PD-1 expression contributes to functional impairment of NK cells in patients with B-CLL

B cell chronic lymphocytic leukemia (B-CLL) is a common leukemia subtype [1] and compromised immune function is a major clinical problem leading to increased mortality among patients [2]. Natural killer (NK) cells recognize and kill transformed or virally infected cells through mechanisms including granule-mediated cell lysis, cytokine release

The immune checkpoint co-expression pattern of PD-1, CTLA-4, LAG-3, TIGIT and TIM-3 was further examined using Boolean gating.The expression of a range of checkpoint combinations was increased in the patient group (Fig. 1B).In particular, the proportions of PD-1/TIGIT; CTLA4/LAG3; PD-1/CTLA4; PD-1/LAG3; PD-1/TIGIT/LAG3; PD-1/CTLA4/TIGIT; and PD-1/CTLA4/TIGIT/LAG3 co-expressing NK cells were all increased in the patient group.Co-expression of the strong inhibitory checkpoints PD-1 and TIGIT was noteworthy as a phenotype not observed in healthy donors but comprising over 2% of the NK cell pool in many patients.Furthermore, it is notable that PD-1 expression is a component in 6 of the 7 checkpoint combinations that are increased on NK cells in patients, implying a potentially important central role for PD-1 in the checkpoint regulation of NK cells in CLL.
Given the importance of PD-1 in combinatorial checkpoint expression we next focused our studies on the phenotypic and functional features of PD-1 + NK cells (PD-1 pos ).The majority of PD-1 pos NK cells were present within the cytotoxic CD56 dim subset, the dominant population within blood and a more differentiated pool compared to CD56 bri NK cells.In particular, within a subset of 10 patients where PD-1 pos cells comprised 8.9% of the total NK pool, PD-1 was expressed on 9.1% of the CD56 dim population compared to only 0.76% of CD56 bri cells (Fig. 2A).
scRNA-Seq was used to compare the transcriptome of FACSsorted PD-1 pos and PD-1 neg NK cells but revealed relatively subtle differences between populations (Supplementary Fig. 3).Transcription of PDCD1, the gene encoding PD-1, was observed in the PD-1 pos subset, indicating that trogocytosis is not the sole determinant of protein expression [15].Of note, the transcriptomes of PD-1 pos and PD-1 neg NK cells did not separate into two distinct clusters, implying that PD-1 expression is not a dominant transcriptional NK lineage marker.Enrichment for genes associated with cytotoxicity and immune synapse formation may reflect high levels of ongoing activation.
In conclusion, the expression of a range of inhibitory checkpoint proteins is increased on NK cells in patients with CLL and PD-1 is a common shared factor in many cases.PD-1 pos cells have impaired functional activity but this may be partially improved by inhibition of ligand engagement.Enhancement of NK cell function may therefore contribute to the clinical utility of antibody mediated PD-1 checkpoint blockade in patients with B-CLL.
Fig. 2 PD-1 NK cells express lower levels of activating receptors and reduced functionality, which can be partially reversed.A The percentages of PD-1 pos NK cells were compared between the CD56 bri and CD56 dim NK populations.B Dot plots to compare the percentages of CD57 pos and NKG2A pos NK cells between PD-1 pos NK cells (red circles) and PD-1 neg NK cells (blue squares).C Dot plots to compare the percentages of DNAM-1 pos, NKp30 pos , NKp46 pos , and NKG2D pos NK cells between PD-1 pos NK cells (red circles) and PD-1 neg NK cells (blue squares).Statistical significance was determined by Wilcoxon matched-paired nonparametric test using GraphPad Prism (*p < 0.05, **p < 0.01, ***p < 0.001).D Bar charts summarizing IFN-γ and TNF-α production by PD-1 pos (red circles) vs. PD-1 neg (blue squares) NK cells from B-CLL patients after stimulation with 721.221 (n = 10).E Bar charts summarizing IFN-γ production by PD-1 pos NK cells (red) and PD-1 neg NK cells (blue) before and after PDL-1/L-2 blockade.Each dot represents a single donor.The differences were determined by the Mann-Whitney test using GraphPad Prism (*p < 0.05, **p < 0.01).F Bar chart to compare CD107a expression after stimulation with 721.221 by PD-1 pos NK cells (red) and PD-1 neg NK cells (blue) before and after PDL-1/L-2 blockade.Data are shown as the percentage of CD107a-expressing NK cell subsets (n = 10).G Summary of CD107a expression in PD-1 pos NK-92 (white bars) and PD-1 neg NK-92 (black bars) cells after stimulation with 721.221 before and after PDL-1/L-2 blockade.H Bar charts to compare the cytotoxicity capacity of PD-1 pos NK-92 (white bars) and PD-1 neg NK-92 (black bars) against 721.221target cells with and without PDL-1/L-2 blockade.Data were analyzed using Mann-Whitney and Wilcoxon matchedpaired nonparametric tests (*p < 0.05, **p < 0.01, ***p < 0.001).

Fig. 1
Fig. 1 Expression pattern of immune checkpoint receptors on NK cells from B-CLL patients.A Surface expression of PD-1, CTLA-4), LAG-3, TIM-3), TIGIT, NKG2A, CD96 and Siglec-7 on NK cells from CLL patients was compared to age-matched healthy controls (HD).Flow plots to demonstrate representative data from one B-CLL patient and dot graphs are the comparisons between B-CLL patients and HD.B Coexpression of PD-1, CTLA-4, LAG-3, TIM-3 and TIGIT on NK cells was analyzed using Boolean gating.Percentages were compared between B-CLL and agematched healthy controls (HD) for double, triple and four and five checkpoint receptor coexpressed NK cell populations.Statistical analysis was performed using the Mann-Whitney nonparametric test (*p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001).