Determining the Cell of Origin in DLBCL

Today we are discussing the way clinicians determine the cell of origin in patients with diffuse large B-cell lymphomas and what to do with this information with Dr. Andrew Zelenetz, MD, PhD, an oncologist who specializes in the treatment of lymphoma at the Memorial Sloan Kettering Cancer Center in New York City. Diffuse large B-cell lymphoma (DLBCL) is an aggressive tumor type that affects a type of white blood cells called B-lymphocytes. It is the most common type of non-Hodgkin lymphoma, accounting for about one-third of all newly diagnosed non-Hodgkin lymphoma in the United States.

-Interviewed by Anna Azvolinsky

Cancer Network: First, DLBCL is known to be quite heterogeneous from patient to patient. What does that mean in the clinic for treating patients?

Dr. Zelentz: We are talking about a disease that can have different presentations. We see isolated presentation in a single lymph node or organ site, or [it] can be a wide-spread disease. Differences in presentation lead to different decisions regarding implementation of treatment, the use of chemotherapy alone or chemotherapy plus radiation. The more advanced stage disease, often, the more difficult it is to treat.

Cancer Network: Researchers, including yourself, have further sub-categorized this tumor type. What are these sub-categories that are referred to as ‘cell of origin’ of DLBCL?

Dr. Zelentz: When we think about diffuse large B-cell lymphoma, we think of specific subtypes such as primary mediastinal large B-cell lymphoma or primary CNS (central nervous system) lymphoma. Most diffuse large B-cell lymphoma is what we would characterize as diffuse large B-cell lymphoma, not otherwise specified.

These tumors can present as I mentioned before, in lymph nodes, either localized or more widespread. Seminal work now nearly 20-years-old identified that diffuse large B-cell lymphoma is actually not all the same thing. In fact, we can see diffuse large B-cell lymphoma arising from cells that are activated B-cells which is called activated B-cell large B-cell lymphoma.

They can arise from types of more differentiated cells called germinal center cells and that subtype is called germinal center B-cell large B-cell lymphoma. We have a group of tumors that are unclassifiable and then we have this group of tumors derived from thymic B-cells, which give rise to the primary mediastinal large B-cell lymphomas.

Cancer Network: How do you diagnose these specific subtypes? Is that done by pathology or by genetics? Are there different clinical characteristics of these different subtypes? What does that mean in terms of current treatment options and clinical trials?

Dr. Zelentz: The original definition of diffuse large B-cell lymphoma and the cell of origin was based on different gene expression profiles and looking at a large number of genes and the pattern of gene expression within a tumor. That led to the classification of either germinal center or activated B-cell or unclassifiable lymphomas.

The technique used originally requires fresh tissue and is a laboratory-based procedure. There have been a large number of attempts to try to reduce this to a simpler test, to use routine immunohistochemistry where we take antibodies to specific proteins and ask if the tumor is expressing or not expressing certain proteins. The most commonly used method is called the Hans algorithm which uses 3 antibodies, CD10, BCL-6 and MUM1 and depending on the expression of these 3 proteins, we can classify tumors broadly into germinal center and non-germinal center.

Unfortunately, this is not a very precise measurement. We took the expression profile of 14,000 genes and reduced it to 3 genes and that results in quite a bit of heterogeneity. There is another test that has been developed which uses some of the key genes that were identified by gene expression analysis.

[The test] uses a fluorescent reporter assay and can be used on formalin-fixed, parafin-embedded tissue. This so-called nano-string assay allows us to much more accurately classify tumors as either germinal center, activate B-cell, or unclassifiable. The immunohistochemistry method cannot separate out the unclassifiable from the activate B-cell tumors.

The reason why this is important is that it has prognostic significance. In many studies, patients who have the germinal center diffuse large B-cell lymphoma have a more favorable outcome than patients with activated B-cell lymphoma. But, what we have seen increasingly, is that there are treatments that are more active in some types vs others.

For instance, the drug lenalidomide appears to be much more active in the activated B-cell large B-cell lymphoma and similarly, a drug called ibrutinib is also more active in this subtype. There have been two large studies comparing standard treatment, rituximab plus CHOP [cyclophosphamide, doxorubicin, vincristine, and prednisone] with or without either ibrutinib or lenalidomide and asking, in the subgroup of tumors that are activated B-cell tumors, can we in fact improve the outcome by adding these drugs to the general mix of rituximab plus CHOP, the standard chemotherapy regimen.

Among the germinal center tumors, they are much more likely to have mutations in the gene called EZH2 and a drug tazemetostat has significant activity in tumors that are mutated in EZH2. Interestingly, there is also activity in the tumors that don’t have the mutations.

The importance of testing for cell of origin is nice when all you have to do is help determine the prognosis. But, if the treatment is going to change, then it becomes essential to test for cell of origin. Increasingly, we will see greater use of the more precise assays such as the Nano-string assay if these clinical trials are positive.

Cancer Network: You mentioned a few clinical trials already. Are there any other specific clinical trials for some of these subtypes of DLBCL that are ongoing now that could be important in either understanding the prognosis of these tumors or for potentially improving the efficacy of treatment for patients?

Dr. Zelentz: I mentioned two of the most important studies, these are very large studies and we are waiting for the results and that is rituximab plus CHOP and lenalidomide vs rituximab plus CHOP and placebo. All patients receive what looks like the exact same treatment but only half of the patients get the lenalidomide. A similarly designed trial with ibrutinib is currently ongoing.

We recently completed a national high-priority study of rituximab and CHOP vs dose-adjusted EPOCH [cyclophosphamide and oral prednisone] to test a chemotherapy regimen developed at the National Cancer Institute by Wyndham Wilson and his colleagues. [This study was conducted] to see if [EPOCH] was better than R-CHOP. 

Unfortunately, the overall outcome that is R-CHOP vs dose-adjusted EPOCH was relatively equivalent and there was no statistically significant difference between the two treatments. But there might be important differences in treatment response based on molecular subtypes and we are waiting for that analysis.

Cancer Network: Lastly, are you and other researchers working to uncover additional DLBCL subtypes? Or do we understand the important molecular heterogeneity of these subtypes?

Dr. Zelentz: Unfortunately, it is a whole lot more complicated than the germinal center and activate B-cell tumors. Two groups have recently shown that we can further subdivide these into tumors based on not only their gene expression profile, which helps us determine cell of origin, but then we can further divide tumors based on molecular features of the tumor.

These papers have just recently been published. One is from Louis Staudt’s group at the NCI [National Cancer Institute] and one is from Margaret Shipp’s group at the Dana Farber Cancer Center in Boston.

The Shipp group reports that there are 5 separate categories of diffuse large B-cell lymphoma that can be categorized, combining cell of origin plus gene expression data. Similarly, we can see molecular subtypes reported by the Staudt group and they report 7 different clusters of large cell lymphoma based on further characterization.

We see now that there are germinal center lymphomas that have very good outcomes and others with poor outcomes so again, they are heterogeneous and are not all behaving the same way.

Interestingly, we are starting to see patterns of gene expression and mutation patterns that might suggest the ability to use new targeted therapies in these subgroups. The major challenge that we have, of course, is that we’ve known about cell of origin since 2000 and we are still not using this routinely to improve the outcomes of patients with diffuse large B-cell lymphoma. These new classifications include not only gene expression profiling but the mutational analysis and also structural variants within the chromosomes.

What we have to do is work on a system that allows us to identify these tumors very rapidly so that if there is going to be a difference in treatment for these subtypes, we can identify those differences in treatment very early and get the right treatment to the right patient. This is really going to be a major challenge going forward on how to molecularly characterize these tumors in a time efficient manner that we can use to change treatment and improve outcomes.

Cancer Network: Thank you so much for joining us today Dr. Zelentz.

Dr. Zelentz: Thank you!