Neutrophils in chronic lymphocytic leukemia are permanently activated and have functional defects

Oncotarget84889 www.impactjournals.com/oncotarget www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 49), pp: 84889-84901 Neutrophils in chronic lymphocytic leukemia are permanently activated and have functional defects Gayane Manukyan1,4, Tomas Papajik2, Petr Gajdos3, Zuzana Mikulkova1, Renata Urbanova2, Gabriela Gabcova1, Milos Kudelka3, Peter Turcsányi2, Pavlina Ryznerova2, Vit Prochazka2 and Eva Kriegova1 1 Department of Immunology, Faculty of Medicine and Dentistry, Palacky University Olomouc, Olomouc, Czech Republic 2 Department of Hemato-Oncology, Faculty of Medicine and Dentistry, Palacky University and University Hospital, Olomouc, Czech Republic 3 Department of Computer Science, Faculty of Electrical Engineering and Computer Science, VSB-Technical University of Ostrava, Ostrava, Czech Republic 4 Laboratory of Molecular and Cellular Immunology, Institute of Molecular Biology NAS RA, Yerevan, Armenia Correspondence to: Eva Kriegova, email: [email protected] Keywords: chronic lymphocytic leukemia, neutrophils, disease activity, flow cytometry, enhanced ROS production Received: May 23, 2017 Accepted: July 25, 2017 Published: August 08, 2017 Copyright: Manukyan et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT A growing body of studies highlights involvement of neutrophils in cancer development and progression. Our aim was to assess the phenotypic and functional properties of circulating neutrophils from patients with chronic lymphocytic leukemia (CLL). The percentage of CD54+ and CD64+ neutrophils as well as CD54 expression on these cells were higher in CLL patients than in age-matched healthy controls. Neutrophils from CLL produced more reactive oxygen species (ROS) compared to controls in both resting and activated conditions. Lipopolysaccharide-induced production of IL-1β and TNF-a as well as reduced TLR2 expression in neutrophils from CLL than in neutrophils from controls suggesting their tolerant state. Finally, phenotypic alterations of neutrophils, particularly elevation of CD64 and CD54 markers, correlated with disease activity and treatment, and low percentage of neutrophils. Taken together, the alterations in percentage and functional characteristics of neutrophils reflect the clinical course of CLL. Our data provide first evidence that neutrophils in CLL are permanently primed and have functional defects. INTRODUCTION Chronic lymphocytic leukemia (CLL) is characterized by an accumulation of abnormal B lymphocytes in the bone marrow, lymphoid tissues and peripheral blood. The expansion of malignant clone(s) may in turn lead to phenotypic alterations of residual hematopoietic bone marrow cells such as those maturing neutrophils, monocytic and/or erythroid cells as well as CD34+ myeloid precursors; and this may be also reflected in circulating cells as shown already in myeloid malignancies [1, 2]. Whether such phenotypic alterations are also present in circulating neutrophils in CLL patients remains to be investigated. Polymorphonuclear neutrophils are the most abundant phagocytes in the circulation and are traditionally associated with innate defense against infection [3, 4]. When activated, they release a spectrum of inflammatory mediators, produce reactive oxygen species (ROS), activate complement and regulate inflammation [5]. Besides, neutrophils mediate antibody-dependent cell-mediated cytotoxicity (ADCC) [6, 7]. A growing body of studies shows that their inappropriate activation is associated with inflammatory, autoimmune processes as well as cancer development and progression [8]. The circulating neutrophils from CLL patients were shown to possess impaired bactericidal activity [9], probably due to myeloperoxidase deficiency and impaired migratory abilities [10]. More recently, neutrophils were shown to differentiate toward a B-cell helper phenotype in murine model of CLL contributing to the formation of survival niches for CLL cells in lymph nodes [11]. Nevertheless, neutrophils in CLL may contribute to treatment response as shown by phagocytosis of about Research Paper Oncotarget84890 www.impactjournals.com/oncotarget 50% of anti-CD20 (obinutuzumab, but not rituximab) opsonized CLL targets by drug-activated neutrophils [12]. In an attempt to further understand how neutrophils may contribute to CLL pathogenesis, we aimed to assess the phenotypic and functional properties of circulating neutrophils from patients with CLL. We investigated expression profiles of major membrane-bound markers associated with cell activation: i) surface adhesion molecules associated with inflammation (CD11b, CD62L, CD54), and ii) Fcγ receptor: FcγRI (CD64) using flow cytometry. Further, we determined the priming ability as well as baseline and stimulated generation of ROS by the neutrophils obtained from CLL patients. RESULTS Surface molecule expression on neutrophils In order to characterize the circulating neutrophils in CLL, we investigated their immunophenotypes using following surface molecules CD54, CD11b, CD62L, and CD64 and compared with those on neutrophils from agematched healthy subjects. Neutrophils were selected in the forward versus side scatter dot plot, and additionally gated as CD15/CD16 positive cells (Supplementary Figure 1 in the on-line Supplement). Figure 1: Distribution of percentage of surface markers (CD64, CD54, CD62L, CD11b) on circulating neutrophils from healthy controls and patients with CLL. Group means are indicated by horizontal bars, error bars indicate 95%CI; P values for differences between two groups are stated. Oncotarget84891 www.impactjournals.com/oncotarget Expression of marker CD54 was increased (P < 0.001) and higher percentage (P < 0.01) of CD54+ neutrophils was observed in CLL patients compared to healthy controls (Table 1, Figure 1, Supplementary Figure 2 in the on-line Supplement). When controls and CLL were compared, the percentage of CD64+ cells was 4-fold higher (P < 0.001) in CLL patients showing marked interindividual variability ranging from 1.2 to 93.3%. The expression (MFI) of CD64 did not differ between CLL vs controls (P = 0.71). Lower percentage of CD62L+ cells (P < 0.001) as well as lower expression of CD62L (P = 0.02) was found in neutrophils from CLL patients comparing to age-matched healthy controls (Table 1, Figure 1, Supplementary Figure 2 in on-line Supplement). Expression of CD11b on neutrophils did not vary between CLL patients and controls (P > 0.05). The comparison of expression of studied surface markers on neutrophils from healthy subjects and CLL patient subgroups according to the treatment history is shown in Supplementary Table 1 in on-line Supplement. Baseline expression of cell surface molecules on neutrophils in CLL subgroups To determine whether disease activity, treatment, IGHV mutational status, Binet stage, blood number counts, and percentage of CLL cells affect the expression level of investigated markers on neutrophils, we compared the expression of CD54, CD11b, CD62L, and CD64 in CLL subgroups. Active disease was associated with upregulated expression of CD54 and CD64 (P < 0.01 and P < 0.01, respectively), increased percentage of CD64 (P < 0.01) and downregulated expression (P < 0.05) and percentage (P < 0.01) of CD62L (Figure 2A, Figure 3A, Supplementary Figure 3 in on-line Supplement). Treated CLL patients displayed simultaneously increased expression of CD64 (P < 0.001), higher percentage of CD64 (P < 0.05) and CD54 (P < 0.01) compared to untreated patients. Remarkable differences were observed for the percentage of CD64 cells, which were almost 4-fold higher in treated vs untreated patients (Figure 2B, Figure 3B, Supplementary Figure 3 in on-line Supplement). Figure 2: Distribution of CD64, CD54 and CD62L positive neutrophils in CLL subgroups: A. non-active vs active disease, B. untreated vs treated disease, C. mutated vs unmutated IGHV gene status, D. CLL patients without infection vs with ongoing infection. Group means are indicated by horizontal bars, error bars indicate 95%CI; P values for differences between two groups are stated. Oncotarget84892 www.impactjournals.com/oncotarget When patient subgroups were compared according to Binet stage, the percentage of CD64+ cells was lower in patients with Binet stage A comparing to stages B and C (P < 0.05) (data not shown). In patients with ongoing infection, lower percentage of CD11b+, CD62L+ cells and expression of CD62L were detected comparing to those with no infection (Figure 2D, Supplementary Figure 3 in on-line supplement). Patients with unmutated IGHV gene status had decreased percentage of CD62L (P < 0.001) and a decreased MFI of CD62L (P < 0.05) compared with mutated IGHV status (Figure 2C, Supplementary Figure 3 in on-line Supplement). In our study, the percentage of circulating neutrophils in CLL showed a negative correlation with CD64 percentage and MFI (r = 0.45, P < 0.01 and r = 0.24, P = 0.07, respectively) and CD54 percentage and MFI (r = 0.34, P < 0.05 and r = 0.43, P < 0.01, respectively). In opposite, neutrophil percentage positively correlated with CD62L (%) and MFI (r = 0.41, P < 0.01 and r = 0.46, P < 0.01, respectively) (Supplementary Figure 4 in on-line Supplement). Figure 3: Percentage of circulating neutrophils and surface markers (CD64, CD54, CD62L, CD11b) in CLL subgroups. Comparison of studied parameters (medians) between A. non-active (light red columns) vs active (dark red columns) disease, B. untreated (light blue columns) vs treated (dark blue columns) disease, C. mutated (light violet columns) vs unmutated (dark violet columns) IGHV gene status. The arrows indicate the increase/decrease in studied parameters in non-active/untreated/mutated CLL subgroups compared to active/treated/unmutated subgroups, respectively. Oncotarget84893 www.impactjournals.com/oncotarget Cellular pattern of peripheral blood in CLL patients Further, we compared percentage of circulating neutrophils, lymphocytes and CLL cells (as assessed by CD5+/CD19+) in patient subgroups according to the disease activity, treatment history and IGHV gene mutational status. Majority of enrolled CLL patients had decreased percentage of circulating neutrophils compared to healthy subjects (Table 1). The absolute number of neutrophils in CLL patients did not differ from those in healthy controls (Table 1), only 3 patients had absolute neutropenia ( < 2.0 x 109/L). Active disease was associated with a increased of neutrophil percentage (P < 0.001) and increased percentage of lymphocytes (P < 0.01) and CLL cells (P < 0.01) comparing to non-active disease. Patients with unmutated IGHV gene status had lower percentage of neutrophils than those with mutated IGHV (P < 0.05), the difference in neutrophils between treated and untreated patients did not reach significance (Table 1, Supplementary Figure 5 in the on-line Supplement). Characteristics of neutrophils associated with CLL and its subgroups To better characterize the neutrophils in CLL, we determined which markers discriminate i) CLL and healthy controls and ii) CLL subgroups according to the disease activity, treatment history and IGHV gene mutational status. We calculated combinations of every two and three markers (percentage and MFI of studied surface markers, percentages of neutrophils, lymphocytes and CLL cells) from individual patients and determined the misclassification error for a particular subgroup. Figure 4: Comparison of percentage of circulating neutrophils and surface markers (CD64, CD54, CD62L, CD11b) in CLL patients and healthy controls. CLL is coloured yellow, control subjects green. Discrimination between CLL and controls using combination of two parameters: A. CD54 (%) - neutrophils (%) and B. CD64 (%) - neutrophils (%). The dots represent the percentage of CD54/CD64 (y-axis) and neutrophils (x-axis) in individual subjects. C. Comparison of studied parameters (medians) between CLL (yellow columns) and healthy groups (green columns). The arrows indicate the increase/decrease in studied parameters in CLL patients compared to healthy controls. Oncotarget84894 www.impactjournals.com/oncotarget Table 1: Relative and absolute neutrophil counts and expression levels of surface markers on neutrophils in: A. healthy controls vs CLL, B. non-active CLL vs active CLL, C. untreated CLL vs treated CLL, D. mutated IGHV gene status vs unmutated IGHV gene status. Oncotarget84895 www.impactjournals.com/oncotarget The best separation of healthy controls and CLL using two markers was achieved for combination of CD64 (%) - neutrophil (%) and CD54 (%) - neutrophil (%), thus further confirming observation that lower percentage of neutrophils is associated with higher percentage of CD64 and CD54 markers (Figure 4). Similarly, combination of CD64 (%) - CD54 (%) - neutrophil (%) showed good separation between studied CLL subgroups, achieving a classification error less than 15 % (Figure 5). Figure 5: Multidimensional clustering of CLL subgroups. Discriminant 3-D models with combination of CD64 (%) - CD54 (%) - neutrophils (%) show the separation of following CLL subgroups: A. non-active (light red clouds) vs active (dark red clouds) disease, B. untreated (light blue clouds) vs treated (dark blue clouds) disease, C. mutated (light violet clouds) vs unmutated (dark violet clouds) IGHV gene status. The dots represent the percentage of CD54 (x-axis), CD64 (y-axis) and neutrophils (z-axis) of individual subjects. The colour bulks represent the probability intervals: the more saturated color the higher probability (intervals: > 90, 90-80, 80-70, 70-60, and 60-50%) of correct classification. ANC: Absolute neutrophil count. The normal range for the ANC = 1.5 to 8.0x109/L. #FC (Fold Change) between group medians *CD54 available in 39 patients and 34 controls Oncotarget84896 www.impactjournals.com/oncotarget Influence of fMLP and PMA on ROS production by neutrophils Having identified increased expression of proinflammatory surface antigens by neutrophils from the patients, we next sought to investigate the basal and induced ROS generation in neutrophils ex vivo. For this purpose, we induced whole blood cells with fMLP (weak chemotactic stimulus) and PMA (non-receptor-dependent activator of protein kinase C) and measured conversion of DHR-123. The fraction of neutrophils with spontaneous (resting burst) ROS production was increased in patients with CLL compared to controls (mean MFI; 34.9 vs 24.2, P < 0.05). The average increase in ROS response to fMLP was detected in CLL group compared to basal secretion level (42.4 vs 34.9, P < 0.01) and compared to healthy control group (42.4 vs 28.7, P < 0.05). Despite PMA markedly induced ROS production by neutrophils in both healthy and CLL groups (P < 0.01 and P < 0.001, respectively), stimulation with PMA has led to a maximum ROS-release in neutrophils from CLL patients that was 4-fold greater than those in healthy controls (1125 vs 281, P < 0.01) (Figure 6A). Surface marker expression on neutrophils after LPS priming To determine whether the observed phenotypical changes of neutrophils in CLL are associated with altered functional responses of the cells, we primed circulating neutrophils with LPS and measured expression levels of CD54, CD11b, CD62L, CD64, TLR2 and TLR4 using flow cytometry. Phenotypic changes induced by LPS ex vivo were selective. LPS upregulated expression of CD11b and TLR2 in healthy cells, and CD11b in CLL group (P < 0.05). In opposite, TLR2 expression after LPS exposure was downregulated in neutrophils from CLL patients (Supplementary Figure 6 in on-line Supplement). The expression of other markers was not influenced by LPS exposure (P > 0.05). Table 2: Patient characteristics. *For more information regarding the treatment in previously treated CLL patients see Supplementary Table 2 in on-line Supplement. Oncotarget84897 www.impactjournals.com/oncotarget IL-1β and TNF-a production in cell culture supernatants after LPS priming and TNF-a levels in serum We further studied the impact of bacterial LPS on IL-1β and TNF-a release by cultured neutrophils from patients with CLL and healthy subjects. Additionally, we investigated the protein levels of TNF-a in sera of CLL patients and healthy controls. LPS stimulation of isolated neutrophils from control group induced production of both pro-inflammatory cytokines (P < 0.05). In contrast, neutrophils from CLL patients failed to induce either IL-1β or TNF-a release from LPS-stimulated cells. Moreover, LPS exposure decreased production of IL-1β by neutrophils isolated from CLL patients (P < 0.05). Both LPS-induced production of IL-1β and TNF-a in CLL group was lower than those released by neutrophils from healthy group (P < 0.05) (Figure 6B, 6C). When comparing the serum levels of TNF-a, higher serum level was observed in CLL patients than in healthy controls (P < 0.001) (Supplementary Figure 7 in on-line Supplement). DISCUSSION A growing body of evidences emphasizes the importance of the complex relationship between the CLL clone(s) and its immune environment. The contribution of neutrophils, effectors in both innate and adaptive immunoregulatory networks, to the pathogenesis of CLL was not established yet. To gain more insights on the phenotype of circulating neutrophils in CLL, we studied expression of membrane-bound markers associated with cell activation (CD54, CD11b, CD62L, CD64). Our data demonstrates for the first time that patients with CLL have increased percentages of circulating CD54+ and CD64+ neutrophils and increased CD54 expression. It is tempting to speculate that activated phenotype of neutrophils in CLL, irrespective of infection, is a result of the presence of systemic inflammatory cytokine milieu [13, 14] which largely promotes pro-survival and proliferation signaling in tumor cells [15]. Our data on increased serum concentrations of TNF-a provide further support for the concept of a systemic inflammation in CLL. Importantly, we detected an increase in the ROSgenerating activity in neutrophils from CLL patients suggesting their functionality. CLL patient neutrophils produced markedly more ROS in both resting condition as well as after stimulation with fMLP and PMA ex vivo than healthy neutrophils. The increased oxidative potential of these cells could be explained by their altered energy metabolism or chronic activation of the immune system, thus deserving further investigations. Generally circulating neutrophils are a major source of ROS in human blood, having main involvement in microbial defense. While keeping the same absolute number of circulating neutrophils in CLL as in healthy controls, the contribution of increased ROS-generating activity in neutrophils from CLL patients may have significant effects on the pathophysiology of CLL far beyond these cells. In line with our observation, there is already evidence about a robust extracellular release of ROS from neutrophils in CLL exposed to anti-CD20 treatment or αCD20-opsonized CLL cells [16]. The same authors proved that the enhanced ROS production by neutrophils and monocytes in CLL may limit the NK cell-mediated ADCC against CLL cells during anti-CD20 treatment, as NK cell ADCC could be partially restored by anti-oxidative agents [17]. Enhanced ROS production from neutrophils may also participate in tumorigenesis [18], induce mutations and genotoxicity generally [19], contribute to drug resistance and aggressive disease course [20] as well as to a systemic Tand NK-cell dysfunction [21]. More studies are warranted to investigate whether ROS released from activated Figure 6: Functional analyses of circulating neutrophils. A. Basal (rest) and fMLPand PMA-induced ROS production by neutrophils from the blood of CLL patients (CLL, n = 18) and healthy controls (control, n = 17). Error bars represent means ± SEM. Induced (*P < 0.05, **P < 0.01, ***P < 0.001) vs resting ROS production; (#P < 0.05, ##P < 0.01) in CLL vs control group. B. IL-1β and C.TNF-a release by neutrophils isolated from the blood of CLL patients (CLL, n = 16) and healthy controls (control, n = 7) after priming with LPS. Error bars are means ± SEM (*P < 0.05, **P < 0.01) vs media; (#P < 0.05, ##P < 0.01) vs control group.