Gene Expression Profiling Reveals The Lymph Node Microenvironment As a Niche For BCR Engagement, NFκB Pathway Activation, and Tumor Proliferation In Mantle Cell Lymphoma

Mantle cell lymphoma (MCL) and chronic lymphocytic leukemia (CLL) are both mature-looking CD5+ B-cell malignancies that invade the peripheral blood (PB), bone marrow (BM), and secondary lymphoid organs. Both are incurable with the exception of transplant. However, in many regards MCL and CLL appear to be different: MCL is aggressive and treated at diagnosis, while CLL is observed and only treated for symptomatic disease. MCL is characterized by the t(11;14)(q13;q32) resulting in cyclin D1 overexpression, anomalies in cell cycle and DNA repair, and recently by the aberrant expression of SOX11. CLL by contrast, lacks a characteristic mutation or translocation but has been recognized for decades to have a bias for certain IGHV genes, indicating that antigen-BCR interaction is a key pathogenic pathway. More recently, MCL also showed a significant IGHVbias and that many MCL cells, like CLL, express virtually identical BCRs, so called “stereotyped BCRs”, that are predicted to recognize distinct antigens (Hadzidimitriou, Blood 2011).

Comparing purified CLL cells isolated concomitantly from the PB, BM, and lymph nodes (LN), we showed that LN resident cells activates SYK and upregulate BCR-target genes (Herishanu, Blood 2011). This indicates that antigenic signaling continues throughout the disease course and that the BCR is engaged primarily in the LN. Here we used gene expression profiling (Affymetrix HU133 plus) to interrogate whether MCL, similar to CLL, displays changes in tumor biology in different anatomic compartments. We collected a total of 100 PB and LN samples from previously untreated CLL and MCL patients. The samples fall into 5 groups: MCL-PB (purified CD19, n=13), MCL-LN (biopsy, n=29), CLL-PB (purified CD19, n=24), CLL-LN (biopsy, n=17), and CLL-LNT (purified CD19, n=17).

An unsupervised hierarchical clustering of all samples grouped them according to tissue derivation and disease entity into MCL-LN, CLL-LN, MCL-PB, CLL-purified tumor (either PB or LNT). The previously defined MCL-LN signature (Rosenwald, Cancer Cell 2003) distinguished MCL (PB and LN) from each of the CLL groups (2-fold change, p<0.0001). Conversely a CLL specific gene signature (Jelinek, Mol Cancer Res 2003) was preferentially expressed in the 3 CLL groups than in MCL groups (3-fold change, p<0.001). There was no statistically significant difference in the expression of these signatures between PB or LN samples of the same disease entity.

To study the effect of the microenvironment (ME) on tumor cells, we compared a set of three well-defined gene signatures (BCR, NFκB, and MCL-proliferation) across our samples spectrum. An individual signature score was generated for each of the five studied groups, and compared to each other using one-way ANOVA.

All three signatures scored higher in LN samples (MCL or CLL) relative to PB (average 2-fold increase, p<0.01-0.0001). These scores were comparable between MCL-LN, CLL-LN, and CLL-LNT (p=ns), suggesting that first, these signatures reflect effects on resident tumor cells, and second, this effect is similar between MCL and CLL. The proliferation signature used is linked to aggressive disease in MCL (Rosenwald et al, Cancer Cell 2003), and unexpectedly scored equally across all LN groups, taking into consideration the indolent nature of CLL. This was validated using the E2F proliferation signature in CLL (Herishanu, Blood 2011). Both proliferation scores were not different between MCL-PB and CLL-PB. However, there was some variability in BCR and NFκB scores within the MCL-PB group: A subset with high BCR and NFκB scores (comparable to LN), and another with low scores (comparable to CLL-PB). The two subsets were well separated suggesting that the BCR signaling may contribute significantly to NFκB activation.

Activation of MCL tumor cells in LN was confirmed by flow cytometry showing an increase in the activation markers (CD38, CD69, and CD80) and a decrease in the chemokine receptor CXCR4 in LN resident compared to circulating cells in paired samples (n=3).

These data suggest that MCL, like CLL, relies on the LN ME for BCR engagement, cell activation and proliferation, and that both may be critically dependent on BCR signaling. This conclusion is underscored by the equally dramatic clinical responses achieved with the BTK inhibitor ibrutinib.

Supported by the Intramural Research Program of NHLBI and NCI, NIH. We thank our patients for donating samples to make this research possible.