We recently described a xenograft model of chronic lymphocytic leukemia (CLL) using NOD/SCID/γc^null (NSG) mice. Adoptive transfer of primary patient PBMCs into these mice results in engraftment and proliferation of CLL cells if autologous activated T cells are present. To date, such CLL-derived T cell proliferation has been achieved by co-transfer of third party antigen presenting cells (APCs). Unfortunately, in this setting, mice succumb to graft-versus-host disease, due to complex interactions between CLL cells, alloAPCs, T cells, and the xenogeneic host. We hypothesized that alternative strategies of autologous T cell activation might refine the model, ultimately providing longer CLL cell engraftment and animal survival.

We describe 3 approaches to achieving engraftment and activation of CLL-derived T cells that support B cell growth in vivo. First, we activated CLL CD3⁺ cells isolated from PBMCs of 4 patients with anti-CD3/28 beads for 72 hours in vitro. Cells were then mixed with CFSE-labeled PBMCs from the same patient at varying ratios (1:50 to 1:1000 CD3⁺ cells:CLL PBMCs) and injected intraorbitally (io) into a total of 17 mice. CD4, CD8 and CD19 cell engraftment, identified by a human CD45 lymphocyte gate (hCD45), and proliferation, assessed by CFSE dilution of labeled cells, were monitored weekly. All mice demonstrated detectable CD3⁺ and CD5⁺CD19⁺ cells from week (wk) 1. The percent (%) CD3⁺ cells, as a proportion of hCD45, increased weekly in all mice receiving anti-CD3/28-activated cells. While overall % CD5⁺CD19⁺ cells decreased weekly, the % proliferating increased and strongly correlated with increasing % of T cells (r²=0.7799, p<0.0001, 45 evaluable data pairs). At the time of reporting (up to 5 wks follow up), 4 animals have died. Death correlated with high % circulating CD8⁺ (mean 57.7% vs. 14.2% prior to death, p<0.0001), but not CD4⁺ cells (mean 31.0% vs 36.3%, p=n/s). Spleen immunohistochemistry analyses revealed follicles containing CD20⁺ CLL cells, based on L chain restriction and RT-PCR for the leukemic IGHV-D-J. Follicles were infiltrated with both CD4⁺ and CD8⁺ cells. These findings were the same as found with our published model.

Our second approach involved engraftment of solely CLL-derived peripheral blood T cells, without associated autologous PBMCs or alloAPCs. Positively selected CD3⁺ populations from 2 patients were activated with anti-CD3/28 beads, expanded in vitro, and 5–10×10⁶ such cells injected into mice io. Cells from both patients yielded engraftment of CD4⁺ and CD8⁺ cells by wk 2. We then utilized two mice engrafted with cells from one patient, 1 with 70% and 1 with 30% CD4⁺ cells to compare the effect of CD4⁺ T cells in CLL cell growth. Injection of CFSE-labeled PBMCs into the mouse with high % of CD4 cells showed a clear circulating CFSE⁺CD5⁺CD19⁺ population at wk 1. By wk 2, 95% of the leukemic B cells had divided with 60% showing ≥6 divisions. In the mouse with low % CD4, CD5⁺CD19⁺ cell engraftment was suboptimal, never being more than 100 detectable events in analysis from any single bleed. No clear CFSE dilution pattern was apparent.

Finally, we have assessed the development of CLL-derived T cells from autologous bone marrow CD34⁺ cells of 4 CLL cases. 48hr old neonatal NSG mice received CD34⁺ cells (≤1×10⁶) intrahepatically. In 1 of the 4 animals, CD19⁺ cells emerged at 2 months but by month 4 the predominant hCD45⁺ population was CD3⁺ (60% of total hCD45). Subsequent io injection of CFSE-labeled PBMCs from the same subject demonstrated engraftment and immediate proliferation (detected by wk1 following injection) of …