Abstract
Introduction: Rituximab, a chimeric IgG1 monoclonal antibody directed at the B-cell antigen CD20, has become a mainstay in the treatment of follicular lymphoma. Unfortunately, not all patients respond to therapy, and therefore the identification of factors that accurately predict response has become an active area of investigation. Recently, gene expression profiling analysis of whole lymph nodes from follicular lymphoma patients resulted in the identification of numerous genes that are differentially expressed in responders versus nonresponders (Bohen, et al. PNAS 100: 1926–1930, 2003). However, as the gene expression profiling technique still requires clinical validation and is also a cost- and time-intensive method, we sought to identify immunohistologic factors that may act as predictors of clinical response to Rituximab. As the gene profiling data suggested a relationship between the tumor immune and clinical responses, we reasoned that characterizing the different immune cell populations in lymph node biopsies by immunohistochemistry might similarly provide prognostic information for Rituximab response.
Methods/Materials: Diagnostic lymph node biopsies from 31 patients diagnosed with follicular lymphoma and for whom the clinical response to Rituximab was known were included in the study. These included lymph nodes from 20 of the 24 patients included in the previously described gene expression profiling study. 12 non-responders, 10 partial responders, and 9 complete responders were evaluated, representing 11 grade 1, 17 grade 2, and 3 grade 3 follicular lymphomas. Numerous immunohistochemical stains were performed, including markers used to diagnose follicular lymphoma (CD20, CD10, bcl-2, Ki-67), as well as additional markers to evaluate tumor-infiltrating T-cells (CD3, CD4, CD8, CD25, TIA-1), follicular dendritic cell networks (CD21, CD23), and monocytes/macrophages (CD163). For each marker, the number, distribution, intensity and variability of staining were scored.
Results: Utilizing multiple logistic regression analyses, we found that increased numbers of intrafollicular CD25+ cells showed a trend towards being associated with partial or complete response (p=0.064); however, the number of intrafollicular CD25+ cells was less reliable in distinguishing partial (p=0.11) from complete responses (p=0.12). The number of intrafollicular CD3+ T-cells was also positively correlated with complete response (p=0.0467), but did not reliably distinguish between partial responders or nonresponders. None of the other markers evaluated showed a statistically significant association with response.
Conclusions: Our data suggest that the clinical response to Rituximab is positively associated with increased numbers of CD25+ tumor-infiltrating cells. This result runs counter to the previously published gene expression profiling data that suggested a negative association between T-cells and clinical response. As our initial sample size was limited, we are presently evaluating additional biopsies and immunohistochemical markers to validate and refine our observations. Nevertheless, these initial observations provide evidence that an immunohistochemical assay may eventually prove useful in predicting clinical responses to Rituximab therapy in follicular lymphoma patients.
Author notes
Corresponding author
