Gene Expression Can Predict the Survival of Lymphoma Patients

Reading time: 5 – 8 minutes

ResearchBlogging.orgA study published last week in the New England Journal of Medicine identified sets of genes in diffuse large B-cell lymphoma that influence the effectiveness of chemotherapy [1]. Researchers found that the tumor microenvironment, consisting of non-cancerous immune and structural cells along with tumor cells, has a significant influence on a patient’s response to treatment.

Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoid malignancy in adults. It is an aggressive form of non-Hodgkin lymphoma, a tumor that originates in lymphocytes, a type of white blood cell. DLBCL tumors often grow in lymph nodes, which are small glands present throughout the body that act as filters for bacteria and are critical for the body’s immune response.

Rituximab targeting a lymphoma cellAlthough phenotypically similar, gene expression profiling has identified two DLBCL molecular subtypes that vary biologically and clinically [2]. The germinal-center B-cell-like (GCB) subtype is more responsive to treatment than the activated B-cell-like (ABC) subtype. The standard chemotherapeutic regime for DLBCL is a combination therapy known as CHOP, consisting of cyclophosphamide, hydroxydoxorubicin (also known as doxorubicin, brand name Adriamycin), Oncovin (also known as vincristine) and prednisone. Recently, the addition of rituximab, an antibody that binds to the surface of malignant B-cells, has improved overall survival of DLBCL patients by 10 to 15%. This combination therapy, called R-CHOP, is able to cure 50% — 60% of patients.

To evaluate the biological basis of survival after therapy for diffuse large-B-cell lymphoma, scientists at the National Cancer Institute (NCI) profiled gene expression in pretreatment biopsy samples obtained from 181 patients treated with CHOP or 233 patients treated with R-CHOP, and searched for gene expression signatures of different aspects of tumor biology that were associated with survival.

The researchers used DNA microarray technology to evaluate the expression level of tens of thousands of genes in the tumor biopsy samples. They found three gene expression signatures that divided CHOP- and R-CHOP-treated patients into subgroups that had long or short survival. To assess whether the signatures were derived from malignant lymphoma cells or from the host environment, the researchers then separated malignant from nonmalignant cells (based on the CD19 antigen) in three biopsy samples.

Genes in the first signature, the germinal-center B-cell (GCB) signature, were more highly expressed in malignant tumor cells than in nonmalignant cells. Previous research has shown that the GCB subtype is more responsive to chemotherapy [3]. In contrast, the other two gene expression signatures, termed stromal-1 and stromal-2, were more highly expressed in nonmalignant tumor cells and were thought to reflect the nature of the tumor microenvironment. The stromal-1 signature expressed genes involved in forming or modifying the extracellular matrix, which provides cellular structural support and regulates tissue structure and function. Additionally, the stromal-1 signature was enriched with genes expressed in monocytes, a type of leukocyte (white blood cell) that is part of the body’s immune system. Tumors with high relative expression of stromal-2 genes were associated with new blood-vessel formation, a process termed angiogenesis that is necessary for tumors to grow. The stromal-1 signature was associated with better prognosis while the stromal-2 signature was associated with poor prognosis.

Study leader Louis M. Staudt, M.D., Ph.D., said [4]:

The ability of a patient with DLBCL to be cured by our current therapy can be predicted by looking at the pattern of gene activity in the tumor biopsy sample taken at diagnosis. In the near term, we need to incorporate gene expression profiling, which measures gene activity, in clinical trials for this disease to allow researchers to standardize results according to the variety of DLBCL tumors included in the trial. In the longer term, new therapies will emerge that are tailored to the particular gene expression profile of a patient’s lymphoma.

This study exemplifies how gene expression can be used as a clinical prognostic factor. Indeed, such studies herald the future of genomic medicine. Specifically for DLBCL, this study suggests that only a subgroup of patients whose tumors are characterized by high relative expression of the stromal-2 signature may benefit from drugs that inhibit angiogenesis.


  1. Lenz et al. Stromal Gene Signatures in Large-B-Cell Lymphomas. N Engl J Med. 2008 Nov 27;359(22):2313-2323. DOI: 10.1056/NEJMoa0802885
    View abstract
  2. Alizadeh et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000 Feb 3;403(6769):503-11.
    View abstract
  3. Rosenwald et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N Engl J Med. 2002 Jun 20;346(25):1937-47.
    View abstract
  4. Variations in Gene Activity Can Predict the Survival of Patients with Lymphoma. National Institutes of Health press release. 2008 Nov 26.
About the Author

Walter Jessen, Ph.D. is a Data Scientist, Digital Biologist, and Knowledge Engineer. His primary focus is to build and support expert systems, including AI (artificial intelligence) and user-generated platforms, and to identify and develop methods to capture, organize, integrate, and make accessible company knowledge. His research interests include disease biology modeling and biomarker identification. He is also a Principal at Highlight Health Media, which publishes Highlight HEALTH, and lead writer at Highlight HEALTH.