Analysis of the Effects of the Bruton's tyrosine kinase (Btk) Inhibitor Ibrutinib on Monocyte Fcγ Receptor (FcγR) Function*

The irreversible Bruton's tyrosine kinase (Btk) inhibitor ibrutinib has shown efficacy against B-cell tumors such as chronic lymphocytic leukemia and B-cell non-Hodgkin lymphoma. Fcγ receptors (FcγR) on immune cells such as macrophages play an important role in tumor-specific antibody-mediated immune responses, but many such responses involve Btk. Here we tested the effects of ibrutinib on FcγR-mediated activities in monocytes. We found that ibrutinib did not affect monocyte FcγR-mediated phagocytosis, even at concentrations higher than those achieved physiologically, but suppressed FcγR-mediated cytokine production. We confirmed these findings in macrophages from Xid mice in which Btk signaling is defective. Because calcium flux is a major event downstream of Btk, we tested whether it was involved in phagocytosis. The results showed that blocking intracellular calcium flux decreased FcγR-mediated cytokine production but not phagocytosis. To verify this, we measured activation of the GTPase Rac, which is responsible for actin polymerization. Results showed that ibrutinib did not inhibit Rac activation, nor did the calcium chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis(acetoxymethyl ester). We next asked whether the effect of ibrutinib on monocyte FcγR-mediated cytokine production could be rescued by IFNγ priming because NK cells produce IFNγ in response to antibody therapy. Pretreatment of monocytes with IFNγ abrogated the effects of ibrutinib on FcγR-mediated cytokine production, suggesting that IFNγ priming could overcome this Btk inhibition. Furthermore, in monocyte-natural killer cell co-cultures, ibrutinib did not inhibit FcγR-mediated cytokine production despite doing so in single cultures. These results suggest that combining ibrutinib with monoclonal antibody therapy could enhance chronic lymphocytic leukemia cell killing without affecting macrophage effector function.

Fc␥ receptors (Fc␥Rs) 3 are critical for antibody-mediated responses because they provide for initial contact of the effector cell to the opsonized target cell and subsequently activate effector signaling pathways that lead to target destruction (1). Activation of Fc␥R also elicits cytokine production by monocytes, which serves to activate other immune effectors, such as natural killer (NK) cells (2)(3)(4)(5)(6)(7)(8). This involves the activation of Src kinases and Syk and then branches to other mediators (9). One such mediator downstream of Syk is Bruton's tyrosine kinase (Btk), which activates the PLC␥/calcium signaling pathway (10) and interacts with multiple proteins and pathways, leading to pleiotropic cellular responses (11,12). Btk is a member of the Tec family of kinases, which are expressed in mammals as well as other organisms such as Drosophila (13).
Btk is well known as a downstream mediator of the B cell receptor (14). This makes it a candidate therapeutic target for autoantibody diseases as well as for B cell lymphomas in which B cell receptor signaling is overactive. Indeed, the Btk inhibitor ibrutinib (PCI-32765) showed efficacy in murine models of arthritis and lupus and in canines with spontaneous B cell non-Hodgkin lymphoma (15). Presently, ibrutinib is approved for clinical marketing by the Food and Drug Administration for the treatment of relapsed mantle cell lymphoma, Waldenström macroglobulinemia and chronic lymphocytic leukemia (http://www.fda.gov/Drugs/InformationOnDrugs/ApprovedDrugs/ ucm432240.htm).
Ibrutinib binds irreversibly to Btk and blocks its kinase activity (15,16). It is also capable of blocking activation of the related Tec kinase Itk (17). Because numerous receptors utilize Btk (and Itk), ibrutinib can dampen the responses of immune cells to activating stimuli. For example, ibrutinib can inhibit degranulation and antibody-dependent cellular cytotoxicity and IFN␥ production by NK cells (18) and block Fc␥R-mediated cytokine production in monocytes/macrophages (19). However, the effects of ibrutinib on monocyte/macrophage Fc␥R have not been fully elucidated, especially within the context of neighboring immune cells.
Here we examined the effects of ibrutinib on monocyte/ macrophage Fc␥R signaling. We found that, although the drug blocked Fc␥R-mediated cytokine production, phagocytic ability was unaffected. Further inquiry revealed that this was because ibrutinib did not inhibit Rac activation downstream of Fc␥R. Interestingly, we also found that priming with IFN␥ prevented the inhibition of Fc␥R function by ibrutinib. In co-cultures of monocytes and NK cells, the endogenous production of IL-12 and IFN␥ following incubation with opsonized tumor cells was sufficient to overcome the negative effects of ibrutinib. Neutralizing either IL-12 or IFN␥ reduced the levels of both and of TNF␣. Therefore, although ibrutinib can block Btk (and Itk) in isolated cultures, proinflammatory intercellular communication is sufficient to overcome its inhibitory effects on monocytes and NK cells. These results suggest that ibrutinib may not be detrimental to patients undergoing antibody therapy and may explain early clinical results showing an enhanced effect when CD20 antibodies such as rituximab or ofatumumab are added to ibrutinib in chronic lymphocytic leukemia (20,21).

Experimental Procedures
Antibodies and Reagents-Ibrutinib and BAPTA-AM were purchased from Selleck Chemicals (Houston, TX). Recombinant human TNF␣ (used at 50 ng/ml) and anti-TNF␣ (used at 5 g/ml) were purchased from R&D Systems (Minneapolis, MN). TRIzol was purchased from Invitrogen. Reverse transcriptase, random hexamers, and SYBR Green PCR mix were purchased from Applied Biosystems (Foster City, CA). Anti-pBtK, anti-pPLC␥2 (Tyr-753 and Tyr-1217), anti-SyK, anti-pErK, anti-pAKT, and anti-pNF-B for Western blotting were purchased from Cell Signaling Technology (Beverly, MA). Anti-Rac2 was obtained from Abcam (Cambridge, MA). Anti-PLC ␥2, anti-Btk, and anti-Syk were purchased from Santa Cruz Biotechnology (Dallas, TX). Antibodies against actin and GAPDH as well as goat and mouse HRP-conjugated secondary antibodies were from Santa Cruz Biotechnology. Rabbit HRP-conjugated secondary antibody was purchased from Cell Signaling Technology. Human and mouse ChromPure whole-molecule IgG were bought from Jackson ImmunoResearch Laboratories (West Grove, PA).
ELISA Measurement of Cytokine Production-96-well plates were coated with 10 g/ml human or mouse IgG diluted in PBS overnight at 4°C. The plates were washed twice with sterile PBS, and cells were added for the incubation periods. Superna-tants were collected and centrifuged at 16,000 ϫ g to clear cells and debris and then analyzed for cytokines using sandwich ELISAs (R&D Systems) according to the protocols of the manufacturer.
Preparation of Heat-aggregated IgG-Human or mouse heataggregated IgG was prepared according to methods described previously (23). In brief, whole human or mouse IgG at a concentration of 350 g/ml in incomplete RPMI medium was incubated at 62°C in a water bath for 30 min and immediately placed on ice.
Culture of Murine Bone Marrow Macrophages-Wild-type and BTK mutant homozygous mice (XID) were housed and euthanized according to institution-approved animal care and use protocols. Bone marrow-derived macrophages were generated as described previously (24). Briefly, mouse bone marrow cells were flushed, cultured in RPMI medium containing 10% fetal bovine serum, and incubated on plates for 1 week with 20 ng/ml recombinant mouse M-CSF (R&D Systems) and 10 g/ml polymyxin B, both replenished on days 3 and 5. Cells were harvested for use on day 7.
MTS Assay-The MTS cell proliferation assay kit was purchased from Promega Corp. (Madison, WI), and the assays were done according to the instructions of the manufacturer. Cells were pretreated with ibrutinib at the indicated concentrations and then incubated in 96-well plates for 24 h. Following addition of CellTiter 96 AQueous One solution reagent and incubation, absorbance at 490 nm was measured using a standard plate reader.
Rac Activity Assay-The Rac1/Cdc42 activation assay kit was purchased from EMD Millipore Corp. (Billerica, MA), and the assay was done in accordance with the provided protocol. Cells were pretreated with ibrutinib or BAPTA-AM and stimulated with the heat-aggregated IgG for the indicated times. Cells were lysed in MLB buffer containing 10 mM Na 3 VO 4 and 10 g/ml each of aprotinin and leupeptin. Protein-matched cell lysates were incubated with GST-PAK1-PBD beads for 1 h at 4°C with gentle rocking. GTP␥S/GDP positive and negative controls were also prepared. After 1 h of incubation, beads were washed with MLB buffer and then boiled in 2ϫ Laemmli sample buffer (60 mM Tris, (pH 6.8), 2.3% SDS, 10% glycerol, 0.01% bromphenol blue, and 1% 2-mercaptoethanol) for 5 min. Western blot analyses were done to measure Rac.
Phagocytosis Assays-Phagocytosis assays were performed as described previously (25). Briefly, sheep red blood cells (Colorado Serum Co., Denver, CO) were fluorescently labeled with PKH-26 (Sigma), opsonized with anti-sheep red blood cell antibody (Sigma), and washed to remove excess antibody. These were then added to the effector cells for the indicated times at 37°C. Following hypotonic lysis of non-phagocytosed sheep red blood cells, effector cells were placed onto slides, and scoring of phagocytosis was done via fluorescence microscopy in a blinded manner. The phagocytic index was defined as the total number of sheep red blood cells ingested by 100 effector cells.
Co-cultures of Monocytes and NK Cells-Human NK cells and PBMs were isolated from leukopacks purchased from the Red Cross (n Ն 3). PBMs were isolated via adherence to Tefloncoated plates, and NK cells were isolated using Rosette-Sep (Siemens SAS, Paris, France) as described previously (26). The breast cancer cell line SKBR3 was incubated for 1 h at 37°C with 100 g/ml Herceptin and then washed and added to 96-well plates. Monocytes and NK cells were pretreated for 4 h with or without 1 M ibrutinib, washed, and then plated in isolation or in co-culture into wells with or without the Herceptin-opsonized SKBR3 tumor cells. Neutralizing antibodies against either human IFN␥ (5 g/ml) or human IL-12 (10 g/ml) (R&D Systems) were added to the appropriate wells. Cultures were incubated for 24 h, and then cleared supernatants were collected and analyzed for IL-12 and IFN␥ by ELISA.
Statistical Analyses-For the experiments that involved placing the cells of each donor across multiple conditions, data were analyzed by using analysis of variance with repeated measures using SAS 9.4 (SAS Inc., Cary, NC). For other experiments with only two groups involved, Student's t tests (paired or unpaired as appropriate) were used to test for statistically significant differences. p Յ 0.05 was considered significant. All error bars represent mean Ϯ S.D.

Ibrutinib Suppresses Fc␥R-mediated Cytokine Production but Not Phagocytosis in Human Monocytes-It has been shown that
Btk inhibition results in reduced Fc␥R-mediated cytokine production in monocytes and macrophages (19,27,28). However, the effect of reduced Btk function on phagocytosis is less clear, with studies showing significant (29 -31) or only marginal (32) effects. To examine this, we began by pretreating monocytes with or without ibrutinib and then treating them for 15 min with heat-aggregated IgG to cluster the Fc␥R. Western blot analyses for phospho-Btk, phospho-PLC␥2, and phospho-Syk were done. Results showed that, as expected, ibrutinib blocked IgG-mediated phosphorylation of Btk and PLC␥2 but not of upstream Syk (Fig. 1A).
Next we examined the effect of ibrutinib on Fc␥R-mediated cytokine production by pretreating cells with or without ibrutinib and then incubating them with or without immobilized IgG for 24 h. Cleared supernatants were assayed for TNF␣ production by ELISA, and the results showed that ibrutinib blocked Fc␥R-mediated cytokine production ( Fig. 1B), in agreement with Chang et al. (19). As part of a later experiment (Fig. 8), we also measured IL-12 as a cytokine known to not require Btk signaling.
Following this, we pretreated monocytes with or without ibrutinib and then tested the effect of ibrutinib on phagocytosis. Results showed that ibrutinib did not significantly affect the phagocytic ability of monocytes (Fig. 1C). As controls for the efficacy and relative specificity of ibrutinib, we also tested the effects of ibrutinib on TLR4-and TLR8-mediated cytokine production. The results showed that it did not inhibit TNF␣ production following LPS treatment but significantly inhibited TNF␣ following treatment with the TLR8 agonist CL075 (data not shown). This is in agreement with previous studies showing that, although Btk is involved with TLR4-and TLR8-mediated cellular activation (33,34), it is not required for TLR4-mediated TNF␣ production (35).
Higher Concentrations of Ibrutinib Do Not Affect Monocyte Fc␥R-mediated Phagocytosis-The disparity between the ability of ibrutinib to inhibit Fc␥R-mediated cytokine production and its inability to block Fc␥R-mediated phagocytosis raised the possibility that a greater concentration of ibrutinib would be required to block phagocytosis. To test this, we pretreated monocytes with 0, 1, 5, or 10 M ibrutinib and then treated them for 15 min with heat-aggregated IgG and measured phosphoproteins as shown in Fig. 1A. The results showed that FIGURE 1. Ibrutinib does not inhibit phagocytosis but suppresses Fc␥R-mediated cytokine production in human PBMs. Human PBMs were isolated and pretreated with 1 M ibrutinib (IB) for 30 min or left untreated (UT). A, following pretreatment, PBM were incubated for 15 min with 350 g/ml heat-aggregated whole human IgG (⌬IgG) or left untreated and then lysed and analyzed by Western blots for phosphorylated Btk, phosphorylated PLC␥2, and phosphorylated Syk. Membranes were reprobed for ␤-actin to verify equivalent loading (n ϭ 3, a representative blot is shown). B, pretreated PBMs were incubated for 24 h in wells coated without (PBS) or with 10 g/ml whole human IgG. Cleared supernatants were collected and analyzed for TNF␣ by ELISA (n ϭ 4). C, the phagocytic ability of pretreated PBMs was measured as described under "Experimental Procedures." Phagocytic Index denotes the total number of opsonized targets ingested by 100 PBMs. Counts were randomized and performed in triplicate for each donor, with a total of five donors. *, p Յ 0.05. Error bars represent mean Ϯ S.D.
the lowest concentration completely inhibited phosphorylation of Btk and PLC␥2 but that no concentration inhibited the phosphorylation of upstream Syk ( Fig. 2A).
We then pretreated monocytes with ibrutinib, incubated the cells for 24 h, and then performed an MTS proliferation assay as a measure of cell survival. As shown in Fig. 2B, none of the ibrutinib concentrations affected survival. Following this, we tested the effects of these increasing ibrutinib concentrations on Fc␥R-mediated cytokine production and phagocytosis. Although Fc␥R-mediated cytokine was blocked even by the lowest concentration of ibrutinib (Fig. 2C), not even the highest concentration significantly affected phagocytosis (Fig. 2D).
Macrophages from Xid Mice Show Defects in Fc␥R-mediated Cytokine Production but Not in Phagocytosis-Xid (CBA/N) mice show an intrinsic defect in B cell development (36) and have a cytosine-to-thymidine mutation at position 219. This results in an arginine-to-cysteine change within the N-terminal region of Btk (37,38), which contains the pleckstrin homology region that interacts with protein kinase C (39). It has been shown previously that peritoneal macrophages from Xid mice do not show defects in their ability to ingest heat-killed Escherichia coli (32). Regarding Fc␥R-mediated phagocytosis, the xid gene was able to rescue the decreased ability of peritoneal macrophages from autoimmune-prone NZB mice to ingest opsonized sheep red blood cells (40). Here we examined the Fc␥R responses of Xid mouse bone marrow-derived macrophages (BMMs) to compare them with the results seen with human monocytes.
Upon stimulation with heat-aggregated IgG, BMMs from Xid mice showed significantly reduced phosphorylation of Btk and the downstream PLC␥2 (Fig. 3A). Similarly, Xid BMMs showed reduced Fc␥R-mediated TNF␣ production (Fig. 3B). However, Fc␥R-mediated phagocytosis was not different between wild-type and Xid mice (Fig. 3C). These results suggest that Fc␥R responses of macrophages from these Btk-defective mice closely resemble the responses seen in ibrutinib-treated human PBMs in that cytokine production, but not phagocytosis, is affected.
Intracellular Calcium Flux Is Required for Fc␥R-mediated Cytokine Production but Not for Phagocytosis-One major signaling event downstream of Btk that is blocked by ibrutinib and other Btk inhibitors is calcium flux (19,28). Here we tested the effects of ibrutinib within the context of calcium signaling and actin polymerization. To begin, we pretreated monocytes with or without the calcium chelator BAPTA-AM for 30 min and then treated the cells for 15 min with heat-aggregated IgG. Western blot analyses showed that BAPTA-AM had no effect on the phosphorylation of Btk or PLC␥2, as expected (Fig. 4A).
To then determine the effect of calcium chelation on Fc␥Rmediated cytokine release, we pretreated monocytes with or without ibrutinib or BAPTA-AM and then incubated the cells on immobilized IgG for 24 h. ELISAs with cleared supernatants showed that both ibrutinib and BAPTA-AM significantly reduced Fc␥R-mediated cytokine production (Fig. 4B), indicating that inhibitor treatment was effective. Next, we pretreated monocytes for 30 min with or without BAPTA-AM or ibrutinib and performed phagocytosis assays. The results showed that neither BAPTA-AM nor ibrutinib could significantly reduce Fc␥R-mediated phagocytosis (Fig. 4C), despite their effect on Fc␥R-mediated cytokine production. These results suggest  Ϫ). A, pretreated PBMs were incubated for 15 min with 350 g/ml heat-aggregated whole human IgG (⌬IgG) or left untreated and then lysed and analyzed via Western blot for phosphorylated Btk, phosphorylated PLC␥2, and phosphorylated Syk. Membranes were reprobed for ␤-actin to verify equivalent loading (n ϭ 3, a representative blot is shown). B, pretreated PBM were incubated overnight, and then cell survival was measured using an MTS assay (n ϭ 4). C, pretreated PBMs were incubated for 24 h in 96-well plates precoated without (PBS) or with 10 g/ml whole human IgG. Cleared supernatants were collected and analyzed by ELISA for TNF␣ (n ϭ 3). D, pretreated PBMs were assayed for phagocytic ability as described under "Experimental Procedures" (n ϭ 4). Error bars represent mean Ϯ S.D.
that, although calcium flux may be required for Fc␥R-mediated cytokine production, the effects of ibrutinib on calcium signaling do not affect Fc␥R-mediated phagocytosis.
Rac is a GTPase that is responsible for actin polymerization following Fc␥R activation and subsequent phagocytosis. To determine whether ibrutinib or downstream Ca 2ϩ mobilization had an effect on Rac, we pretreated monocytes with or without ibrutinib for 30 min, followed by treatment with heat-aggregated IgG for 5 min. We then performed a Rac activation assay and found that neither ibrutinib (Fig. 5A) nor chelation of Ca 2ϩ (Fig. 5B) affected Fc␥R-mediated Rac activation. Collectively, these results suggest that the Btk inhibitor ibrutinib blocks Fc␥R-mediated cytokine production, likely by its effect on calcium signaling, but this has little to no effect on Rac activation. Therefore, ibrutinib shows minimal effects on monocyte-mediated phagocytosis. This is summarized in Fig. 5C and is in agreement with Patel et al. (41), in whose work Vav could activate Rac independently of Ca 2ϩ .
IFN␥ Rescues Fc␥R-mediated Cytokine Production following Ibrutinib Treatment-It has been well established that priming of monocytes with IFN␥ can markedly enhance Fc␥R-mediated cytokine production as well as other functions, such as phagocytosis and reactive oxygen species production. This is important in the context of antibody therapy because NK cells are known to secrete substantial quantities of IFN␥ upon engaging an antibodycoated target. This led us to test whether IFN␥ pretreatment could overcome the effects of ibrutinib and whether the IFN␥ produced by NK cells could act in a paracrine fashion on monocytes to combat the inhibition by ibrutinib. To begin, we pretreated monocytes overnight with or without IFN␥, treated them with ibrutinib or left them untreated, and then incubated them for 24 h in 96-well plates with or without immobilized IgG. The results from the ELISAs showed that ibrutinib blocked Fc␥R-mediated cytokine production as in the figures above but that pretreatment with IFN␥ rescued this inhibition (Fig. 5D).
We then proceeded to test the effects of ibrutinib on downstream mediators of Fc␥R activation to determine which were affected and which might be rescued by IFN␥. Ibrutinib has been shown to inhibit the activation of NF-B, Erk, and Akt (42)(43)(44), so we tested whether IFN␥ priming could rescue their phosphorylation following Fc␥R activation in ibrutinib-pretreated monocytes. To begin, we incubated PBMs with ibruti- FIGURE 3. BMMs from XID mice show defects in Fc␥R-mediated cytokine production but not in phagocytosis. BMMs were generated from XID mice (n ϭ 3) and wild-type controls (n ϭ 4) as described under "Experimental Procedures." A, BMMs were incubated for 15 min either without (R, resting) or with 350 g/ml heat-aggregated IgG (⌬IgG), lysed, and analyzed via Western blot for phosphorylated Btk (pBtk) and phosphorylated PLC␥2 (pPLC␥2). Membranes were reprobed for total Btk and total PLC␥2 to verify equivalent loading. B, wild-type or XID BMMs were incubated for 24 h in 96-well plates precoated without (PBS) or with 10 g/ml whole mouse IgG. ELISAs were done to measure TNF␣ in cleared supernatants (n ϭ 3). *, p Յ 0.05 between WT and XID TNF␣ values. C, BMMs from wild-type and XID mice were assayed for phagocytic ability as described under "Experimental Procedures" (n ϭ 4). The phagocytic index was counted as the number of opsonized targets ingested by 100 BMMs. Error bars represent mean Ϯ S.D. FIGURE 4. Intracellular calcium flux influences cytokine production but not phagocytosis in human PBMs. A, PBMs were pretreated without (untreated, UT) or with 5 M of the calcium chelator BAPTA-AM or 1 M ibrutinib and then incubated for 15 min either without (R, resting) or with 350 g/ml heat-aggregated whole human IgG (⌬IgG). Cells were lysed and assayed for phosphorylated Btk (pBtk) and phosphorylated PLC␥2 (pPLC␥2) via Western blotting. Membranes were reprobed for total Btk and total PLC␥2 to verify equivalent loading (n ϭ 3, a representative blot is shown). B, PBMs were pretreated with 5 M BAPTA-AM or 1 M ibrutinib (IB) or left untreated. Cells were then incubated for 24 h in 96-well plates precoated without (PBS) or with 10 g/ml whole human IgG. ELISAs were done to measure TNF␣ in the cleared supernatants (n ϭ 4). C, cells were pretreated as in B and then subjected to phagocytic assays as described under "Experimental Procedures" (n ϭ 3). The phagocytic index is a count of the number of opsonized targets ingested by 100 PBMs. *, p Յ 0.05. Error bars represent mean Ϯ S.D. nib for 1 h and then treated them with heat-aggregated IgG to activate the Fc␥R or left them untreated. Protein lysates were collected to measure the phosphorylation of Btk itself, along with NF-B, Akt, and Erk. As shown in Fig. 6, pretreatment with ibrutinib led to significant reductions in the phosphorylation of all four proteins. Next, to test whether IFN␥ could prevent this block of phosphorylation, we primed monocytes with or without IFN␥ overnight, treated them with ibrutinib for 1 h, and stimulated the cells with heat-aggregated IgG or left them untreated. The results showed that IFN␥ did not block the activity of ibrutinib against Btk, NF-B, or Akt (Fig. 7, A-C, respectively) because significant decreases in Fc␥R-mediated phosphorylation were still seen with ibrutinib. However, IFN␥ prevented the ibrutinib-mediated reduction in Fc␥R-mediated Erk phosphorylation (Fig. 7D, p ϭ 0.259). These results suggest that IFN␥ can rescue Erk pathway activation downstream of Fc␥R.
Because IFN␥ could reverse the suppression of Fc␥R-mediated cytokine production that was seen with ibrutinib, we tested whether measurable IFN␥ would be produced when NK cells and monocytes were co-cultured in the presence of antibody-coated target cells. Simultaneously, because it has been shown in macrophages from Xid mice that Btk deficiency leads to enhanced IL-12 production (45), we also wished to determine whether IgG-stimulated monocytes would produce IL-12 even after ibrutinib treatment. This IL-12, if produced, might serve to further activate the IgG-stimulated NK cells.
To examine this, PBMs and autologous NK cells were treated for 4 h with ibrutinib or left untreated and then incubated either separately or in co-culture. Each single and co-culture incubation was done for 24 h with or without Herceptin-coated SKBR3 breast cancer cells (Fc␥R-activating stimuli), and coculture wells were also treated with neutralizing antibodies against IL-12 or IFN␥ or left untreated. Following these incubations, cleared supernatants were analyzed for IL-12 and IFN␥ by ELISA. Results showed that, as predicted, ibrutinib did not block monocyte IL-12 production following incubation with opsonized target cells (Fig. 8A). In addition, production of IL-12 (Fig. 8A) along with IFN␥ and TNF␣ (Fig. 8, B and C, respec- then incubated for 5 min either without (R, resting) or with 350 g/ml heat-aggregated whole human IgG. Cell lysates were then subjected to a Rac activity assay as described under "Experimental Procedures," with blots for Rac2 serving as a readout for activated Rac. GTP and GDP were positive and negative controls, respectively (Ctl). representative blot is shown (n ϭ 4). Bottom panel, densitometry values for all four donors. B, PBMs were left untreated or pretreated with 5 M BAPTA-AM and then lysed and subjected to a Rac activity assay as in A. A representative blot is shown (n ϭ 5). Bottom panel, densitometry values for all five donors. C, proposed model of the separate signaling pathways for phagocytosis versus cytokine production downstream of PI3K. D, PBMs (n ϭ 3 donors) were treated overnight with 10 ng/ml IFN␥ or left untreated and then incubated for 24 h in 96-well plates precoated without (PBS) or with 10 g/ml whole human IgG. Cleared supernatants were collected and analyzed by ELISA for TNF␣. *, p Ͻ 0.05 versus untreated cells. Error bars represent mean Ϯ S.D. tively) was significantly higher in co-cultures than in the single cultures, with the exception of monocyte TNF␣ versus co-culture with TNF␣. This increase was likely due to the intercellular communication between monocytes and NK cells because the neutralizing antibodies against IL-12 and IFN␥ each reduced the levels of the other (Fig. 8, A and B, respectively). Importantly, ibrutinib had little to no effect on Fc␥R-mediated cytokine production in these co-cultures. The levels of IL-12 and TNF␣ were unchanged, whereas IFN␥ showed a small but statistically significant decrease. Because ibrutinib did not reduce the levels of TNF␣ in co-cultures but neutralizing antibody against IFN␥ did, it appears as though the level of IFN␥ that was produced in co-cultures was still high enough to rescue monocytes from the effects of ibrutinib. Collectively, these results suggest that ibrutinib may not inhibit Fc␥R activity in heterogeneous immune cell populations, such as in patients undergoing treatment.

Discussion
Ibrutinib is approved for the treatment of relapsed mantle cell lymphoma, Waldenström macroglobulinemia, and chronic lymphocytic leukemia. Because of the potential negative influence on NK cells published by our group and others (18,29), its role in combination therapy with monoclonal antibodies such as rituximab has been questioned. However, NK-cell mediated antibody-dependent cellular cytotoxicity is only one mechanism by which therapeutic antibodies deplete tumor cells. In  addition to direct killing and complement-mediated killing, monocyte-and macrophage-mediated, antibody-dependent phagocytosis represents another relevant mechanism of tumor elimination. Here we examined in detail the effects of ibrutinib on monocyte Fc␥R function, which has also been shown to be important for antibody-mediated tumor clearance. The results demonstrated that ibrutinib did not adversely affect the phagocytic ability of monocytes despite its ability to block Fc␥Rmediated TNF␣ production by monocytes. We also tested the effects of the more selective Btk inhibitor CGI-1746 (28) and found that it was equal in its ability to block Fc␥R-mediated cytokine production in monocytes (data not shown). This is in contrast to the findings with NK cells, where ibrutinib, but not CGI-1746, blocked NK-cell antibody-dependent cellular cytotoxicity because activated NK cells express Itk, which is not inhibited by CGI-1746 (18). Our results regarding TNF␣ pro-duction are in agreement with others both for primary human and murine cells. For example, Horwood et al. (46) have shown that TNF␣ mRNA was reduced in mononuclear cells from Xid mice following LPS treatment. Chang et al. (19) have shown that ibrutinib reduced Fc␥R-induced TNF␣, IL-1␤, and IL-6 in primary monocytes.
The lack of ibrutinib effect on monocyte/macrophage-mediated phagocytosis is in agreement with Mangla et al. (32), who found that phagocytosis of heat-killed fluoresceinated E. coli was not affected in macrophages from Xid mice. This lack of effect is likely not due to insufficient levels of ibrutinib because we tested with levels of 1 and 10 M, and we saw effective blocking of Btk phosphorylation even at 1 M. In addition, it has been shown that as little as 1-10 nM is sufficient to block Fc␥Rmediated cytokine production in monocytes and THP-1 cells (19), which suggests that Btk kinase activity (and, most notably, its downstream Ca 2ϩ mobilization) is not required for phagocytosis. Indeed, it appears more likely that the lack of ibrutinib effect on phagocytosis is due to the non-dependence of phagocytosis on Ca 2ϩ mobilization (47)(48)(49)(50). Ca 2ϩ mobilization is a major downstream event following Btk activation, and both genetic and pharmacologic reduction of Btk activity reduce the mobilization of Ca 2ϩ . For example, mast cells of Xid mice are defective in Ca 2ϩ mobilization (51), as are B cells, primary monocytes, and THP-1 cells treated with ibrutinib (19).
Our results with human PBMs did not show a consistent increase in Fc␥R-mediated IL-12 production, but, perhaps as importantly, we did not observe any decrease. It has been shown previously by Mukhopadhyay et al. (52) that Xid mice displayed a macrophage-dependent skewing toward Th1 characteristics that led to a delay in the clearance of filarial infection. Follow-up work by the same group has shown that the decreased nitric oxide induction in Btk-deficient macrophages permitted higher levels of IL-12 production (45,53). Therefore, in the context of multiple immune cells, it is possible, and even perhaps likely, that ibrutinib treatment would not block the production of IL-12. This, in turn, might serve to elicit IFN␥ from neighboring NK cells. The NK cells would produce biologically relevant (albeit perhaps small) amounts of IFN␥, which, in turn, would work toward ameliorating the effects of ibrutinib on monocytes. Along with this, ibrutinib has already been shown to promote Th1 responses by inhibiting Itk in Th2 cells (17), which could serve as an additional source of IFN␥ for the monocytes.
We also found that ibrutinib could not block TNF␣ production in monocytes primed with IFN␥ (Fig. 5D) or in co-cultures of monocytes and NK cells (Fig. 8C). Therefore, within a mixed cell setting, it is likely that Fc␥R-activated monocytes can still produce not only IL-12 (discussed above) but also other chemokines/cytokines that could serve to activate neighboring cells.
When taken collectively, these data suggest that, although ibrutinib can effectively inhibit proinflammatory responses in isolated cells, its ability to do so within the whole organism is limited. This may explain why early clinical results showed an enhanced effect when adding rituximab or ofatumumab to ibrutinib in treatment for chronic lymphocytic leukemia (20,21). Anti-CD20 antibodies such as rituximab and obinutuzumab are key components of therapy for diseases such as A-C, autologous PBMs and NK cells were isolated, treated for 4 h without (ϪIB) or with (ϩIB) 1 M ibrutinib, and then cultured either separately or in monocyte-NK cell co-cultures as indicated. Cells were incubated for 24 h without (ϪTarget) or with (ϩTarget) Herceptin-coated SKBR3 cells and with or without neutralizing antibodies against IL-12 or IFN␥ (␣IL-12 and ␣IFN␥). Cleared supernatants were collected and analyzed for IL-12 (A), IFN␥ (B), and TNF␣ (C). Graphs show mean values from three separate donors, and error bars represent mean Ϯ S.D. Significant differences for single and double asterisks would be p Ͻ ϭ 0.05. *, significant differences between co-culture versus each single-culture and neutralizing-antibody condition; **, significant difference between co-cultures without versus with ibrutinib; ns, no significant difference versus co-culture. Error bars represent mean Ϯ S.D. chronic lymphocytic leukemia or non-Hodgkin lymphoma, and these results suggest a potential mechanism by which their effectiveness would not be compromised by concurrent treatment with a Btk inhibitor.
Author Contributions-J. P. B. and S. T. designed the study and wrote the manuscript. L. R., A. C., S. G., H. F., S. E., K. F., P. M., A. S., and B. F. R. performed experiments and acquired data for the figures. X. M., J. C. B., W. E. C., J. P. B., and S. T. analyzed and interpreted the data. J. C. B. and W. E. C. performed critical revisions of the manuscript. All authors reviewed and approved the final version of this manuscript.