DNA Microarrays in Lymphoid Malignancies

DNA Microarrays in Lymphoid Malignancies

ABSTRACT: Gene expression profiling using cDNA microarrays has the potential to improve current lymphoma classification schemes by establishing a molecular diagnosis of these malignancies. The use of this technology led to the discovery of biologically and clinically distinct subtypes of diffuse large B-cell lymphoma (DLBCL). Gene expression data can also be used to formulate powerful mathematical algorithms that predict the clinical outcome in patients with DLBCL and mantle cell lymphoma. In B-cell chronic lymphocytic leukemia, gene expression profiling identified ZAP70, an important prognostic marker whose expression correlates with the mutational status of the immunoglobulin heavy chain gene and, therefore, with survival in these patients. These examples illustrate that gene expression profiling may pave the way for detailed molecular characterization of lymphoid malignancies that will ultimately lead to tailored, disease-specific therapies.

The diagnosis of lymphoid malignancies currently relies on the morphologic appearance of the tumor cells, their immunophenotype, and the genetic and clinical aspects of the disease.[1] By including genetic information and the expression of a handful of molecular markers, this approach can be viewed as a first attempt to define malignant lymphomas on a molecular level. However, many diagnostic entities are still heterogeneous in their clinical course or in their response to a particular therapy. Even in lymphoma subgroups that share a common genetic alteration, a marked divergence in their clinical outcome is frequently observed. For example, mantle cell lymphoma (MCL) is now recognized as a distinct entity characterized by the chromosomal translocation t(11;14), which leads to overexpression of cyclin D1, a key regulator of the G1/S cell-cycle checkpoint.[2] Despite the presence of this translocation in more than 90% of MCL cases, some patients die within 1 year of diagnosis whereas others live for more than a decade.[3] It is likely, therefore, that this variability is due to additional molecular differences between the tumors that have not yet been captured by current classification criteria. The information provided by the Human Genome Project in concert with advances in high-throughput technologies have paved the way for a comprehensive molecular characterization of lymphomas and cancers in general. In particular, the development of cDNA microarray technology has made it possible to measure the expression of thousands of genes simultaneously in a single experiment.[4] Using this technology, molecular portraits of tumors can be established, and the gene expression profiles of multiple tumor specimens can be compared. This review will illustrate the effectiveness of gene expression profiling in defining new subtypes of lymphoma, in predicting the clinical outcome of lymphoma patients, and in discovering novel predictive markers that may be useful for clinical application. Novel Subtypes of Diffuse Large B-Cell Lymphoma Approximately 22,000 patients are newly diagnosed with diffuse large B-cell lymphoma (DLBCL) in the United States each year.[5] Current chemotherapy regimens, usually involving anthracyclines-eg, CHOP (cyclophosphamide [Cytoxan, Neosar], doxorubicin HCl, vincristine [Oncovin], prednisone)-achieve durable remissions in less than 50% of patients. Multiple attempts have been made in the past to improve clinical outcome in the majority of DLBCL patients, but these attempts have been largely unsuccessful.[ 6] Histopathologic subclassification of DLBCL into various morphologic subtypes suffers from a lack of interobserver reproducibility, and a clinical relevance has not been convincingly demonstrated.[7] The striking heterogeneity in response to chemotherapy among DLBCL patients raises the possibility that this entity actually encompasses more than one disease on the molecular level; however, despite its well-recognized biologic and clinical heterogeneity, it is still considered a single lymphoma entity in the most recent World Health Organization classification of lymphoid malignancies.[1] Gene Expression Studies
Over the past 3 years, gene expression profiling using cDNA microarrays has been used to decipher the underlying heterogeneity responsible for the varying clinical outcome in DLBCL patients on a molecular level.[8-10] An initial gene expression study, in which lymph node specimens from previously untreated DLBCL patients were analyzed, demonstrated a remarkable degree of heterogeneity in gene expression in this disease.[10] However, genes that are typically expressed in the germinal center stage of B-cell differentiation could be used to define two major subgroups of DLBCL patients, namely the "germinal center B-cell-like" type and the "activated B-cell-like" type.[9,10] The germinal center B-cell DLBCL subtype expresses CD10, BCL-6, JAW1, and other genes at high levels characteristic of this B-cell differentiation stage, whereas the activated B-cell DLBCL subtype expresses these genes at relatively low levels and instead expresses genes that are induced during activation of peripheral blood B cells (eg, by mitogenic stimulation), such as cyclin D2, IRF-4, and CD44 (Figure 1). A large follow-up study that included gene expression profiling of 274 DLBCL patients confirmed the existence of the two subgroups and identified a third subset- type III DLBCL.[10] These data suggest that the two subtypes of DLBCL may originate from different cells (germinal center B-cell vs non-germinal center B-cell) and may be pathogenetically distinct. This hypothesis is supported by analysis of the mutational status of the immunoglobulin heavy chain (IgVH) gene. IgVH mutations occur in B cells during the germinal center reaction in a process called affinity maturation. Although most, if not all cases of DLBCL appear to carry mutated IgVH genes, only germinal center Bcell DLBCL cases display the phe- nomenon of "ongoing somatic mutations"- a hallmark of germinal center B cells-suggesting that these DLBCL cases retain characteristics of normal B cells at this stage of differentiation.[ 11] In contrast, activated B-cell DLBCL cases do not display intraclonal variation in their mutated IgVH genes and may, therefore, be derived from normal B cells that have already passed through the germinal center (eg, post-germinal center B cells).[11] In addition to differences in the proposed normal B-cell counterpart, the two DLBCL subgroups also use distinct oncogenic mechanisms. In particular, the chromosomal translocation t(14;18), leading to upregulation of the bcl-2 oncogene and genomic amplification of the short arm of chromosome 2 (including the c-rel oncogene) are exclusively detected in germinal center B-cell DLBCL cases but not in activated B-cell DLBCL cases.[12,13] In contrast, activated B-cell DLBCL cases were shown to have constitutive activity of the nuclear factor-kappaB pathway, which protects cells from programmed cell death. In vitro interference with this pathway was toxic to activated B-cell DLBCL-like cell lines, but no effect was seen in germinal center B-cell DLBCL-like cell lines, suggesting that this pathway might represent an attractive therapeutic target in patients with activated B-cell DLBCL.[14] Most importantly, however, the two DLBCL gene expression subgroups had distinct overall survival rates following CHOP-based therapy. The 5-year survival rates for germinal center B-cell and activated B-cell DLBCL patients were 60% and 35%, respectively.[10] Outcome Prediction by Gene Expression Profiling Gene expression profiling can also be used to create mathematical models that predict overall survival or response to a given therapy. For example, germinal center B-cell DLBCL patients have a more favorable clinical outcome compared to activated B-cell DLBCL patients, and several lines of evidence suggest that these two DLBCL subtypes represent pathogenetically distinct entities. However, this distinction does not fully capture the clinical variability of this disease, as 40% of the patients in the favorable group (germinal center B-cell DLBCL) die within 5 years of diagnosis, and 35% of those in the prognostically unfavorable group (activated B-cell DLBCL) are still alive at 5 years. This suggests that additional biologic features may influence overall survival in patients with this disease. Gene Expression Signature
With a supervised analysis approach, gene expression data can also be used to search for individual genes, the expression levels of which correlate with the length of survival. These genes can then be grouped into biologic "signatures" that represent groups of genes expressed coordinately in a particular cell type during a particular state of cell activation or in response to extracellular stimuli.[15] In DLBCL, this "signature" approach identified five features that influence survival in these patients. As demonstrated previously, the germinal center B-cell signature predicted a favorable outcome, as did the major histocompatibility complex class II signature and the lymph node signature.[10] The favorable predictive value of the latter two signatures suggests that the immunologic response of the host to the tumor cells plays an important role in determining the response to chemotherapy. Conversely, the proliferation signature and the expression of bone morphogenetic protein-6 were unfavorable prognostic indicators. By combining gene expression levels of representative genes from the four gene expression signatures and bone morphogenetic protein-6, a multivariate outcome predictor for DLBCL patients could be created.[10] This predictor divided patients into quartiles with strikingly different 5-year survival rates of 73%, 71%, 34%, and 15% (Figure 2). Because the predictor involves only a limited number of genes (17 genes), a diagnostic test (eg, a multiplexed reverse transcriptase-polymerase chain reaction [RT-PCR] assay or diagnostic miniarrays) could easily be developed for routine clinical application. Mantle Cell Lymphoma
Gene expression profiling has also provided insights into the pathogenesis and clinical behavior of mantle cell lymphoma (MCL).[16] Most cases of MCL are characterized by a common chromosomal translocation, t(11;14), which juxtaposes the cyclin D1 gene to the IgVH gene locus.[2] Gene expression profiling demonstrated that MCL is characterized by signature genes, which distinguish it from other non-Hodgkin's lymphomas.[ 16] Moreover, this MCL gene expression signature was used to identify a novel subset of MCLs that lack the t(11;14) translocation and cyclin D1 overexpression but resemble "classic" cyclin D1-positive MCLs morphologically and clinically. Despite the homogeneous gene expression signature of all MCLs, a search was made for individual genes whose expression levels correlate with survival. Much of the variability in survival could be accounted for by differences in the proliferation gene expression signature, with higher expression of this signature associated with a worse overall survival. Using this measure of the tumor cell proliferation rate, MCL patients could be subdivided into quartiles with median survival times of 0.8, 2.3, 3.3, and 6.7 years (Figure 3). Two oncogenic mechanisms were identified that accounted for some of the variability in the proliferation and survival rates. Some of the more proliferative MCLs express higher levels of cyclin D1 mRNA due to the preferential expression of a more stable isoform of cyclin D1 mRNA.[16] In addition, deletions of the INK4a/ARF tumor-suppressor locus are common among the highly proliferative MCLs. Both of these oncogenic events were independently associated with shorter survivals, but a statistical model that combined these two events did not predict length of survival as well as proliferation gene expression signature alone. Thus, the proliferation signature can be viewed as a quantitative integrator of multiple oncogenic events that affect the clinical course of MCL patients. Novel Prognostic Marker in B-Cell Chronic Lymphocytic Leukemia B-cell chronic lymphocytic leukemia (B-CLL) is regarded as a relatively indolent but incurable form of leukemia. The clinical course of these patients, however, can be highly variable: Some patients will never require any therapeutic intervention and are more likely to die with their disease rather than from it; others experience rapid clinical progression despite intensive treatment.[17] The clinical staging systems by Rai[18] and Binet[19] are widely used to assess the prognosis of B-CLL patients. In many patients, however, B-CLL is diagnosed at an early stage, for example, during a routine medical examination. This represents a challenging situation for both the patient and the physician because, with earlystage B-CLL, the Rai and Binet classifications are unable to predict whether a patient falls into the favorable or poor prognostic group-ie, whether a particular case of B-CLL is stable or likely to progress. Mutated and Unmutated IgVH Genes
The dramatic clincal heterogeneity of B-CLL patients is likely to be reflected in molecular differences in tumor cells and, therefore, a search for molecular correlates of the stable and progressive form of B-CLL was undertaken in recent years. In 1999, two landmark studies described a link between the presence or absence of somatic mutations of the B-cell receptor IgVH gene and the clinical course of the disease.[20,21] In particular, B-CLL patients with unmutated IgVH genes had a median survival of only 95 months and a tendency toward advanced clinical stage, atypical morphology, and rapid disease progression.[ 21] In contrast, patients with mutated IgVH genes frequently did not require therapy and had a median survival of 293 months.[21] The finding that B-CLL cells can carry either somatically mutated or unmutated IgVH genes led to the concept that these two subtypes may be derived from different stages of B-cell differentiation. Because somatic mutations of the IgVH genes occur during the germinal center reaction,[22] B-CLL cells with mutated IgVH genes are thought to be derived from post- germinal center B-cells, whereas B-CLL cells with unmutated IgVH genes are derived from pre-germinal center B cells. Aside from different cells of origin, the two B-CLL subtypes are characterized by different genetic alterations, which might also contribute to their distinct clinical behavior. The prognostically unfavorable cytogenetic deletions of 11q and 17p are usually acquired by IgVHunmutated B-CLL cells, while IgVHmutated tumor cells frequently carry prognostically favorable deletions in 13q.[23] Given the differences between the clinical courses of B-CLL patients with IgVH-mutated and those with IgVH-unmutated tumor cells, it would be clearly beneficial to determine the IgVH mutational status in each B-CLL patient as part of the routine work-up. Most clinical laboratories, however, do not have the ability to routinely sequence the IgVH genes; moreover, the procedure is time-consuming and costly. ZAP70
Gene expression analysis in B-CLL revealed that all tumor cells, regardless of their IgVH mutational status, share a common gene expression program, suggesting that despite its clinical heterogeneity B-CLL should be considered a single disease.[13,24] However, roughly 175 genes were discovered for which gene expression levels correlate with IgVH mutational status and, therefore, with clinical outcome.[ 13,24,25] ZAP70, a tyrosine kinase critical for signaling through the T-cell receptor and not previously demonstrated to be expressed in B cells, was found to be the most discriminating gene between the two IgVH mutational subgroups of B-CLL patients, with higher expression of ZAP70 present in IgVH-unmutated B-CLL cells (Figure 4).[13,25] Possible technologic platforms for the clinical application of ZAP70 expression in B-CLL cells include semiquantitative or quantitative RTPCR assays, immunohistochemistry (IHC), and flow-cytometric analysis. Quantitative PCR assays are already in use in routine clinical laboratories and could accurately discriminate between the two B-CLL subtypes. However, because ZAP70 is highly expressed in T cells, this approach would require purification of leukemic cells, eg, by magnetic separation of CD19- positive cells. Protein expression of ZAP70-for example, detected by IHC-also correlates well with IgVH mutational status.[25] This technique is easy to perform and does not require purification of the tumor cells. On the other hand, IHC is only semiquantitative in nature and may lead to misinterpretation in B-CLL cases with intermediate expression of ZAP70. Probably the most feasible technique for use in a routine clinical setting is the measurement of ZAP70 protein expression by multiparameter flow-cytometric analysis.[26] This approach offers the advantage that ZAP70 expression can be selectively analyzed in B-CLL cells, T cells, and natural killer cells. Moreover, flowcytometric analysis of tumor cells is part of the routine work-up of B-CLL patients and, therefore, staining for ZAP70 could easily be included in the diagnostic procedure. Conclusions These examples illustrate that gene expression profiling can identify new molecular subgroups of malignant lymphomas. Hopefully, this approach will lead to an improved molecular classification of lymphoma subgroups that are more uniform in their biologic and clinical behavior. Ultimately, the value of a molecular diagnosis lies in its utility in the clinical setting-ie, by identifying the optimal therapy for a given patient. In addition, detailed molecular characterization of lymphomas will hopefully lead to the discovery of pathways that are aberrantly activated or suppressed in these malignancies and that will guide future therapeutic drug development.


The author(s) have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.


1. Jaffe ES, Harris NL, Stein H, et al: Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France, IARC Press, 2001.
2. Campo E, Raffeld M, Jaffe ES: Mantlecell lymphoma. Semin Hematol 36:115-127, 1999.
3. Swerdlow SH, Williams ME: From centrocytic to mantle cell lymphoma: A clinicopathologic and molecular review of three decades. Hum Pathol 33:7-20, 2002.
4. Schulze A, Downward J: Navigating gene expression using microarrays—a technology review. Nat Cell Biol 3:E190-195, 2001.
5. Coiffier B: Diffuse large cell lymphoma. Curr Opin Oncol 13:325-334, 2001.
6. Fisher RI: Diffuse large-cell lymphoma. Ann Oncol 11(suppl 1):29-33, 2000.
7. Harris NL, Jaffe ES, Diebold J, et al: Lymphoma classification—from controversy to consensus: The R.E.A.L. and WHO Classification of lymphoid neoplasms. Ann Oncol 11(suppl 1):3-10, 2000.
8. Shipp MA, Ross KN, Tamayo P, et al: Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning. Nat Med 8:68-74, 2002.
9. Rosenwald A, Wright G, Chan WC, et al: The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N Engl J Med 346:1937-1947, 2002.
10. Alizadeh AA, Eisen MB, Davis RE, et al: Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503-511, 2000.
11. Lossos IS, Alizadeh AA, Eisen MB, et al: Ongoing immunoglobulin somatic mutation in germinal center B cell-like but not in activated B cell-like diffuse large cell lymphomas. Proc Natl Acad Sci U S A 97:10209-10213, 2000.
12. Huang JZ, Sanger WG, Greiner TC, et al: The t(14;18) defines a unique subset of diffuse large B-cell lymphoma with a germinal center B-cell gene expression profile. Blood 99:2285-2290, 2002.
13. Rosenwald A, Alizadeh AA, Widhopf G, et al: Relation of gene expression phenotype to immunoglobulin mutation genotype in B cell chronic lymphocytic leukemia. J Exp Med 194:1639-1647, 2001.
14. Davis RE, Brown KD, Siebenlist U, et al: Constitutive nuclear factor kappaB activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J Exp Med 194:1861-1874, 2001.
15. Shaffer AL, Rosenwald A, Hurt EM, et al: Signatures of the immune response. Immunity 15:375-385, 2001.
16. Rosenwald A, Wright G, Wiestner A, et al: The proliferation gene expression signature is a quantitative integrator of oncogenic events that predicts survival in mantle cell lymphoma. Cancer Cell 3:185-197, 2003.
17. Hamblin T: Chronic lymphocytic leukaemia: One disease or two? Ann Hematol 81:299-303, 2002.
18. Rai KR, Sawitsky A, Cronkite EP, et al: Clinical staging of chronic lymphocytic leukemia. Blood 46:219-234, 1975.
19. Binet JL, Auquier A, Dighiero G, et al: A new prognostic classification of chronic lymphocytic leukemia derived from a multivariate survival analysis. Cancer 48:198-206, 1981.
20. Damle RN, Wasil T, Fais F, et al: Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood 94:1840-1847, 1999.
21. Hamblin TJ, Davis Z, Gardiner A, et al: Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood 94:1848-1854, 1999.
22. Rajewsky K: Clonal selection and learning in the antibody system. Nature 381:751- 758, 1996.
23. Krober A, Seiler T, Benner A, et al: V(H) mutation status, CD38 expression level, genomic aberrations, and survival in chronic lymphocytic leukemia. Blood 100:1410-1416, 2002.
24. Klein U, Tu Y, Stolovitzky GA, et al: Gene expression profiling of B cell chronic lymphocytic leukemia reveals a homogeneous phenotype related to memory B cells. J Exp Med 194:1625-1638, 2001.
25. Wiestner A, Rosenwald A, Barry TS, et al: ZAP-70 expression identifies a chronic lymphocytic leukemia subtype with unmutated immunoglobulin genes, inferior clinical outcome, and distinct gene expression profile. Blood 101:4944-4951, 2003.
26. Crespo M, Bosch F, Villamor N, et al: ZAP-70 expression as a surrogate for immunoglobulin- variable-region mutations in chronic lymphocytic leukemia. N Engl J Med 348:1764-1775, 2003.
Loading comments...
Please Wait 20 seconds or click here to close