Abstract

Abstract 3875

Introduction:

Immunologic environment influences progression of lymphoid malignancies. Specifically, shifts in subsets of natural killer (NK) and T cells as well as tumor expression of inhibitory ligands may contribute to ability to evade host detection. Immune dysfunction may be particularly important in CLL/SLL, as prevalent circulating tumor cells engage in persistent, widespread interactions with immune cells; commonly-used mAb therapies (e.g. rituximab, alemtuzumab) rely upon ADCC mediated by NK cells and other innate effectors; and disease course is highly variable and not fully accounted for by tumor-intrinsic prognostic factors. Therefore, to better characterize the immune system in CLL/SLL, we prospectively assessed NK and T cell frequency, phenotype, and function in a series of CLL/SLL patients.

Methods:

Serial blood samples (up to 3 samples each, 3–6 months apart) were collected from 31 untreated CLL/SLL patients (median age 66) and 15 healthy age-matched controls (HC), and peripheral blood lymphocytes (PBL) analyzed directly ex vivo by multiparameter flow cytometry (160 distinct parameters evaluated, primarily on T and NK cells). NK cell-mediated natural and antibody-dependent cytotoxicity were also assessed by CD107a degranulation assay following PBL co-culture with rituximab, 721.221 EBV-transformed lymphoma cells, or both. Differences in parameters between patients and controls, or between progressors and non-progressors [categorized based on updated NCI-WG criteria (Blood 2008;111:5446)] were analyzed by Wilcoxon rank-sum test. All subjects signed IRB approved informed consent forms.

Results:

CLL/SLL VS. HC: CLL/SLL samples displayed a marked decrease in the ability of the cytolytic CD56dim NK cells to degranulate in response to tumor, both with or without rituximab (Table 1). CD56dim NK cells from CLL/SLL patients also displayed a more immature phenotype (↓CD57, ↓NKG2D, ↑CD27, ↓KIR) than those from HC, suggesting either a block in differentiation or elimination of the most-differentiated cells. NK cell expression of NKp44, CD69, CD62L, CD137, granzyme B, perforin, or PD-1, as well as tumor-induced NK cell production of IFNγ, did not differ. CLL/SLL patients had increased total T cells with a decreased CD4:CD8 ratio, associated with increased total number of CD8 T cells, greater activation of naive CD4 T cells and transition to a memory phenotype. Treg (CD4+CD25+FoxP3+) frequency was significantly higher in CLL/SLL patients (4.5% vs. 1.8% of CD4 T cells, p=0.005), as was PD-1 expression on both CD4 and CD8 T cells, while CD137 and ICOS expression was similar in both groups. PROGRESSORS VS. NON-PROGRESSORS: With median follow-up of 16.5 months (range 1–37), 7 of 31 patients have met criteria for progression. Compared to non-progressors, progressors showed changes in the CD56bright NK cell compartment suggestive of increased activation and accelerated differentiation, with increased expression of CD69, granzyme B, perforin, CD16, and KIR. However, no significant functional differences in NK cells, or consistent differences in T cell subsets, have been observed to date.

Conclusions:

CLL/SLL patients have a shift toward less mature NK cells, associated with deficits in NK cell degranulation against tumor targets, compared with healthy donors. Those CLL/SLL patients who progressed had greater CD56 bright NK cell phenotypic aberrancies than non-progressors, though these findings require confirmation with a larger cohort. Taken together, our findings support the hypothesis that immune dysfunction in CLL/SLL may be due in part to a block in NK cell differentiation or loss of more mature cells, and current studies are exploring these possibilities and potential mechanisms. Given these findings, along with the immunosuppressive changes observed in the T cell compartment (↑Tregs, ↑PD-1), these data support therapeutic strategies in CLL/SLL aimed at augmenting NK and/or T cell function.

Table 1.

Parameters Differentially Expressed in cytolytic CD56 dimNK cells

ParametersCLL/SLL (median)HC (median)P-Value
%CD107a+ (PBL + tumor) 14.9% 37.4% 0.001 
%CD107a+ (PBL + rituximab + tumor) 37.5% 62% 0.0008 
CD57 MFI 4937 20,640 0.03 
CD27 MFI 368 190 0.04 
NKG2D MFI 4187 5378 0.006 
%KIR2DS4+ 0.3 10.2 0.008 
%KIR2DS1+ 8.0 16.8 0.02 
%KIR3DL1+ 7.9 19.9 0.008 
ParametersCLL/SLL (median)HC (median)P-Value
%CD107a+ (PBL + tumor) 14.9% 37.4% 0.001 
%CD107a+ (PBL + rituximab + tumor) 37.5% 62% 0.0008 
CD57 MFI 4937 20,640 0.03 
CD27 MFI 368 190 0.04 
NKG2D MFI 4187 5378 0.006 
%KIR2DS4+ 0.3 10.2 0.008 
%KIR2DS1+ 8.0 16.8 0.02 
%KIR3DL1+ 7.9 19.9 0.008 
Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.