The lack of BTK does not impair monocytes and polymorphonuclear cells functions in X-linked agammaglobulinemia under treatment with intravenous immunoglobulin replacement

The lack of BTK in X-linked agammaglobulinemia (XLA) patients does not affect monocytes and polymorphonuclear cells (PMN) phenotype and functions. In this study, we show that XLA patients had an increased frequency of the intermediate monocytes subset and that BTK-deficient monocytes and PMN had a normal expression of receptors involved in the activation and cellular responses. We demonstrate that BTK is not required for migration, phagocytosis and the production of reactive oxygen species (ROS) following engagement of FC gamma receptors (FcγR). XLA monocytes and PMN showed an efficient calcium (Ca2+)-independent activation of oxidative burst, suggesting that oxidative burst is less dependent by Ca2+ mobilization. The phagocytosis was functional and it remained unaltered also after Ca2+ chelation, confirming the independence of phagocytosis on Ca2+ mobilization. Intravenous immunoglobulin (IVIg) infusion exerted an anti-inflammatory effect by reducing the frequency of pro-inflammatory monocytes. In monocytes, the IVIg reduce the oxidative burst and phagocytosis even if these functions remained efficient.


Introduction
X-linked agammaglobulinemia (XLA) is a symptomatic primary antibody deficiency (PAD) caused by mutations in the Bruton's tyrosine kinase (BTK) gene located on the long arm of Xchromosome encoding the cytoplasmic BTK [1,2]. These mutations lead to the nearly complete arrest of B cell differentiation in the bone marrow at the pre-B cell stage resulting in absent or very low peripheral B cells, agammaglobulinemia, absence of B cell-dependent lymphoid tissues and inability to mount specific antibody responses [3][4][5]. BTK is involved in regulation of proliferation, differentiation and survival of B cells and in functions of innate 20-60 years; mean age: 36.7 ± 15.4 years) and ten age-matched male healthy donors (HD) (age range of 27-58 years; mean age: 39.6 ± 9.8 years) were enrolled for the study. All XLA patients were on replacement treatment, with a cumulative monthly dosage of 400-600 mg/kg of IVIg administered every three weeks (S1 Table). The infusion time ranged from 2 to 3 hours. The infusion speed was established according to the individual tolerability. This study was approved by the Ethics Committee of the Sapienza University of Rome. All participants gave written informed consent prior to inclusion.

Blood samples preparation
Heparinized whole blood samples were collected from 10 HD and 6 XLA patients immediately before and one hour after IVIg administration. These time points were chosen based on our previous observations [23,38], taking into consideration that the highest increase of several cytokines plasma concentration occurs within one hour after IVIg infusion [39]. Total peripheral blood monocytes and neutrophils count were determined from blood cell counts and white blood cell differentials. For evaluating circulating cell without the harm of cell loss related to the density gradient centrifugation procedure, peripheral red blood cells was lysed using lysing buffer (Becton Dickinson, BD). Samples were washed twice before staining with various combinations of fluorochrome-labeled antibodies. All antibodies were obtained from BD Biosciences. Flow cytometric analysis was done with a FACSCalibur instrument (BD) using CellQuest (BD) and FlowJo (TreeStar, Ashland, Ore) software. The cytometer stability and sensitivity were checked before each acquisition session by using microbeads designed to control the efficiency, the coefficient of variation of scatter and fluorescence signals and the time delay calibration (Nile Red Fluorescent particles and Calibrite APC Beads, all from BD). Results were expressed as geometric mean Fluorescence Intensity (MFI) of any given marker within the defined population. 30.000 events were counted per sample.

Phenotypic analysis of monocyte subpopulations
Whole blood samples were first treated to lyse red blood cells and then washed twice and stained at 4˚C for 30 min with combinations of fluorochrome-labeled antibodies. Samples were washed, suspended in ice cold PBS and analyzed by a 4-color flow cytometry single platform. Monocytes subpopulations were phenotypically selected by gating on CD14 + HLA-DR + monocytes and then classified according to their expression of CD14 and CD16 into classical (CD14 ++ CD16 -), intermediate (CD14 ++ CD16 + ) and non classical monocytes (CD14 + CD16 ++ ). An isotype control (IgG 1 , BD) with the same fluorochrome of CD16 antibody was run in parallel in order to set the boundary between classical and intermediate monocytes [23]. The surface expression of CD181, CD11b, CD11c and Siglec 9 receptors was evaluated on monocytes samples from erythrocytes-lysed whole blood, performing a staining at 4˚C for 30 min with combinations of fluorochrome-labeled antibodies. In parallel, we used an isotype control (IgG 1 , BD) for each receptor analyzed. Results were expressed as percentage of cells that stained positive for a given marker.

Monocytes and PMN stimulation by Escherichia coli
One hundred μL of whole blood were added to 20 μl of pre-cooled opsonized not labeled whole Escherichia coli (E. coli) at a concentration of 1-2x10 9 /ml (Glycotope, Biotechnology). Samples were incubated in water bath for 20 min at 37˚C and erythrocytes were lysed for 15 min at room temperature. Cells stimulated were stained at 4˚C for 30 min with fluorochromelabeled antibodies in various combinations to evaluate the receptors expression.

Monocytes and PMN oxidative burst activity
The leukocyte oxidative burst was analyzed by using the PHAGOBURST assay (Glycotope, Biotechnology). Whole blood samples (100 μl) were incubated in a water bath for 20 min at 37˚C with opsonized E. coli (1-2x10 9 /ml), PMA (1.62 mM) and wash solution as negative control. The percentage and MFI of cells that produced ROS was analyzed. The intracellular production of superoxide anions and hydrogen peroxide in monocytes and PMN in response to phagocytosis of bacteria was tested by using the fluorescence probes dihydrorhodamine 123 (DHR 123). We repeated the analysis of the oxidative burst using non-opsonized E. coli (1.5x10 9 /ml) as control in order to exclude the FcγR-independent phagocytosis. In parallel, we performed the study of oxidative burst by using BAPTA-AM (Unimed Scientifica) as calcium chelator, in order to verify the Ca 2+ -dependence production of ROS triggered by FcγR stimulation. Whole blood samples were pre-treated with 100 μM of BAPTA-AM for 30 min and then they were incubated in a water bath for 20 min at 37˚C with opsonized E. coli (1-2x10 9 /ml). The oxidation lead to fluorescence detected by flow cytometry, by using the blue-green excitation light (488 nm argon-ion laser).

Monocytes and PMN phagocytosis
The leukocyte phagocytosis was determined by using a PHAGOTEST assay (Glycotope, Biotechnology). Whole blood samples (100 μl) were incubated in a water bath for 10 min at 37˚C with opsonized FITC-labeled E. coli (2x10 9 /ml). As negative controls, 100 μl of each sample were incubated in an ice bath for 10 min with opsonized FITC-labeled E. coli (2x10 9 /ml). As control, the same experiment was done by using non opsonized FITC-labeled E. coli (2x10 9 / ml). The percentage and MFI of granulocytes and monocytes having performed phagocytosis was analyzed. In parallel, we tested the effect of calcium chelator on phagocytosis in order to confirm the Ca 2+ -independence of the process. Whole blood samples were pre-treated with 100 μM of BAPTA-AM for 30 min and then incubated in a water bath for 10 min at 37˚C with FITC-labeled E. coli (2x10 9 /ml). Cells were analyzed by flow cytometry by using the blue-green excitation light (488 nm argon-ion laser).

PMN migration
PMN chemotaxis was determined by the MIGRATEST CHEMOTAXIS assay (Glycotope, Biotechnology). Heparinized whole blood samples (1000 μl) were added to leukocyte separation medium for isolation of leukocyte-rich plasma (LRP), for 40 min at room temperature. After isolation, 100 μl of LRP were added into cell culture insert (3.0 μm pore size) and incubated in a water bath for 30 min at 37˚C with chemotactic peptide N-formyl-Met-Leu-Phe (fLMP). After incubation, the cell suspensions were added in two separate tubes and placed on ice. 20 μl of cell suspension from the cell culture insert of the negative control were mixed with 180 μl of incubation buffer and then transferred on a third tube. 20 μl of antibody and counting reagent were added to LRP and incubated for 10 min in an ice bath in the dark. Finally, 20 μl of vital DNA staining solution were added per tube and incubated 5 min on ice. The number of migrated PMN was analyzed by flow cytometry by using the blue-green excitation light (488 nm argon-ion laser).

PMN elastase assay
The PMN elastase was quantified by Human PMN Elastase Platinum ELISA (Affymetrix eBioscience). 100 μl of each plasma sample were added to a microwell plate coated with polyclonal antibody to human PMN elastase and incubated at room temperature for 1 hour. After washing, samples were incubated with TMB Substrate Solution at room temperature for 20 min in the dark. 50 μl of stop solution were added to stop the reaction. The absorbance was determined at 450 nm on a microplate reader (Multiskan™ FC Microplate Photometer, Thermo Scientific). Concentrations were expressed as ng/ml.

IL-8 plasma dosage
Interleukin-8 (IL-8) was measured by using a commercially available kit ELISA (CliniSciences). Plasma was collected from blood samples unstimulated and stimulated with E. coli by centrifugation for 15 min at 1000 x g, diluted and tested in duplicates. 100 μl of each plasma sample were added to a microwell plate coated with monoclonal antibody to human IL-8 and incubated at 37˚C for 90 min. After washing, samples were incubated with Biotin-Conjugate Solution at 37˚C for 1 h. Then, each well was washed several times and incubated with 100 μl of Streptavidin-HRP at 37˚C for 30 min. After washing, 100 μl of Substrate Solution were added to each well and incubated for 10-20 min a 37˚C. 100 μl of stop solution were added to stop the reaction. The absorbance was determined at 450 nm on a microplate reader (Multiskan™ FC Microplate Photometer, Thermo Scientific). Concentrations were expressed as pg/ml.

Statistical analysis
Data were analyzed using the Mann-Whitney U test for unpaired two groups or by paired Wilcoxon test. Correlations were calculated by means of linear regression analysis.
Data were analyzed with StatView 5.0.1 software (SAS Institute, Cary, NC). A p value equal or less than 0.05 was considered as statistically significant.

Monocyte subsets
Monocytes subsets were identified according to their expression CD14 and the CD16 (Fig 1A  and 1B). Absolute counts of peripheral blood total monocytes, classical and non classical monocytes did not show any difference between XLA and HD. Classical and non classical monocyte were similarly represented in XLA and HD while the frequency of intermediate monocytes was increased in XLA ( Fig 1C). In Fig 1D we show the increase of monocytes subsets from XLA patients compared to HD expressed as percentage.
After IVIg, the frequency of intermediate monocytes rapidly decreased remaining higher than HD (Fig 1C and 1D) while the frequency of non classical and classical monocyte subsets did not change (Fig 1C).
patients and HD. Moreover, IVIg administration did not induce any variation of their expression (Fig 2A and 2D). The stimulation with opsonized E. coli induced an increase of all receptors similarly in XLA patients and HD (Table 1). This overexpression induced by E. coli was preserved after IVIg, showing that IVIg replacement did not affect the monocytes ability to respond to opsonized bacteria (Table 1).
PMN. The activation status of PMN strongly influence the respiratory burst, phagocytosis and adhesion processes. As shown in Fig 3A and 3F and in Table 2, PMN from XLA and HD showed an overlapping expression of all receptors with the exception of CD66b that was overexpressed in XLA patients ( Fig 3B). However, this increase was not correlated with a physiological dysfunction, as demonstrated by the normal PMN ability to release elastase, a proteinase needed to degrade foreign organisms and host tissue during inflammation. In fact, at baseline, the elastase release was similar in XLA and HD (Fig 4A and 4B). After E. coli stimulation, the elastase release rapidly increased in XLA and in HD. After IVIg infusion, PMN maintained their degranulation ability ( Fig 4B).  In vivo effects of IVIg infusion on innate immune cells of XLA Moreover, also the PMN migration induced by fMLP was comparable in XLA patients and HD. After IVIg, the PMN migration remained unaltered (Fig 5A and 5D). Thus, the explanation for the increased CD66b remained obscure.
E. coli stimulation induced an overexpression of all receptors in XLA patients and HD, with the exception of CD181, that was reduced, in line with a previous study [28].
After IVIg infusion, the expression of all receptors in unstimulated conditions and their overexpression induced by E. coli remained unaltered, demonstrating that IVIg did not influence the phenotype of PMN, as previously shown in CVID patients [38] and did not affect the PMN ability to respond to opsonized bacteria ( Table 2).

Phagocytosis
Phagocytosis contributes to the first line of defense against infection trough the engulfment and destruction of invading microorganisms. It also plays a key role in tissue homeostasis and remodeling.
Monocytes. Monocytes phagocytosis in XLA was normal at baseline (Fig 7A), but it was decreased about 25% after IVIg (Fig 7A). The phagocytosis induced by a non-opsonized E. coli was similar to the negative control confirming that E. coli was internalized through a FcγR pathway.
PMN. PMN from XLA displayed a normal phagocytosis (Fig 7B). After IVIg, it remained unaltered (Fig 7B), demonstrating that IVIg administered at replacement dosage did not affect the ability of PMN to perform an efficient phagocytosis. Also in PMN, we found that phagocytosis using non-opsonized E. coli was negligible.

Oxidative burst
The oxidative burst reflects the killing ability of innate immune cells and it involves many cellular pathways. The lack of BTK activity might influence the respiratory burst, a process requiring Ca 2+ mobilization from internal storage. We verified this hypothesis on monocytes and neutrophils subsets.
Monocytes. The monocytes ROS production after stimulation with opsonized E. coli or PMA was comparable in XLA and HD (Fig 8A and 8B). After IVIg infusion, the oxidative burst induced by E. coli slightly decreased about 13%, while it remained unaltered when PMA was used as stimulus (Fig 8A and 8B). As control, we analyzed the oxidative burst using nonopsonized E. coli in order to evaluate the oxidative burst resulting from the FcγR-independent phagocytosis. As expected, we found that the oxidative burst was comparable to negative control, confirming that opsonized E. coli induced the oxidative burst through a FcγRs pathway. The positive correlation between oxidative burst and phagocytosis (Panels A and B in S2 Fig) confirmed the dependence of oxidative burst process from phagocytosis [40][41][42][43].
PMN. The PMN ROS production after E. coli and PMA stimulation was comparable in XLA patients and HD (Fig 8C and 8D). After IVIg, the E. coli-induced and PMA-induced ROS  In vivo effects of IVIg infusion on innate immune cells of XLA production remained unchanged (Fig 8C and 8D). The PMN oxidative burst induced by nonopsonized E. coli was negligible.
All our observations were done in XLA patients under IVIg treatment since many years and thus an interference of this chronic treatment on our data could not be ruled out. In order to verify our findings, we analyzed a newly diagnosed untreated XLA patient (XLA naïve). We found that XLA naïve patient showed receptors expression, migration, phagocytosis and oxidative burst functions similar as XLA patients under chronic IVIg replacement.  In vivo effects of IVIg infusion on innate immune cells of XLA

Calcium chelation
The observation that BTK is not essential for the ROS production and that ROS production requires Ca 2+ mobilization [44,45] prompted us to verify if lack of expression of functional BTK in XLA patients might be bypassed by alternative mechanisms able to activate Ca 2+ mobilization or by Ca 2+ -independent mechanisms.
Phagocytosis. To verify the independence of monocytes phagocytosis by Ca 2+ mobilization [36], in XLA and HD we used the calcium chelator BAPTA-AM. Data confirmed that after calcium chelation the phagocytosis did not changed (Fig 9A and 9B). The independence of phagocytic process by Ca 2+ mobilization was observed also in PMN (Fig 9C and 9D).
Oxidative burst. E. coli-induced oxidative burst by monocytes was more inhibited by calcium chelation in HD than XLA (Fig 10A). The average reduction in XLA was about 50% ( Fig  10B) while in HD was about 75% (Fig 10C). This finding suggests that in XLA patients the Ca 2+ mobilization required for the ROS production was necessary but less critical than in HD. Also in PMN, BAPTA-AM induced a mild reduction of the E. coli-induced oxidative burst in Monocytes and PMN phagocytosis-Ca 2+ chelation. Whole blood samples were pre-treated with 100 μM of the calcium chelator BAPTA-AM for 30 min and incubated in a water bath for 20 min at 37˚C with opsonized E. coli (1-2x10 9 / ml). The conversion of DHR 123 to R 123 was used to evaluate the intracellular ROS production. FSC and SSC characteristics were used to identify the monocytes population. A regular phagocytosis was observed on monocytes and PMN from HD and XLA even after calcium chelation demonstrating that the process is Ca 2+ independent. Results are expressed as Mean Fluorescence Intensity. Histograms denote mean values and bars standard deviation. In (A and B) are shown monocytes and in (C and D) PMN phagocytosis of representative XLA patient: black peak BAPTA-AM(-) and gray peak BAPTA-AM(+). https://doi.org/10.1371/journal.pone.0175961.g009 In vivo effects of IVIg infusion on innate immune cells of XLA XLA than HD (Fig 11A). The average reduction in XLA was about 50% (Fig 11B), while in HD was about 75% (Fig 11C). By contrast, the Ca 2+ chelation did not affect PMN oxidative burst induced by PMA. Indeed, PMA induced a normal oxidative burst even when calcium was chelated (Fig 11D and 11F). This observation was expected since the PMA induces the ROS production through a direct binding to protein kinase C (PKC) [46,47].

Discussion
Mutations in BTK gene lead to an impaired B cell development resulting in absent or very low number of B-cells and profound hypogammaglobulinemia in patients affected by XLA [1,2]. Previous studies on patients with XLA [7][8][9] showed the involvement of BTK in innate immune cell responses [48] and its role in human DC responses upon TLR9 engagement [49]. Moreover, studies on the immune function of XLA patients who lack functional BTK might also provide information concerning the consequences of drugs acting as BTK inhibitors on innate immunity of patients under these new treatments [50][51][52][53][54].
In this study, we show that monocyte and PMN from XLA patients normally express those receptors involved in their activation, suggesting that they did not exhibit an altered activation status. Despite the lack of BTK, monocytes and PMN rapidly overexpressed surface receptors after stimulation with opsonized bacteria, suggesting that BTK could be dispensable or bypassed in pathways triggered by FcγR clustering. Moreover, the study of the oxidative burst and phagocytosis provided us information on the killing ability of innate immune cells in XLA patients. In agreement with previous studies [44,55], the phagocytosis and respiratory burst induced through FcγR were preserved in monocytes and PMN confirming that the BTK kinase activity is bypassed during the engagement of FcγR. These results also suggested that BTK might not be essential for Ca 2+ mobilization, an event normally required for ROS production. Indeed, activated BTK induce a rapid Ca 2+ release trough the phosphorylation of PLCγ2 that generates diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3). IP3 binds IP3 receptors and activates Ca 2+ -release channels on the endoplasmic reticulum (ER), triggering the calcium flux [14,56]. Both DAG and Ca 2+ are able to activate the PKC, the enzyme responsible for NADPH oxidase activation [46,47]. The lack of BTK in XLA patients might be bypassed by an efficient activation of PLCγ2 performed by Syk, as we show in Fig 12 [57]. In monocytes, PLCγ2 is activated by FcγRI and FcγRIIA whereas in neutrophils the engagement of FcγRIIA is sufficient for its activation [57]. Moreover, we found that BAPTA-AM exerted a greater inhibition of the oxidative burst in HD than XLA, suggesting that ROS production in XLA patients is less dependent on Ca 2+ mobilization. It is possible that the residual ability to produce ROS in XLA patients depend on a greater efficacy of DAG-dependent stimulation of PKC, an enzyme normally maintained in an inactive conformation [58][59][60]. In agreement with this finding, we showed that the Ca 2+ chelator BAPTA-AM did not affect the ROS production when PMA was provided as stimulus since phorbol esters are able to bind PKC in the same In vivo effects of IVIg infusion on innate immune cells of XLA binding site of DAG and thus inducing kinase activation [46,47,61]. Our observations are supported by others studies reporting different pathways for Ca 2+ mobilization [62][63][64] and are also in agreement with the conclusions made by Ren et al [36] who hypothesized that the inhibition of BTK kinase activity within a whole organism could have limited effects on the FcγR-mediated processes. Our results are also in agreement with the study of Marron et al [44] who showed a ROS production not augmented or compromised in neutrophils from patients with XLA but in contrast with the results of Honda et al [9] who described an excessive ROS production. One possibility that could explain these discrepancies is related to differences in the methods of PMN collection and in the stimuli used during the experimental procedures, in that we used a more physiological stimulus such as the opsonized E. coli or a higher PMA concentration. Moreover, we found that the FcγR-mediated phagocytosis performed by monocytes and PMN was fully functional and it remained unaltered when monocytes and PMN were pre-treated with a Ca 2+ chelator, confirming the independence of phagocytosis by Ca 2+ mobilization [65][66][67][68]. Thus, the BTK kinase activity or the Ca 2+ mobilization downstream BTK did not appear to be required for this process, as previously shown using a BTK inhibitor [36]. Schematic representation of signaling pathways for phagocytosis and oxidative burst. As a result of FcγR clustering, two separate pathways lead to phagocytosis, through a calcium independent way, or to the oxidative burst in a calcium dependent manner. Syk induce the activation of Vav and the downstream signaling that lead to the phagocytosis process. The main pathway for the activation of the oxidative burst consist in the activation of Btk by Syk and the consequent phosphorylation of PLCγ2. PLCγ2 activated produce DAG and IP3 from PIP3. DAG directly activate PKC while IP3 bind IP3R on endoplasmic reticulum inducing Ca 2+ mobilization that lead to PKC activation and finally to the oxidative burst. The lack of BTK in XLA patients is effectively bypassed by the direct stimulation of PLCγ2 by Syk. The inhibition of Ca 2+ mobilization by BAPTA-AM strongly reduce the oxidative burst. PMA is able to bind PKC to the same binding site of DAG and bypass both the production of DAG and the calcium flux from endoplasmic reticulum. https://doi.org/10.1371/journal.pone.0175961.g012 In vivo effects of IVIg infusion on innate immune cells of XLA The only abnormality found was the overexpression of CD66b, suggesting an hyperactivation status of PMN. CD66b is able to stimulate neutrophils and its crosslinking is able to induce the secretion of preformed interleukin-8 (IL-8) [69]. The dosage of the IL-8 in the plasma of patients was normal in unstimulated condition. After stimulation with E. coli, IL-8 showed a rapid increase, suggesting that PMN of XLA patients effectively respond to opsonized bacteria in terms of cytokine secretion. However, our data show that the altered expression of CD66b on PMN surface did not correlate with IL-8 and elastase secretion and therefore it needs to be better clarified.
Also others functions involved in functionality of PMN, such as migration and elastase degranulation, were well preserved in XLA patients, confirming that all the fundamental steps involved in PMN killing ability were fully functional.
XLA patients received a long-life substitutive treatment with polyvalent IgG as a prophylaxis against infections [20,21]. IVIg infusions could exert an anti-inflammatory function, even if administrated at replacement dosages, acting on monocytes as we recently showed in monocyte from CVID [23]. Interestingly, in line with our previous observation, XLA patients showed an increased percentage of intermediate monocyte, possibly caused by their inflammatory condition. This observation suggest that a greater frequency of intermediate monocytes could be a common feature of different PAD. The prompt reduction of the number of intermediate monocytes observed after IVIg is in line with the observation made in CVID patients [23] confirming the normalizing effect of IVIg infusion on monocytes. In addition, we show that the in vivo IVIg administration did not affect the expression of any surface receptors considered. Consistent with our previous study [23], we show that IVIg administration induced a decrease of monocytes' oxidative burst and phagocytosis. We hypothesize that the transient binding of IVIg to FcγR might reduce the availability of monocyte receptors for the binding with opsonized E. coli, resulting in a slight reduction of phagocytosis and consequently in a reduced ROS production.
In summary, the lack of BTK kinase activity in XLA did not affect monocytes and PMN functions. IVIg infusion exerted a dampening effect on monocytes compartment by reducing the pro-inflammatory subset and by reducing their killing ability.