The NSG – Human CLL Xenograft Model Recapitulates the Human Lymph Node Microenvironment in Regards to B-Cell Activation and Tumor Proliferation

Abstract 973

Chronic lymphocytic leukemia (CLL) cells depend on signals from the tissue microenvironment for proliferation and survival. A comparative analysis of lymph node (LN) and peripheral blood (PB) derived CLL cells revealed dynamic B-cell receptor (BCR) and NF-kB activation in this microenvironment (Herishanu et al, Blood 2011). To dissect the pathogenic role of different signaling pathways in vivo, it is crucial to develop a model system that can reproduce the tumor biology encountered in the human LN. We hypothesized that secondary lymphoid compartments (such as the bone marrow (BM) or spleen) of mice may recapitulate the human LN microenvironment. In order to study the effect of the murine microenvironment on human CLL cells we chose to use the recently established NOD scid gamma null (NSG) - human CLL xenograft model (Bagnara et al., Blood 2011), with slight modifications: no adoptive transfer with normal human CD34 or CD14 positive cells was undertaken, and NSG mice were conditioned with 25 mg/kg busulfan instead of g-irradiation. 24 hours later, 1 × 10⁸ CFSE labeled CLL PB mononuclear cells (PBMCs) were injected through the tail vein. Mice were bled weekly and sacrificed between two and six weeks post xenograft. The PB CLL cell count (evaluated as hCD45+, hCD19+, hCD5+ cells) one week after injection averaged 300 cells/μL but then rapidly declined and became virtually undetectable by week four. In contrast, CLL cells persisted in the spleen and BM. We next sought to determine if tumor proliferation was occurring in the xenograft model. We found that CLL proliferation (measured by the fraction of CFSE low cells) was barely apparent early (weeks 1–2) but increased significantly by weeks 3–4 (p=0.02). A similar trend was also observed for T-cells. There was no significant difference in the fraction of CFSE low cells among the three different compartments; however, the number of CLL cells in the spleen was significantly higher than in the PB or BM, suggesting increased homing to the spleen. We next assessed if the mouse spleen microenvironment would modulate the expression of CD38, CD69 and CD184 (CXCR4). Xenografted CLL cells extracted from murine spleens showed significantly increased expression of CD38 (p=0.01) and CD69 (p<0.001) and significantly decreased expression of CD184 (p<0.001) compared to PB cells, recapitulating findings previously reported in patients. We have previously shown that LN-resident CLL cells display gene expression signatures indicating BCR and NF-kB activation. We therefore sought to determine if the mouse spleen microenvironment would modulate gene expression in CLL cells. We compared patient derived PB-CLL cells to either matched patient derived LN-CLL cells or matched xenografted CLL cells isolated from the murine spleen and evaluated expression of BCR, NF-kB, and MYC target genes using quantitative RT-PCR (pre-designed Taqman Gene Expression assays). BCR (AICDA, EGR1, EGR3, GFI1, KLF10, OAS3) and NF-kB (CCL4, CCL3, RGS1, TNF, CCND2) target genes were upregulated in CLL cells from both the human LN and murine spleen. In addition, genes associated with proliferation (CDT1, PCNA, RRM2) were upregulated in CLL cells from both LN and murine spleen; consistent with the proliferation measurements using CFSE. However, MYC target genes (MYC, DUSP1 and HSPD1) were upregulated only in CLL cells from LN but not in the murine spleen. It has been hypothesized that blocking CXCR4 on tumor cells could inhibit their ability to home and interact with the tissue microenvironment. We therefore tested the effect of AMD3100 – a CXCR4 antagonist (5 mg/kg/day by continuous subcutaneous infusion) in this model. Unexpectedly, there was no effect on leukemic cell proliferation or on the CLL count in the PB of treated as compared to untreated mice. However, we found a dramatic increase in CXCR4 expression on CLL cells in the treated mice, indicating adequate drug delivery. Taken together, our results suggest that the murine spleen microenvironment can adequately recapitulate that of the human LN. These results provide further justification for the use of the NSG - human CLL xenograft model to study both the pathogenic mechanisms that contribute to disease progression in the tissue microenvironment and as a pre-clinical model for drug development and assessment.

Supported by the Intramural Research Program of NHLBI