burkitt lymphoma, with a loss of threonine at residue 58. Unlike Burkitt lymphoma expressing wild-type Myc, tumors with transactivation mutants had wild-type levels of ARF and p53 but failed to upregulate Bim, a proapoptotic Bcl-2 family member. Consistent with a loss of apoptosis induction by the mutants, expression of P58A Myc led to more rapid induction of lymphoma than wild-type Myc. These data led to the hypothesis that Myc-induced apoptosis requires a threshold of activation that is dependent on both ARF/p53 and Bim, and inactivation of either arm will abrogate the oncogene checkpoint.178
The role of Epstein-Barr virus in the disease is unclear. Nearly 100 percent of endemic Burkitt lymphoma and about 30 percent of sporadic cases contain clonal episomes of the Epstein-Barr virus genome. In endemic and AIDS-related Burkitt lymphoma, the virus is latent in the cells, and only EBNA-1 and the two small RNAs, EBER-1 and EBER-2, are expressed. The form of latency in sporadic Burkitt lymphoma is type II, in which latent membrane proteins (LMPs) are also expressed. It remains unclear how Epstein-Barr virus contributes to the lymphomagenesis in Burkitt lymphoma.63
As shown in Table 59.1-2, a number of entities are considered under the heading of mature T cell lymphomas. In general, lymphomas of mature T cell origin are not associated with signature chromosomal translocations. T cell lymphomas are more rare and have been less well studied. Like B cell lymphomas, they are identifiable as monoclonal through the expression of clonal T cell receptor gene rearrangements. Cell surface markers shared by mature peripheral T cells are often noted, such as CD2, CD3, CD5, and CD7. In many cases, one of these will be undetected, usually CD7, and this is a clue to the malignant nature of the infiltrate.64 The tumors often possess bizarre karyotypes with multiple abnormalities often involving the T cell receptor gene loci on chromosomes 7q35 (β chain genes) and 14q11 (α and δ chain genes). The chromosomes most often altered in structure are 1, 6, 2, 4, 11, 14, and 17. Breakpoints at 6q23 are common. Aneuploidy typically involves trisomies of 3 or 5 and an extra X chromosome.65 The diversity of these findings has not allowed the construction of a consistent model of molecular pathogenesis.
Two entities are exceptions to this conclusion: anaplastic large cell lymphoma (see below) and T cell prolymphocytic leukemia (T-PLL). T-PLL is rare, accounting for <2 percent of cases of small lymphocytic leukemia. The disease always involves the bone marrow and peripheral blood, and it frequently produces hepatosplenomegaly, adenopathy, and skin infiltrations. In 80 percent of patients, the tumor cells bear a t(14;14)(q11;q32.1), a reciprocal tandem translocation that joins the T cell receptor α genes to a region containing a pair of oncogenes called TCL1 and TCL1b. 66 Alternative ways of activating TCL1 include chromosomal inversion inv(14)(q11q32.1) and translocating it to the β chain genes on chromosome 7 [t(7;14)(q35;q32.1)]. The Tcl1 protein is an eight-stranded β-barrel protein. A structural homologue of Tcl1 called Mtcp1 maps to the X chromosome and is also involved in translocations to chromosome 14 in some cases of T-PLL [t(X;14)(q28;q11)].67 The function of β-barrel proteins is not defined, but expression of both gene products in transgenic mice leads to a T cell leukemia at the age of 15 months.68,69
Abnormalities of chromosome 8 [deletions, trisomy 8q, t(8;8)p11-12;q12)] are also seen in 70 to 80 percent of cases of T-PLL.70 Deletions and mutations at the ATM gene (mutated in ataxia telangiectasia) at 11q23 are also common, 71,72 implying a role for abnormal DNA repair in the pathogenesis of the disease.
Anaplastic Large-Cell Lymphoma
Anaplastic large-cell lymphoma (ALCL) is derived from peripheral T cells that are pleomorphic and large and have horseshoe-shaped nuclei. The majority of patients (70 percent) have advanced-stage disease with extensive peripheral and abdominal adenopathy, often with involvement of bone marrow and other extranodal disease sites such as skin, soft tissue, and lung. High fever is often seen. The tumor cells often express CD30, a member of the tumor necrosis factor receptor family that is a signal transduction molecule expressed on activated T cells. The major differential diagnosis is with Hodgkin disease, but Hodgkin disease often expresses CD15 and B cell markers that ALCL cells lack, and ALCL cells express EMA and granzyme B that Hodgkin cells lack.
Ninety percent of ALCLs show clonal rearrangement of the T cell receptor genes, and in the remainder both immunoglobulin and T cell receptor genes are in the germ-line configuration (null immunophenotype).73 The most frequent genetic abnormality (found in 70 to 80 percent) in ALCL is t(2;5)(p23;q35), a translocation in which the NPM gene for nucleophosmin on 5q35 is fused to ALK, a gene called anaplastic lymphoma kinase on 2p23.74 ALK encodes a tyrosine kinase receptor belonging to the insulin receptor family and is normally expressed during neural development, but not in lymphoid cells. Nucleophosmin is a nucleolar phosphoprotein involved in transporting preribosomal complexes from the nucleus to the cytoplasm, and in the regulation of cell division, DNA repair, transcription, and genomic stability. The fusion truncates the N-terminal nuclear localization signal of nucleophosmin, and it is thought that the fusion protein NPM-ALK (p80) localizes to the nucleus through dimerization with wild-type nucleophosmin. The ALK component of the fusion protein consists mainly of the intracellular catalytic domain of the kinase whose dimerization and activation are facilitated by NPM sequences. The activated kinase associates with Src and Jak3, leading to the activation of a number of downstream signaling pathways, including STAT3 and STAT5, Ras/MAPK, and PI3K/Akt. Retroviral introduction of a chimeric gene into mice causes T cell lymphomas, 75 and mice harboring NPM-ALK transgenes develop lymphoma.156
A number of other ALK-activating translocations have been detected in the 20 percent or so of ALCL patients lacking t(2;5). These include rearrangement to the tropomyosin 3 gene on chromosome 1, a TRK-fusion gene on chromosome 3, the ATIC (5-aminoimidazole-4-carboxamide-ribonucleotide transformylase-inosine monophosphate cyclohydrolase) gene on chromosome 2, the clathrin heavy chain gene on chromosome 17, and an unidentified partner on chromosome 19.76 ALK-negative ALCL is unusual and has not been carefully studied.
Mediastinal Large B Cell Lymphoma
Mediastinal or thymic large B cell lymphoma arises in the mediastinum and consists of sheets of large pleomorphic tumor cells with abundant pale cytoplasm associated with dense fibrotic strands that compartmentalize the tumor. The disease is often localized to the chest and affects young women disproportionately. The tumor cells express CD19 and CD20, and their immunoglobulin genes are clonally rearranged; however, no immunoglobulin molecules are expressed.77 The tumor cells are often hyperdiploid with gains in chromosome 9p and amplification of REL, 78 consistent with gene expression profiling demonstrating that mediastinal large B cell lymphoma has an activated NF-κB target gene signature that promotes cell survival.170 Furthermore, in addition to sharing clinical features with classic Hodgkin lymphoma, it also has a similar overall gene expression profile.230 In addition, the tumor cells appear to overexpress a gene called MAL that encodes a proteolipid normally found in association with glycosphingolipids in myelin and in T cells.79 It is unclear how MAL expression relates to the biology of the tumor.
Lymphoplasmacytic lymphoma produces the clinical syndrome known as Waldenström macroglobulinemia. It is a tumor of small lymphocytes, some with features of plasma cell differentiation, that affects the lymph nodes, bone marrow, and spleen. The median age of patients is 63 years, with men and women roughly equally affected. The clinical features are dominated by the IgM paraprotein secreted by the tumor. Symptoms are related to hyperviscosity; neuropathy, if the specificity of the secreted antibody is for a myelin sheath antigen; coagulopathy, if the antibody interferes with clotting factors, platelets, or fibrin; and cryoglobulinemia. A proportion of patients are infected with hepatitis C, but it is not clear whether the virus is promoting lymphoid cell growth or the immune system is responding to the viral infection or a particular viral protein. A disproportionate number of the tumors appear to use the VH 1-69 gene in the tumor immunoglobulin, and it has been shown that antibodies to hepatitis C E2 protein also frequently use this gene.80 The possibility that a viral antigen is driving the tumor is intriguing.
The immunoglobulin gene V regions contain somatic mutations indicating that the cell is of post–germinal center origin. However, the cell has not undergone class switching. Chromosome 6 deletions involving 6q21-q23 have been observed in approximately half of patients with Waldenström macroglobulinemia, and some data suggest that these patients have more aggressive disease.211,233 About 50 percent of cases were reported to have a translocation t(9;14)(p13;q32) bringing the PAX5 gene on chromosome 9 under the influence of the immunoglobulin heavy chain gene promoter.81 However, other investigators were not able to confirm this result using IgM–fluorescence in situ hybridization analysis.179 No other recurrent genetic lesions have been identified in this tumor.
Hodgkin disease is now divided into two disease entities: nodular lymphocyte-predominant Hodgkin disease (NLPHD) constitutes about 5 percent of all cases, and classic Hodgkin disease (CHD) constitutes the other 95 percent (with nodular sclerosis and mixed cellularity accounting for 90 percent). For many years, the fundamental nature of Hodgkin disease eluded study because in all subtypes of this disease, the malignant cell is a rare cell within a mass of normal polyclonal reactive inflammatory and immune cells. However, the ability to microdissect tumor tissue and harvest and analyze gene expression in individual tumor cells has markedly advanced our understanding.
NLPHD is a monoclonal B cell tumor characterized by a nodular pattern of growth and scattered expansion of neoplastic cells with nuclear contours that resemble popcorn kernels called L&H cells (for lymphocytic and histiocytic Reed-Sternberg cell variants). A network of follicular dendritic cells establishes the background, and these cells are associated with small B cells and numerous CD57+ T cells. The disease tends to cause localized adenopathy, usually in the neck, and it has a slow natural history. Local radiation therapy often cures the disease but in the small subset who relapse, subsequent therapy is usually effective at inducing subsequent remissions such that very few patients die from the disease.
L&H cells typically express CD20, CD79a, Bcl6, and CD45 and usually express J (joining) chain of immunoglobulin M and CD75. They are usually CD30- and CD15-negative, distinguishing them from Reed-Sternberg cells in CHD. The cells usually express both the Oct-2 and Bob-1 chains of the transcription factor that ensures immunoglobulin gene transcription. This, too, is distinct from Reed-Sternberg cells that are always missing one (20 percent) or both (80 percent) factors, which, along with the expression of Notch1, is likely an important cause for the repression of B lineage transcription.82 The immunoglobulin genes are clonally rearranged and carry a heavy load of somatic mutations with evidence of intraclonal variations that document ongoing mutation.83,84 The inference from these data is that the L&H cell is derived from the dark zone of the germinal center where the somatic mutation cell machinery is active. Progression to DLBCL is seen in about 3 percent of cases, and the lymphoma is clonally related to the NLPHD.85
CHD is characterized by malignant cells (Reed-Sternberg cells) that are multinucleated and reside in a cellular infiltrate composed of nonneoplastic lymphocytes, eosinophils, neutrophils, macrophages, plasma cells, fibroblasts, and fibrosis. The disease commonly originates in cervical lymph nodes and spreads contiguously to adjacent lymph node groups. CHD commonly involves the mediastinum, and about one-third of patients have spleen involvement, which implies hematogenous spread because the spleen has no afferent lymphatics. The marrow is involved in less than 5 percent of cases. About 90 percent of patients with CHD are curable with therapies widely available today.
Reed-Sternberg cells are nearly always CD30- and CD15-positive and are usually negative for CD45, the J chain, and CD75. CD20 is present on some tumor cells in about 40 percent of cases. The cells are readily distinguishable from CD30-positive ALCL because they express BSAP, a B cell–restricted protein, and do not express ALK. Reed-Sternberg cells express a wide range of cytokines and chemokines. In about 98 percent of cases, the Reed-Sternberg cells contain clonally rearranged immunoglobulin genes, and the genes contain somatic mutations, but no immunoglobulin transcription takes place and no intraclonal variation is noted; this suggests that mutations are not ongoing.86 A small number of cases appear to have clonal T cell receptor rearrangements.87
Genetic studies of Reed-Sternberg cells have documented hypertetraploidy consistent with the multinucleated nature of the cells, but no recurrent specific chromosomal changes have been documented in CHD.88 Comparative genomic hybridization demonstrates recurrent gains on chromosomes 2p, 9p, and 12q and amplifications on 4p16, 4q23-24, and 9p23-24. 89 Appropriate models to approach the problem of lymphomagenesis in Hodgkin disease are lacking. However, one reasonable inference from the presence of functional immunoglobulin gene rearrangements that do not lead to gene transcription is that there is a problem with the apoptosis pathway. Normal B cells that fail to express surface immunoglobulin die. During the process of somatic hypermutation in the germinal center, loss of immunoglobulin expression, or failure to increase affinity for antigen, leads to activation of programmed cell death pathways. A number of genetic alterations have been identified in Reed-Sternberg cells that help to explain why these cells survive despite the lack of immunoglobulin expression. These include constitutive activation of the NF-κB transcription factor leading to overexpression of c-FLIP and XIAP, proteins that inhibit the extrinsic and intrinsic pathways of apoptosis, respectively.166,190,201 Activation of NF-κB may occur through amplification of the RelA locus at chromosome 2p16 or mutation of the IkBβ repressor protein142,188 An important NF-κB target in Reed-Sternberg cells is Bmi-1, a member of the Polycomb repressor complex 1 (PRC-1) responsible for initiating the transcriptional silencing of developmentally regulated genes. Bmi-1 upregulation in CHD appears to play a role in silencing of the B cell differentiation program as well as silencing potential tumor suppressor genes including ATM, the product of the ataxia-telangiectasia DNA repair pathway protein.167 An additional survival and proliferative pathway activated in CHD is the JAK/STAT pathway, caused by overexpression of the JAK2 locus at 9p24, or mutation of the SOCS-1 (suppressor of cytokine signaling) repressor.186,237,248 These genetic alterations, combined with the cytokine-rich milieu provided by infiltrating inflammatory cells within the Reed-Sternberg microenvironment, lead to brisk activation of the JAK/STAT pathway. About 50 percent of patients with mixed cellularity CHD have clonal Epstein-Barr virus episomes detected, and the association is stronger in developing countries. When Epstein-Barr virus is present, the tumor cells appear to express LMP1, LMP2a, and EBNA-1 (see below). LMP2a is a transmembrane protein that harbors a cytoplasmic ITAM motif that can recruit cytoplasmic kinases that normally bind to the B cell receptor, thus mimicking signaling pathways required by developing B cells.143 LMP1 mimics an activated CD40 receptor, providing another mechanism for activation of NF-κB in a subset of Reed-Sternberg cells.192 Additional studies are necessary to clarify the relationship between Epstein-Barr virus and CHD.
Chronic Lymphoid Leukemia
CLL is the most common leukemia in the West. It accounts for 90 percent of chronic leukemias of lymphoid origin. Typically, the bone marrow contains infiltrates of small lymphoid cells and the peripheral lymphocyte count is >10,000/mm3. The morphology of the circulating cells is similar to that of normal small lymphocytes. When the peripheral lymphocyte count is not elevated but lymphadenopathy and/or splenomegaly are present, the patient is said to have small lymphocytic lymphoma. The main difference between CLL and small lymphocytic lymphoma is the pattern of disease involvement. The difference is most likely related to the expression of an adhesion molecule or a chemokine receptor that permits the neoplastic cells to home to lymph nodes.
CLL causes death by replacing the marrow, thereby inhibiting normal hematopoiesis. About 50 percent of patients with CLL die from infection. About 25 percent of patients with CLL develop autoimmune hemolytic anemia or thrombocytopenia that can mimic marrow failure, but when the cytopenias are due to an autoimmune mechanism, they are readily controlled with glucocorticoids and do not exert an adverse influence on survival. The autoantibody is usually not the tumor immunoglobulin.
CLL is not a homogeneous disease, despite the appearance of the neoplastic cells. About 50 percent of cases contain cells bearing mutated immunoglobulin, suggesting a follicular center origin, and about 50 percent of cases contain cells bearing unmutated immunoglobulin, suggesting a mantle zone origin.5,6 The presence of unmutated immunoglobulin genes predicts for a shorter natural history and more resistance to treatment. The cells in both types of CLL express CD5, low levels of surface IgM, and CD23. The expression of zeta-associated protein 70 (ZAP-70) correlates with unmutated VH sequences and may be an even better predictor of the time from diagnosis to disease progression requiring treatment.224 ZAP-70 is expressed on a subpopulation of normal splenic and tonsillar B cells that express an activated phenotype, 210 and although the activation of the BCR and downstream pathways in CLL is controversial, some data indicate that CLL B cells that express ZAP-70 are likely to respond to IgM cross-linking with increased ZAP-70-mediated tyrosine phosphorylation and calcium flux, leading to increased tumor cell proliferation and survival.151 Controversy surrounds the issue of the role of CD38. Without question CD38 expression has an adverse influence on survival in CLL.90 However, it is not completely clear whether CD38 expression predicts for unmutated immunoglobulin genes. One can usually distinguish CLL from MCL (another CD5-bearing malignancy) by the presence of cyclin D1 in MCL but not in CLL.
Trisomy 12 has been noted in about 20 to 25 percent of cases of CLL 91 and tends to identify cases with more aggressive natural history, but the correlation with unmutated immunoglobulin genes has not yet been assessed. About 50 percent of cases show deletions at 13q14;92,93 it is thought that such deletions involve tumor suppressor gene(s) that may be important in the pathogenesis of CLL (see below). In 5 percent of cases, CLL or small lymphocytic lymphoma may become DLBCL, a change that signals a more aggressive natural history and is usually heralded by mutations in p53.94 About 10 percent of CLL patients have tumors with p53 mutations and no progression to DLBCL. A small number of CLL patients have mutations in ATM. Deletions in 6q are seen in patients with somewhat more aggressive B cell prolymphocytic leukemia.95
Aberrant expression of microRNAs may play a pathogenic role in CLL. MicroRNAs are 19 to 25 nucleotides long and are excised from a hairpin structure of 60 to 110 nucleotides that is cleaved from a large primary transcript. Processing by the Drosha and Dicer RNAse complexes leads to the functional species that target specific mRNAs for degradation or inhibition of translation. CLL cells have a distinct pattern of microRNA expression from normal CD5+ B cells.146 A study of 94 CLL patients whose ZAP-70 and Ig mutational status, as well as treatment history, was known found that a signature set of 13 mature microRNAs could discriminate between patients with ZAP70+/unmutated and ZAP-70–/mutated phenotypes.147 A 9-member subset of these 13 could predict the time from diagnosis to the initiation of therapy. Furthermore, the tumor suppressor locus at the common deleted segment of 13q14.3 encodes microRNAs miR-15a and miR-16-1, whose normal function is to negatively regulate expression of Bcl-2, thus enhancing the normal apoptotic program of developing B cells. 145,152 Hence one pathway for CLL induction is the deletion of specific microRNA genes leading to Bcl-2 upregulation and aberrant survival of clonal CD5+ B cells.
ALL occurs mainly (75 percent) in children under age 6 years; 80 to 85 percent of ALLs are of precursor B cell origin and 15 to 20 percent are of precursor T cell origin. These precursor lymphoid cell neoplasms usually present with fatigue and weakness from anemia. In the vast majority of cases most of the bone marrow is replaced by the malignant cells with leukemia blasts present in the peripheral blood. A small fraction of patients will present with mass lesions and less than 25 percent marrow involvement; these cases are called lymphoblastic lymphoma. Lymphoblastic lymphoma is more commonly of T cell than B cell origin. T cell lymphoblastic lymphoma can produce a large mediastinal mass as it originates from the thymus. All the leukemias derived from precursor B or T cells are of the FAB L1 or L2 type morphologically. FAB L3 is Burkitt leukemia or B cell ALL, so called because it is derived from a mature peripheral surface immunoglobulin-expressing B cell. Acute leukemias of both lineages can spread to the CNS and require intensive chemotherapy for treatment together with maintenance therapy and CNS prophylaxis.
Despite clinical similarities, precursor B cell tumors differ from precursor T cell tumors in immunophenotype and genetic abnormalities. Precursor B cell tumors of the earliest stage are human leukocyte antigen (HLA)-DR-positive and express CD19, cytoplasmic CD22, and cytoplasmic CD79a. Intermediate-stage or common ALL is characterized by CD10 expression. In the most mature pre-B cell stage, blasts express cytoplasmic μ heavy chains. Genetic lesions commonly associated with precursor B cell neoplasms include the following: hyperdiploidy (51 to 65 chromosomes) is noted in about 20 to 25 percent of cases, and t(12;21)(p13;q22), the TEL/AML1(CBFA2) translocation, is associated with 16 to 29 percent of cases; both of these genetic lesions are associated with a good prognosis (85- to 90-percent long-term survival).96 Four other genetic lesions are associated with a poor prognosis, and each is noted in 3 to 6 percent of all cases: hypodiploidy; t(9;22)(q34; q11.2), the BCR/ABL translocation; t(4;11)(q21;q23), the AF4/MLL translocation; and t(1;19)(q23;p13.3), the PBX/E2A translocation.How these changes lead to leukemia and why they occur in pre-B cells is unclear. ALL with t(12;21) shows high levels of expression of CD10 and HLA-DR and no CD20.
The TEL gene is a promiscuous partner in leukemias. When translocated to the PDGFRb gene on chromosome 5, the disease that develops is chronic myelomonocytic leukemia.97 When translocated to the ABL gene on chromosome 9, the disease that develops is acute myeloid leukemia.98 When TEL translocates to the core binding-factor component gene CBFA2 (often called AML1), pre-B cell ALL develops. TEL (also known as ETV6) is member of the ETS family of transcription factors that contains a conserved 90-amino-acid winged helix-loop-helix DNA binding domain. AML1 is a component of core binding factor. It contains a runt domain (so called because of homology to Drosophila runt protein) that serves as its DNA binding site, recognizing an enhancer motif associated with a number of important genes in hematopoiesis.99 The runt domain also facilitates heterodimerization. The C-terminus of CBFA2 contains a transactivating domain that binds with p300, a transcriptional activator involved in acetylation of histone proteins and chromatin remodeling.100 Unlike other fusions with AML1, the fusion with TEL retains essentially the full-length AML1 including the runt domain and the transactivation domain. It has been proposed that the chimeric TEL/AML1 protein acts as a dominant negative to interfere with function of the remaining wild-type AML1.101 The wild-type TEL allele is nearly always deleted in cells bearing the translocation, suggesting that the abrogation of the normal functions of both TEL and AML1 is necessary for transformation.102 In murine models, introduction of Tel-AML1 into CD34+ hematopoietic precursors induces a preleukemic state characterized by increased self-renewal of B cell precursors; cooperation with other oncogenes is necessary for the full leukemic phenotype. Mutational analysis indicates that several domains from both TEL and AML1 are required for fusion protein activity. 208,209,244 The HLH domain of TEL likely converts AML1 from a transcriptional activator to a repressor through recruitment of N-Cor and associated histone deacetylases.169 TEL-AML1-mediated gene repression can be reversed by histone deacetylase inhibitors, suggesting a potential therapeutic strategy for B-ALL patients whose leukemic blasts harbor this genetic lesion.
The MLL gene on chromosome 11q23 encodes a histone methyl-transferase with homology to trithorax family proteins; it is required for normal hematopoiesis through its ability to maintain transcription of genes encoding homeobox transcription factors (Hox genes) that regulate multiple developmental programs.193 MLL is a promiscuous partner in acute leukemia, fusing with over 50 other genes and leading to AML, ALL, and mixed-lineage leukemias. MLL gene rearrangements are seen in about 80 percent of infant ALL cases and in about 50 percent of infant acute myeloid leukemia.105 ALL associated with MLL translocations are usually CD10- and CD24-negative and CD15-positive.106 In adult and pediatric ALL, the MLL translocation to AF4 is the most frequent, with fusions to ENL, AF9, AF10, and AF6 also common in younger patients. However, MLL translocation can also occur secondary to the use of topoisomerase inhibitors to treat other forms of cancer and, as with de novo MLL translocations in ALL, is associated with a poor prognosis. 103,150 Nearly all the translocations involving MLL disrupt the gene between exons 5 and 11, creating a chimeric fusion gene that encodes the amino-terminal portion of MLL and the C-terminal portion of the translocated partner.104 A partial tandem duplication of a variable region spanning exons 5–12 also converts wild-type MLL into an oncogene.144 One role for the diverse fusion partners is to confer dimerization upon the truncated MLL protein; in some cases, however, the fusion partners encode nuclear proteins with chromatin remodeling activities and probably contribute novel gene regulatory functions to the oncogenic fusion protein. 168,212,238 Oncogenic MLL fusion proteins lead to aberrant expression of Hox genes—e.g., Hoxa9, Hoxa10, Meis 1—and evidence suggests that some MLL fusion proteins are capable of reprogramming gene expression in committed progenitors such that a portion of the normal stem cell transcription program is expressed in the context of a more differentiated cell. 160,182,203 Leukemia stem cells with self-renewal capacity and a limited differentiation program may characterize all leukemias and are the subject of intense investigation.
The ALL cells associated with the t(1;19) translocation contain cytoplasmic β heavy chains. In this translocation, the E2A transcription factor important for B-lineage commitment on chromosome 19 is fused to a homeobox gene called PBX on chromosome 1 that is normally not expressed in B cells. The fusion severs the C-terminal basic helix-loop-helix DNA binding domain but leaves the amino-terminal transactivating domains of E2A. PBX1 has no transactivating capability on its own, but its portion of the fusion protein dictates the interaction of E2A with other homeobox factors and allows for DNA binding and activation of genes controlled by PBX1-HOX complexes.107 A small number of cases of ALL bear a t(17;19)(q22;p13), in which the hepatic leukemia factor (HLF), a basic leucine zipper transcription factor, moves to the E2A site.108 HLF bears homology with the proapoptotic protein of Caenorhabditis elegans called CES-2. However, the HLF-E2A fusion appears to have strong antiapoptotic effects that may contribute to leukemogenesis.109
The features of the t(9;22) translocation gene product that are important in producing chronic myeloid leukemia are well defined and include the activities of the SH2 domain, the tyrosine kinase domain, and the F-actin binding domain.110 However, it is unclear how this translocation leads to ALL. About 50 percent of adult ALLs with this translocation generate a fusion protein called p210; the other half generate an alternative form of the protein, p190. In children with ALL, the p190 form predominates. The differences between the diseases promoted by p210 and p190 also are not clear.
The lymphoblasts in T cell ALL are terminal transferase positive and variably express CD1a, CD2, CD4, CD8, CD7, and cytoplasmic CD3. Myeloid antigens such as CD13 and CD33 may be expressed. Translocations may be present in about one-third of T cell ALL, bringing a number of different genes under the transcriptional influence of the T cell antigen receptor genes, the α and δ loci on 14q11.2, the β locus at 7q35, or the γ locus at 7p14-15 (Table 59.1-4). Partner genes include the transcription factors MYC (8q24); TAL1 (1p32); TAL2 (9q34); RBTN1 or LMO1 (11p15); RBTN2 or LMO2 (11p13); LYL1 or ENL (19p13); and HOX11 (10q24), as well as the src family kinase LCK (1p34.3-35).111 In about 25 percent of cases, deletions in the 5′ regulatory sequence of TAL1 promote its expression without translocation. In both B cell and T cell ALL, deletions interfere with the expression of the cyclin-dependent kinase inhibitors p15 and p16; these may be detected in up to 60 percent of T cell ALL and 25 percent of B cell ALL.112
Less than 1 percent of T-ALL patients harbor a t(7:9) that leads to the activation of Notch1, a plasma membrane protein involved the regulation of cell differentiation, including fate decisions such as B versus T lineage and maintenance of stem cell self-renewal capacity. This rare activating translocation, which is oncogenic in murine models, led to the discovery that most T-ALLs have an activated Notch1 signaling pathway despite lacking the translocation.218 Activating somatic mutations of Notch1 were found in 50 percent of T-ALL leukemic blasts. 174,247 Notch mutations were localized to two sites: (1) the autoinhibitory domain, which regulates cleavage by the γ-secretase complex, and (2) the PEST domain, leading to an increase in half-life of the cleaved Notch-intracellular domain, a transcription factor that associates with the CSL and coactivator complexes to turn on genes known to regulate T cell differentiation, including c-Myc, and the NF-κB and mTOR pathways. 214,234 Approximately 20% of T-ALLs harbor both mutations in cis. Further demonstration of the importance of Notch activation in the pathogenesis of T-ALL is the finding that an additional 30 percent of T-ALLs (all of which lack Notch mutations) have inactivation of Fbxw7, an F-box protein component of the Skip-Cullen-F-Box E3 ubiquitin ligase responsible for the inactivation of Notch-ICD (O’Neil et al, 2007). Hence, like PEST-domain mutations, this leads to dysregulated Notch signaling. T-ALL patients with activated Notch1 have a poor prognosis;51 gamma secretase inhibitors may have efficacy in patients with mutant Notch1 alleles, but leukemia cells harboring Fbxw7 mutations are resistant (O’Neil et al, 2007).
Table 59.1-4: Chromosomal Rearrangements in Precursor T-Cell Neoplasms |Favorite Table|Download (.pdf) Table 59.1-4: Chromosomal Rearrangements in Precursor T-Cell Neoplasms
|Gene-Activation Rearrangement ||Rearranged Gene at Breakpoint ||Activated Gene Near Breakpoint ||Encoded Protein Domain |
|t(8;14)(q24;q11) ||TCRα (14q11) || c-MYC (8q24) ||bHLH |
|t(1;14)(p33;q11) ||TCRδ (14q11) || SCL-TAL/TCL5 (1p33) ||bHLH |
|t(1;7)(p33;q35) ||TCRβ (7q35) || SCL/TAL/TCL5 (1p33) ||bHLH |
|t(1;3)(p33;q21) ||TCTA (3p21) || SCL/TAL/TCL5 (1p33) ||bHLH |
|t(7;9)(q35;q34) ||TCRβ (7q35) || TAL2 (9q34) ||bHLH |
|t(7;19)(q35;q13) ||TCRβ (7q35) || LYL1 (19p13) ||bHLH |
|t(11;14)(p15;q11) ||TCRδ (14q11) || RBTN1/TTG1 (11p15) ||LIM |
|t(11;14)(p13;q11) ||TCRα/δ (14q11) || RBTN2/TTG1 (11p13) ||LIM |
|t(7;11)(q35;p13) ||TCRβ (7q35) || RBTN2/TTG1 (11p13) ||LIM |
|t(10;14)(q24;q11) ||TCRα (14q11) || HOX11 (10q24) ||Homeobox |
|t(7;10)(q35;q13) ||TCRβ (7q35) || HOX11 ||Homeobox |
|t(7;9)(q34;q34) ||TCRβ || TAN1 (9q34.3) ||Notch |
|t(1;7)(p34;q34) ||TCRβ (7q34) || LCK (1p34) ||Receptor tyrosine kinase |
| Gene Fusion Rearrangement || Fusion Gene || Genes Involved in Fusion || Encoded Protein Domains |
|t(11;19)(p23;p13.3) || ALL1/ENL || ALL1/MLL/HRX (11q23) ||Trithorax |
| || || ENL (19q13.3) ||Zinc finger |
|t(X;11)(q13;q23) || ALL1/AFX || ALL1/MLL/HRX (11q23) ||Trithorax |
| || || AFX (Xq13) ||Zinc finger |
Despite all this descriptive biology, we do not fully understand the number of genetic lesions necessary to develop ALL, nor do we understand the order in which the lesions appear. The development of an effective treatment that interferes with the kinase activity of the BCR-ABL fusion protein (Gleevec) raises some hope that targeting a gene fusion translocation can have therapeutic efficacy.113 However, in the instance of BCR-ABL, strong evidence was available in animal models that the kinase activity was necessary and sufficient to develop the neoplasm. Gamma-secretase inhibitors that should prevent the cleavage of Notch and prevent its activation are in clinical development. Most of the translocations in ALL involve gene-activating rearrangements, and it is not clear how such gene products would be selectively targeted. In addition, complete dependence of the tumor on the translocation product may be involved in Burkitt lymphoma/leukemia, but little supportive data have been developed for most of the other genetic lesions.
Plasma cell disorders include multiple myeloma (MM); extramedullary plasmacytoma; solitary plasmacytoma of bone; immunoglobulin deposition diseases such as amyloidosis and osteosclerotic myeloma [associated with the POEMS syndrome: polyneuropathy (sensorimotor demyelination), organomegaly (hepatosplenomegaly), endocrinopathy (diabetes, gynecomastia, testicular atrophy, impotence), monoclonal gammopathy, and skin changes (hyperpigmentation, hypertrichosis)]; and the heavy chain diseases (γ, μ, α). Myeloma is by far the most common of these disorders.
Myeloma is a multifocal bone marrow disease characterized by a monoclonal immunoglobulin in the serum (a tumor product); decreased synthesis and an increased catabolism of other normal immunoglobulins; destructive osteolytic bony lesions without an osteoblastic component; pathologic fractures; bone pain; hypercalcemia; renal failure (from hypercalcemia, recurrent infections, and damage from the tumor immunoglobulin); and anemia. The most important differential diagnosis is with monoclonal gammopathy of uncertain significance (MGUS), which affects about 6 percent of people over age 70 years. MGUS typically has lower levels of serum paraprotein, less than 10 percent plasma cells in the bone marrow, and absence of lytic lesions or other signs and symptoms associated with MM. The risk of MGUS evolving to myeloma is about 1 percent per year, which is a surprisingly low number given the observation that MGUS plasma cells share many of the same genetic lesions seen in early MM (see below).
The malignant cells in myeloma are plasma cells that contain clonally rearranged and usually class-switched immunoglobulin, and the variable region genes are mutated in a manner consistent with derivation from a post–germinal center B cell.114 The cells have lost bcl6 expression but have cytoplasmic immunoglobulin and express CD38- and CD138 (syndecan). In some patients, light chains alone are secreted, causing renal failure; tumor cells sometimes delete their heavy chain genes or portions of them.
The initiating genetic event in approximately 50 percent of MM and MGUS patients is the translocation of novel genetic sequences to the IgH switch region on 14q32.117 The four most common translocations are t(11;14)(q13;q32) involving CCND1 (cyclin D1);153 t(6:14)(p21;q32) that encodes cyclin D3;235 t(14;16)(q32;q23) deregulating expression of the transcription factor c-MAF;154 and t(4;14)(p16;q32), which leads to aberrant expression of fibroblast growth factor receptor 3 (FGFR3, a transmembrane tyrosine kinase receptor) on the derivative 14, and the MMSET transcription factor on der(4).155 MMSET expression may be the more important component of this balanced translocation, as FGFR3 expression is often absent and the der(14) chromosomal locus is sometimes deleted in the MM cells.155 Both t(4;14)(p16;q32) and t(14:16)(q32;q23) are associated with poor prognosis and the downstream upregulation of cyclin D2.
Aneuploidy is a nearly universal finding in myeloma cells, including gains in chromosomes 3, 5, 7, 9, 11, 15, and 19, and losses on 8, 13, 14, 17, and X.118 Given that plasma cells have undergone three processes involving double-strand DNA breaks (VDJ rearrangement, somatic hypermutation, and switch rearrangement), it is not surprising that so many structural errors are detected and the errors seem to increase in patients over time.119 What is surprising is the paucity of information on the role of the genetic instability in the behavior of the disease. Chromosome 13 deletion is present in approximately 50% of MGUS and MM patients and is associated with shorter survival and lower response rate to treatment, 164,173 possibly due to haploinsufficiency of the RB1 locus present in the deleted segment. Investigators have separated MM into two groups based on genetic profiles: hyperdiploid MM with a low prevalence of IgH translocations and nonhyperdiploid MM with frequent IgH translocations.172
Gene expression profiling of MM patients has confirmed the importance of translocations and cyclin D expression in MM classification.140 Furthermore, two studies have identified a profile associated with lytic bone disease.177,222 One study identified upregulation of four genes including DKK-1, an inhibitor of Wnt signaling, which may be at least partially responsible for the loss of osteoblast activity in MM bone disease. Upregulation of the chemokine MIP1α also has been linked to bone lesions; interestingly, there is overexpression of CCR1, the receptor for MIP1α, in MM cells with high levels of c-maf.183
MM cells constitutively express the NF-κB pathway; inhibitors of this pathway, including bortezomib, have clinical utility. Multiple genetic lesions have been identified in MM cells that lead to increased NF-κB activity. 139,191 These include gain-of-function mutation (by amplification, point mutation, or translocation) of receptors (CD40, TAC1, LTβR), kinases (NIK, NF-κB-inducing kinase), or NF-κB family members (p52/p100, p50/p105) that upregulate the classic and/or alternative pathways of NF-κB signaling. Loss of inhibitory regulators of the pathway, such as TRAF2, TRAF3, cIAP1, cIAP2, and the deubiquitinating enzyme CYLD, also was discovered. Restoration or downregulation of mutant genes in vitro using molecular techniques confirmed the functional importance of these genetic lesions.
About 40 percent of patients with myeloma contain activating mutations of N- or K-ras, 115 and 25 percent have deletions in p53; their absence in MGUS suggests that ras and p53 mutations are late events associated with disease progression.116