Fc gamma receptors promote antibody-induced LILRB4 internalization and immune regulation of monocytic AML

Abstract The immune checkpoint leukocyte immunoglobulin-like receptor B4 (LILRB4) is found specifically on the cell surface of acute monocytic leukemia (monocytic AML), an aggressive and common subtype of AML. We have developed a humanized monoclonal IgG1 LILRB4-blocking antibody (h128-3), which improved immune regulation but reduced cell surface expression of LILRB4 in monocytic AML models by 40–60%. Interestingly, most of this effect was neutralized by mutation of the Fc region of the antibody (h128-3/N297A), which prevents interaction with Fc gamma receptors (FcγRs). This suggested that there is FcγR-dependent antigenic modulation underlying h128-3’s effects, a mechanism known to alter the function of antibodies targeting B-cell malignancies. We disrupted the Fc-FcγR interaction pharmacologically and with stable CRISPR-Cas9-mediated genetic knockout of FcγRs in monocytic AML cell lines to investigate the role of FcγR-dependent antigenic modulation in the regulation of LILRB4 by h128-3. When FcγRI is inhibited or removed from the surface of monocytic AML cells, h128-3 cannot optimally perform its blocking function, resulting in activation of the LILRB4 inhibitory receptor and leading to a 15–25% decrease in T-cell-mediated cytotoxicity in vitro. In the absence of FcγRI, scaffolding by FcγRIIa allows h128-3 to maintain LILRB4-blocking function. Here we define a FcγR-dependent antigenic modulation mechanism underlying the function of an immunoreceptor blocking antibody for the first time in myeloid malignancy. This research will facilitate the development of safe, precision-targeted antibody therapeutics in myeloid malignancies with greater potency and efficacy.


INTRODUCTION
Acute myeloid leukemia (AML) is the most common acute blood and bone marrow cancer in adults and is characterized by rapid growth of abnormal myeloblastic cells that build up in the bone marrow and blood, interfering with normal hematopoiesis [1].Approximately 10% of AML cases belong to the monocytic AML (M5-AML) subtype, which carries a poor prognosis.NCI SEER data collected checkpoint that suppresses T-cell proliferation and as a tumor-promoting molecule that enhances tumor migration [3].Disrupting the interaction of LILRB4 with its functional ligand ApoE by an efficacious humanized antibody h128-3 reverses this T-cell suppression phenotype and blocks AML development [5].Thus, targeting LILRB4 with antibodies represents a tumor-specific and effective therapeutic strategy for treatment of monocytic AML.
Here we report that the anti-LILRB4 mAb h128-3 induced internalization and degradation of LILRB4 on monocytic AML cells in an Fcγ RI-dependent manner.We also confirm that the low-affinity Fcγ RIIa can scaffold h128-3 and internalize the h128-3/LILRB4 complex, particularly in the absence of functional high-affinity Fcγ RI, which may be saturated by circulating IgG present at a serum concentration of 7-16 mg/ml in the physiologic setting [18].This work characterizes a novel mechanism of Fcγ R-mediated antigenic modulation of LILRB4 in monocytic AML and will aid the development of a new generation of therapeutic antibodies optimized to take advantage of this unique biology in myeloid malignancy while maintaining Fc-mediated immune effector functions such as ADCC and ADCP.

Construction of stable HEK293T overexpression cell lines
HEK293T expressing human LILRB4 with or without Fcγ RI were generated using lentivirus.Briefly, the receptor cDNA genes (Sino Biological HG16742 and HG10256) were cloned into pCDH-CMV-MCS-IRES-Puro (for LILRB4) or pCDH-CMV-MCS-IRES-Bls (for Fcγ RI) vectors downstream of the CMV promoter to create the transfer plasmids.The HEK293T cell lines were generated by transducing with packaged lentivirus (generated using the transfer plasmid, pCMV-VSV-G (Addgene 8454), pCMV delta R8.2 (Addgene 12263)).Cells expressing the transgene were selected by 2.5 μg/ml puromycin (Gibco) with or without 5 μg/ml blasticidin (Gibco) until a sufficient number of cells with transgene emerged.For the cells that co-express LILRB4 and Fcγ RI, the LILRB4-HEK293T stable cell line was first established using lentivirus and selected under 2.5 μg/ml puromycin.Then, the lentivirus particles with Fcγ RI were used to transduce LILRB4-HEK293T.The double-positive cell lines were selected with 2.5 μg/ml puromycin and 5 μg/ml blasticidin until a sufficient number of cells with transgene emerged.Expression of LILRB4 and Fcγ R on these stable cell lines was detected by flow cytometry.

Generation of N297A mutated antibody
In our previous study, we constructed the humanized IgG 1 anti-LILRB4 mAb h128-3 (wild-type) [5].The variant containing a modified Fc region was made using site-directed mutagenesis PCR.The modification made was an alanine (A) substitution at amino acid N297 of the CH2 region in the h128-3 heavy chain.Mutated heavy-chain and wild-type light-chain constructs were cotransfected into human embryonic kidney freestyle 293 (HEK293F) cells using transfection reagent PEI (Sigma).After 7 days of expression, supernatants were harvested and antibodies were purified by affinity chromatography using protein A resin, as we previously reported (Repligen) [19].

ELISA binding assay
Corning 96-well EIA/RIA plates were coated for 18 h at 4 • C with LILRB4 recombinant proteins (1 μg/ml) and blocked for 2 h at 37 • C with 5% non-fat milk.After washing with PBST three times, 100 μl of serially diluted h128-3 antibodies were added and incubated for 45 min at 37 • C. Subsequently, the plates were washed with PBST and incubated for 30 min with HRP-conjugated goat antihuman F(ab') 2 (Jackson ImmunoResearch Laboratories).The immunoreactions were developed with TMB substrates (Sigma) and stopped by the addition of 2 M sulfuric acid before the plate was read at 450 nm.

Western blot
Cell lysates prepared from cell culture were subjected to SDS-PAGE separation.Gels were transferred to a polyvinylidene fluoride membrane by standard procedures.Membranes were blotted using a goat anti-rabbit Fchorseradish peroxidase (Sigma) and images were detected with FluorChem M imager (Cell BioSciences) using enhanced chemiluminescence substrate (Denville Scientific).A stalk region linear epitope targeting mAb generated in house (R8, 2 μg/ml), was used to detect LILRB4.Membranes were incubated with antibody for 18 h at 4 • C.

Flow cytometry analysis
For analysis of cell surface receptors, cells were run on the Sartorius iQue3 instrument based on the manufacturer's instructions.Briefly, 2-5 × 10 5 cells were dispensed in 100 μl aliquots and blocked with 300 μg/ml hIgG on ice for 1 h.Primary antibodies (5 μg/ml) were then added for 1 h on ice, followed by the addition of labeled secondary detection antibodies as needed.After washing with PBS buffer, the cells were analyzed for fluorescence inten-sity.Irrelevant rabbit or human IgG was used as negative control.

LILRB4 internalization assay
Monocytes were seeded in 24-well plates (2 × 10 5 cells/well, 1 ml) and incubated with antibodies (0.1-10 μg/ml) for 24 h (or different time points) at 37 • C. To check LILRB4 internalization, cells were blocked with 300 μg/ml hIgG, Fc fragment (Jackson ImmunoResearch Laboratories) before staining of surface LILRB4 with 5 μg/ml non-competitive rabbit anti-LILRB4 antibody R193 (generated in-house) and performed by FACS.APC-conjugated goat F(ab ) 2 anti-rabbit F(ab ) 2 (Jackson ImmunoResearch Laboratories) was diluted 1/200 and used as detection antibody.Internalization was expressed as percentage of surface LILRB4 change treated with anti-LILRB4 antibodies relative to surface LILRB4 treated with irrelevant IgG control: For examination of antibody-induced LILRB4 internalization, a method has previously been reported that detects the internalized antibody using a pH-dependent fluorescence probe.The probe enables maximum fluorescence signals of antibody under intracellular environment [20].The internalization of anti-LILRB4 antibody was detected using this method.Briefly, antibodies were conjugated with pHAb Amine Reactive dyes (Promega) and then diluted with cell culture media.Monocytic AML cells were seeded into a 24-well plate (4 × 10 5 cells/well).A total of 100 μl of medium containing different concentrations of pHAbconjugated antibodies were added into each well.After incubation at 37 • C for 18 h, the internalization of anti-LILRB4 antibodies was measured by flow cytometry of fluorescence in the far-red emission channel following the manufacturer's instructions.The contribution of cis and trans interactions to internalization of anti-LILRB4 antibody was also detected using this method.In brief, pHAb-conjugated anti-LILRB4 or irrelevant hIgG antibodies (10 μg/ml) were incubated with donor monocytes (GFP + ) for 1 h at RT. Acceptor monocytes (GFP − ) were then added 1:1, and co-cultured cells were incubated at 37 • C for various time points.Fluorescence of pHAbconjugated anti-LILRB4 and irrelevant hIgG antibodies by donor or acceptor monocytes was measured relative to untreated donor or acceptor monocytes by flow cytometry of gated GFP + or GFP − cells, respectively.For example, internalization in donor cells was expressed as a fold change relative to untreated cells: Normalized pHAb internalization = MFI of pHAb treated GFP + cells in sample x MFI of untreated GFP + cells .
After blocking with Duolink Blocking Solution (Sigma) for 1 h at 37

LILRB4 blocking assay
THP-1 cell lines were seeded in 6-well plates (1 × 10 7 cells/well, 2 ml) and serum-starved for 18 h at 37 • C to induce cell cycle synchronization.The cells were then incubated with serum-free media supplemented with PBS, hIgG or h128-3 (10 μg/ml) for 1 h at 37 • C. ApoE2 (5 μg/ml, Peprotech) and anti-HLA-DR antibody (5 μg/ml, L243, Biolegend) diluted in PBS were plated on non-treated 6-well tissue culture plates for 1 h at 37 • C, blocked with 2% BSA/PBS for 30 min at RT and washed twice with sterile PBS at RT.The treated THP-1 cells were then stimulated with bound ApoE2 and anti-HLA-DR antibody for 15 min at 37 • C and lysed for 10 min in 1% NP-40 buffer (Alfa Aesar) at 4 • C in the presence of protease and phosphatase inhibitors (cOmplete, PhosSTOP).Five percent of the total protein lysate was collected for western blot detection.LILRB4 was immunoprecipitated from the remaining lysed protein for 18 h at 4 • C with high-affinity rabbit anti-LILRB4 antibody (R8, 5 μg/ml) using the Dynabeads Protein A Immunoprecipitation Kit (Thermo Fisher).Input protein lysates and immunoprecipitated LILRB4 were subjected to SDS-PAGE separation, proteins were transferred to polyvinylidene fluoride membrane by standard procedures and membranes were blocked with 5% BSA/TBS-T for 30 min at RT then immunoblotted with mouse anti-pTyr antibody 4G10 (1:1000; Cell Signaling) for 18 h at 4 • C.
Membranes were washed with TBS-T, stained with Tidy-Blot Western Blot Detection Reagent:HRP (1:400; Bio-Rad) for 1 h at RT; then, images were detected.Membranes were then stripped with mild stripping buffer (Glycine, 0.2 M, pH 2.2) for 15 min at RT, washed with TBS-T and re-blocked with 5% BSA/TBS-T for 30 min at RT. Stripped membranes were then re-probed with rabbit anti-LILRB4 antibody R8 (2 μg/ml) for 1 h at RT and stained with TidyBlot Western Blot Detection Reagent:HRP (1:400; Bio-Rad) for 1 h at RT before image detection.Images were captured with a FluorChem M imager (Cell BioSciences) using enhanced chemiluminescence substrate (Denville Scientific).
Baseline surface LILRB4 levels in two LILRB4-expressing monocytic AML cell lines (THP-1 and Mono-mac-6) were measured by flow cytometry with the same LILRB4targeting antibody R193 (Fig. 1A).There were high levels of surface LILRB4 on both cell lines.To confirm potential antibody-mediated LILRB4 downregulation, we carried out a series of in vitro studies in these two cell lines.We first treated THP-1 with 10 μg/ml of h128-3 or control hIgG.At different time points, cells were collected and surface LILRB4 was measured by flow cytometry.As shown in Fig. 1B, surface LILRB4 on THP-1 cells was reduced by about 60% following treatment with h128-3 for 24 h.The internalization of h128-3 started immediately following antibody treatment and reached a plateau at 24-48 h.Western blot analysis revealed that total LILRB4 was similarly reduced following treatment with h128-3 (Fig. 1C).We repeated this internalization study in Mono-mac-6 and confirmed that LILRB4 downregulation occurred in this monocytic AML cell line as well (Fig. 1D).However, surface LILRB4 on Mono-mac-6 cells was reduced more slowly by h128-3, reaching a maximum of about 40% reduction following treatment with h128-3 for 48 h.Western blot of lysates prepared from these cells revealed that total LILRB4 was similarly reduced (Fig. 1E).We next measured the dose-response of h128-3 for inducing LILRB4 internalization in THP-1 and Mono-mac-6.As shown in Fig. 1F and G, h128-3 has high potency in triggering LILRB4 internalization on THP-1 and Mono-mac-6 cells, respectively.To further confirm LILRB4 internalization, we next labeled h128-3 with a pH-dependent fluorescence probe (pHAb).This probe enables maximum APC-channel fluorescence signals of antibody in the intracellular environment but very weak signal in neutral conditions (Fig. 1H).We treated THP-1 and Mono-mac-6 cells with different concentrations of conjugated antibodies and as shown in Fig. 1I and J, the internalization of h128-3/pHAb is significantly higher than that of isotype control antibody hIgG/pHAb in THP-1 cells but not in Mono-mac-6 cells.This result suggests that the h128-3/LILRB4 complex was internalized in THP-1 cells but h128-3 was not internalized with LILRB4 as a complex in Mono-mac-6 cells.
On the other hand, h128-3-induced antigenic modulation is Fcγ R-independent in Fcγ R low Mono-mac-6.A majority of the LILRB4 internalization observed in Mono-mac-6 cells is likely to be mediated by receptor clustering and clathrin-dependent endocytosis.

Fcγ RI is involved in LILRB4 internalization induced by h128-3
As Fcγ RI was expressed at moderate to high levels in both monocytic AML cell lines, we wanted to focus on this receptor and study the role of Fcγ RI independently.

The low-affinity Fcγ RIIa plays a role in h128-3-induced LILRB4 internalization
Fcγ RI clearly plays a significant role in the h128-3-induced internalization of LILRB4 on Fcγ R high monocytic AML THP-1 cells, but we could not rule out the possibility that Fcγ RIIa can also bind the h128-3 Fc and internalize the h128-3/LILRB4 complex.To characterize the potential role of Fcγ RIIa in this biological phenomenon, we generated stable Fcγ RI, Fcγ RIIa and dual Fcγ RI/Fcγ RIIa genetic knockouts in Fcγ R high THP-1 by CRISPR-Cas9.We specifically generated these knockouts using efficient lentiviral particle delivery of vector DNA encoding doxycycline-inducible Cas9 and Fcγ R-targeting sgRNAs.We used a vector with a non-targeting sgRNA (NT CTRL) to control for off-target effects of CRISPR-Cas9.Cas9 was doxycycline-induced to perform double-stranded breaks on genes specifically encoding Fcγ RI and/or Fcγ RIIa, and knockout cells were negatively selected by FACS (Fig. 4A-C).In Fcγ R high THP-1, h128-3-induced LILRB4 internalization upon Fcγ RI knockout at 24 h was significantly decreased relative to internalization in nontargeting control (NT CTRL) cells as expected (Fig. 4D).Interestingly, the h128-3-induced LILRB4 internalization in Fcγ RIIa KO THP-1 cells was similarly decreased relative to NT CTRL cells (Fig. 4E).In Fcγ RI/Fcγ RIIa DKO cells, h128-3-induced LILRB4 internalization was reduced relative to NT CTRL cells by a comparable 25-30% (Fig. 4F).LILRB4 internalization induced by h128-3 in Fcγ R high THP-1 can thus be mediated by independent contributions from Fcγ RI or Fcγ RIIa to achieve its maximal effect.

h128-3/LILRB4 internalization is not driven by Fcγ Rs on neighboring cells
Though studies of rituximab-induced Fcγ R/CD20 interactions have demonstrated that the antibody-antigen complex can be internalized by Fcγ R in both cis (Fcγ R on same cell) and trans (Fcγ R on neighboring cell) conformations [6][7][8][9], we sought to determine the contribution of cis and trans Fcγ Rs to h128-3-induced Fcγ R/LILRB4 interactions.We investigated the trans contribution to these interactions empirically using THP-1 Fcγ RI/Fcγ RIIa DKO (GFP − ) donor cells opsonized with 10 μg/ml h128-3/pHAb or hIgG/pHAb and THP-1 WT (GFP + ) acceptor cells.We incubated cells in co-culture at 37 • C, and pHAb fluorescence in the acidic intracellular compartment was measured by flow cytometry at 4 and 24 h.MFI in the far-red channel of GFP − (donor THP-1 DKO) and GFP + (acceptor THP-1 WT) was normalized to that of untreated co-cultured cells (Fig. 4G).Evidently, there are very low levels of internalization of h128-3/pHAb mediated by trans Fcγ Rs, while there is significant relative internalization occurring in the THP-1 Fcγ RI/Fcγ RIIa DKO cells.This indicates that h128-3-induced LILRB4 internalization is predominantly mediated by cis interactions, with minimal contributions from trans Fcγ Rs.   anti-hFcγ RIIa.As illustrated in Fig. 4I, there was a very low PLA colocalization signal detected on the NT CTRL cells stained with rabbit control IgG and mouse anti-hFcγ RIIa.On the NT CTRL cells stained with rabbit anti-hLILRB4 and mouse anti-hFcγ RIIa, there was also very low PLA signal, indicating that h128-3induced colocalization of LILRB4/Fcγ RIIa does not occur on the surface of these cells after 2 h.However, when we stained h128-3-treated Fcγ RI KO THP-1 cells with rabbit anti-hLILRB4 and mouse anti-hFcγ RIIa, we found that the PLA signal of LILRB4/Fcγ RIIa on these cells was relatively high.Analysis of five representative images each from three independent experiments demonstrated a significant increase in normalized LILRB4/Fcγ RIIa PLA colocalization signal intensity on Fcγ RI KO THP-1 as compared to NT CTRL.These results indicate that LILRB4 and Fcγ RI are rapidly colocalized by h128-3 in Fcγ R high THP-1 prior to h128-3/LILRB4 complex internalization but LILRB4 and Fcγ RIIa are not.LILRB4 and Fcγ RIIa colocalization does occur, but it occurs more slowly unless Fcγ RI is prevented from interacting with LILRB4.

Interactions between LILRB4 and Fcγ RI are essential for optimal blocking function of the anti-LILRB4 mAb h128-3 and consequent T-cell-mediated cytotoxicity
As we have reported, the LILRB4 blocking antibody h128-3 inhibits monocytic AML development by multiple mechanisms, including reversal of LILRB4-mediated AML tissue infiltration and local T-cell suppression [5].However, when the Fc region of h128-3 is modified at the N297 residue to reduce Fcγ R-mediated effector functions, the antibody is not able to optimally induce these phenotypes in Fcγ R high monocytic AML xenograft mouse models [3,5].As others have demonstrated, Fcγ R on immune cells may serve as a scaffold for enhancing antibody-induced target antigen clustering and effector functions [10][11][12][13].
Fcγ RI binding of h128-3 may similarly guide the antibody to its target, allowing for optimal blocking of LILRB4.
To determine if this occurs, we serum-starved Fcγ R high THP-1 cells to cell-cycle-synchronize them and specifically activated LILRB4 on these cells by receptor clustering using a plate-bound LILRB4 ligand (ApoE, 5 μg/ml).
In the presence of LILRB4 ligand and a plate-bound immune stimulant (anti-HLA-DR mAb L243, 5 μg/ml) that can cluster and activate a neighboring ITAM-bearing receptor (HLA-DR), there is tyrosine phosphorylation of LILRB4 that can be detected by western blot of LILRB4 proteins immunoprecipitated from these stimulated THP-1 cells.We first incubated serum-starved THP-1 cells with the LILRB4-blocking antibody h128-3 (10 μg/ml) for 1 h at 37 • C to induce h128-3-mediated interactions between LILRB4 and Fcγ R, before seeding the cells on a plate with pre-bound ApoE and anti-HLA-DR antibody for 15 min at 37 • C. The cells were gently lysed in the presence of protease and phosphatase inhibitors at 4 • C, and LILRB4 was immunoprecipitated.As expected, western blot detection of tyrosine phosphorylation on LILRB4 immunoprecipitated from the stimulated THP-1 cells pretreated with h128-3 decreased relative to that of cells pre-treated with isotype control hIgG (Fig. 5A).After confirming that tyrosine phosphorylation on LILRB4 was decreased by h128-3, we sought to determine if Fcγ RI-scaffolding would improve the blocking function of the antibody.We performed three independent LILRB4 stimulation assays, this time using NT CTRL or Fcγ RI KO THP-1 cells.Each cell line was serum-starved to induce cell cycle arrest at G0/G1, incubated with h128-3 for 1 h at 37 • C and seeded on a plate with pre-bound ApoE and anti-HLA-DR antibody for 15 min at 37 • C before gentle cell lysis and immunoprecipitation of LILRB4 protein from each antibody-treated cell population.Each time, the addition of h128-3 to NT CTRL THP-1 cells impeded tyrosine phosphorylation on immunoprecipitated LILRB4 relative to isotype control hIgG as anticipated.However, in Fcγ RI KO THP-1 cells, h128-3 pre-treatment could not prevent tyrosine phosphorylation relative to the isotype control (Fig. 5B).This indicated that Fcγ RI scaffolding plays a crucial role in the blocking function of the anti-LILRB4 mAb h128-3 and consequently, its ability to influence Tcell activation and cytotoxicity.To validate this, we isolated CD3 + T-cells from healthy donor PBMC and expanded them for 48 h in media enriched with anti-CD3/CD28 agonistic antibodies and IL-7/IL-15 cytokines.As depicted in Fig. 5C, we then treated NT CTRL or Fcγ RI KO THP-1 cells with h128-3 or hIgG isotype control (10 μg/ml) or Staphylococcal enterotoxin B (SEB, 2 μg/ml), a positive control superantigen treatment that induces T-cell hyperactivation and near-total annihilation of neighboring cells [25].After treatment and T-cell co-culture for 24 h, we assessed T-cell-mediated cytotoxicity of each treatment population relative to that of untreated co-cultured cells.
In support of our earlier findings, h128-3 stimulated potent T-cell-mediated cytotoxicity of NT CTRL THP-1 cells in six biologically independent experiments at E:T ratios of 4:1 (35-65%) and 8:1 (35-70%).Though cytotoxicity of Fcγ RI KO THP-1 co-cultured with expanded CD3 + T cells was observed with h128-3 treatment, the observed cytotoxicity was diminished relative to the cytotoxicity of NT CTRL cells at E:T ratios of 4:1 (15-45%, p = 0.1054) and 8:1 (5-45%, p = 0.0068).This demonstrated that Fcγ RI on monocytic AML supports h128-3's ability to influence T-cell activation and cytotoxicity of these malignant cells.Fcγ RI binding of h128-3 can scaffold the antibody to its LILRB4 target and enhance blocking of its immunosuppressive signaling.Reduction of this inhibitory signal ultimately promotes activation of T cells in the tumor immune microenvironment and T-cell-mediated cytotoxicity of malignant cells.

DISCUSSION
Despite increased understanding of the underlying biology of AML, the standard intervention of non-targeted cytotoxic chemotherapy followed by consolidative therapy such as bone marrow transplant has not changed in the past 40 years.As many as 70% of patients 65 years or older die of their disease within a year of diagnosis [26,27].Moreover, immune checkpoint inhibitor therapeutics, such as those targeting CTLA-4 and PD-1/PD-L1, have not yielded clinical benefits in AML patients with weakened immune function [28,29].Recently, our group and others identified LILRB4 as a surface marker for the monocytic AML subtype as it is expressed at significantly higher levels on these cells than on their normal counterparts [3,30].Most importantly, LILRB4 expression is inversely correlated with the overall survival of patients with monocytic AML.Binding of ApoE, one of the functional ligands of LILRB4, is coupled with T-cell suppression and tumor infiltration through LILRB4-mediated downstream signaling in AML cells [3].LILRB4 is thus a promising target for antibody-based therapies for monocytic AML.
We have developed a humanized LILRB4 blocking mAb h128-3, which disrupts the interaction of LILRB4 with ligands including ApoE [5].h128-3 showed potent activity in blocking the development of monocytic AML in several models, including a xenograft mouse model, a syngeneic immunocompetent mouse model and a PDX mouse model of disseminated AML.We further demonstrated that h128-3 blocks monocytic AML development through reversal of T-cell suppression, inhibition of AML tissue infiltration, ADCC and ADCP [5].However, in monocytic AML PDX mouse models, we found downregulation of surface LILRB4 following treatment with h128-3 [3].This phenomenon indicated that h128-3 may induce LILRB4 internalization in an Fcγ R-mediated fashion.For the receptor CD20 on B-cell lymphoma, Fcγ RIIb-mediated CD20 internalization with rituximab treatment decreases therapeutic efficacy in vivo [7].Fcγ R-mediated LILRB4 internalization with h128-3 treatment may represent a different case.As LILRB4 signaling is critical for immune suppression and AML cell tissue infiltration, LILRB4 internalization and degradation induced by h128-3 may block AML cell migration and reverse T-cell suppression permanently.This internalization of LILRB4 induced by h128-3 also suggests that treatment with h128-3-based ADCs may be a useful strategy to deplete monocytic AML cells [31].
Antibody-induced receptor internalization is a commonly occurring biological phenomenon in many cell types that can occur through receptor dimerization and clathrin-mediated endocytosis [32][33][34].In this study, we discovered that Fc and Fcγ R interaction promotes strong h128-3-induced LILRB4 internalization.Disrupting the interaction of Fc and Fcγ Rs by pre-incubation of monocytic AML cells with control hIgG or by CRISPR-Cas9 genetic knockout of Fcγ RI or Fcγ RIIa reduced LILRB4 internalization in Fcγ R high cells such as THP-1.These results are similar to those from the rituximab studies.Rituximab-mediated interactions between CD20 and Fcγ RIIb on B-cell lymphoma increased the internalization of the rituximab/CD20 immune complex and prevented immune regulation of rituximab-opsonized Bcell lymphoma [6][7][8][9].As monocytic AML is derived from monocytes, it canonically expresses moderate to high levels of Fcγ RI and Fcγ RIIa and low levels of Fcγ RIIb.We confirmed this with flow cytometry, detecting very low levels of Fcγ RIIb and Fcγ RIII on monocytic AML cells, which is congruent with other studies [17,35,36].Fcγ RIII is expressed on NK cells and neutrophils and is responsible for the ADCC effector functions of these cells [37].We further demonstrated that scaffolding of h128-3 by Fcγ RI contributes to the internalization of LILRB4.Scaffolding of h128-3 by Fcγ RIIa also contributes to LILRB4 internalization in the absence of functional highaffinity Fcγ RI, which may be saturated by circulating IgG in the physiologic setting.Our results suggest that h128-3 can induce interactions between LILRB4 and Fcγ RI or Fcγ RIIa on monocytic AML cells, resulting in internalization of the h128-3/LILRB4 complexes and irreversible degradation of LILRB4.However, when THP-1 cells were cultured with h128-3/N297A, a low level of Fcγ R-independent h128-3-induced LILRB4 internalization was observed.Substitution of glutamine at position 297 to alanine (h128-3/N297A) decreased antibody binding to Fcγ Rs.However, this mutated antibody retains some ability to bind Fcγ Rs [38] and may have retained some Fcγ R-mediated internalization of LILRB4 relative to isotype control hIgG.
Depending on the Fcγ R that is scaffolding the h128-3 mAb, the effects on immune regulation may vary.Fcγ RI and Fcγ RIIa, the Fcγ Rs that scaffold and internalize the h128-3/LILRB4 complex on monocytic AML cells, both have intracellular immunoreceptor tyrosine-based activating motifs (ITAM) that canonically drive the expression and release of pro-inflammatory cytokines and chemokines (e.g.TNF, IL-1β and IL-8) that alter the effector function, migration and survival of leukocytes [17,36].Fcγ RIIb, the only inhibitory receptor, is expressed on many types of immune cells including B cells, DCs, monocytes, macrophages, mast cells and basophils [39].The inhibitory Fcγ RIIb has an intracellular ITIM domain and recruits phosphatases to its ITIM following receptor crosslinking and phosphorylation by Src family kinases.These ITIM-recruited phosphatases, such as SHIP-1 and SHP-2, canonically inhibit immune-activating signals.
Crosslinking of Fcγ RIIb on effector cells by antibody Fc dampens activation of immune cells induced by activating Fcγ Rs through recruitment of these phosphatases [40][41][42].However, in some cases, crosslinking of Fcγ RIIb on immune cells by antibody Fc increases the efficacy of antibodies [10][11][12][13].In one of these cases, enhancing the binding affinity of Fc and Fcγ RIIb by mutation of some key amino acid residues at the interface increased the therapeutic efficacy of anti-CD40 antibodies in vivo [13].
Here, Fcγ RIIb expressed on tissue-infiltrating immune cells served as a scaffold for enhancing antibody-induced target antigen clustering and cellular effector function.As summarized in Fig. 6, LILRB4-targeting mAb h128-3 takes advantage of similar Fcγ R-mediated scaffolding effects to bind its target receptor and internalize it.This mechanism leads to improved antagonism of the target inhibitory receptor and activation of effector T cells in the tumor immune microenvironment.
Biologically, our study showcases an example that the functional modification of an inhibitory receptor can be modulated by a local activating receptor.There are several potential directions for future development of LILRB4based antibody therapeutics.To begin with, we need to balance the antibody-induced LILRB4 internalization on target cells and the effects on immune cells mediated by crosslinking of Fcγ Rs by antibody Fc.For naked therapeutic IgG 1 mAb development, increasing the interaction of Fc with Fcγ RI/Fcγ RIIa on malignant cells or Fcγ RIII on immune effector cells by generating Fc-engineered mutants or Fc-modified afucosylated variants [43] may be good choices.For development of LILRB4-targeting ADCs [31], we may need to disrupt the paradigm and increase or maintain interactions of Fc with Fcγ RI/Fcγ RIIa to obtain significant LILRB4 internalization.In the long run, defining the mechanisms underlying the interaction of therapeutic antibody Fc and Fcγ Rs as we have reported here will help us to develop therapeutics targeting LILRB4 and other immune checkpoints on myeloid malignancies with greater potency and efficacy.

Figure 4 .
Figure 4. (Continue) (H, I) THP-1 NT CTRL and Fcγ RIIa KO cells (n = 2) (H) or THP-1 NT CTRL and Fcγ RI KO cells (n = 3) (I)were treated with h128-3 (5 μg/ml; 2.5 × 10 5 cells) for 2 h and fixed on slides.Fixed cells were blocked for 1 h at RT, stained for 18 h at 4 • C with mouse anti-hFcγ RI/anti-hFcγ RIIa and control Rabbit anti-hIgG (all 5 μg/ml) or mouse anti-hFcγ RI/anti-hFcγ RIIa and rabbit anti-hLILRB4 (all 5 μg/ml) and then stained with the Sigma Duolink Orange (Cy3) PLA kit.Nuclei were stained with ToPro3 and slides were sealed with ProLong Gold Antifade Mountant (ThermoFisher) before imaging with the 63× objective of a Leica TCS SP5 confocal microscope.Five representative images were collected for quantitative analysis of each sample.Image analysis was performed with Leica LAS X software.PLA colocalization signal (Cy3) intensity was normalized to nuclear stain (ToPro3) intensity on all samples.Experimental sample intensities were normalized to mean IgG control sample intensity.

Figure 5 .
Figure 5. Fcγ RI crosslinking enhances the LILRB4 blocking function of h128-3, leading to improved T-cell-mediated cytotoxicity of Fcγ R high monocytic AML.(A, B) THP-1 WT (A), NT CTRL and Fcγ RI KO (B) cells (1 × 10 7 ) were serum-starved for 18 h and then incubated with PBS, hIgG or h128-3 (10 μg/ml) for 1 h at 37 • C. The cells were then seeded on non-treated tissue culture plates with pre-bound ApoE2 (5 μg/ml, Peprotech) and anti-HLA-DR antibody (5 μg/ml, L243, Biolegend) for 15 min at 37 • C to stimulate rapid LILRB4 activation before lysis at 4 • C in the presence of protease and phosphatase inhibitors.LILRB4 was immunoprecipitated from the lysed cells overnight at 4 • C using a high-affinity anti-LILRB4 antibody (R8), and western blot was run on the immunoprecipitated LILRB4.The membrane was probed overnight at 4 • C with anti-pTyrosine antibody (4G10, Cell Signaling), stripped with mild stripping buffer and then re-probed with anti-LILRB4 (R8) for 1 h at RT for loading comparison.(C) To determine T-cell-mediated cytotoxicity induced by h128-3, CD3 + T cells were first isolated from healthy donor PBMCs by negative selection and expanded for 48 h at 37 • C in medium enriched with ImmunoCult Human CD3/CD28 T-Cell Activator (25 μl/10 6 cells/ml, Stemcell Technologies), rIL-7 and rIL-15 (10 ng/ml, Peprotech).GFP + THP-1 NT CTRL and Fcγ RI KO cells were then seeded in 96 well U-bottom plates in normal R10 media (untreated) or R10 supplemented with isotype control hIgG (10 μg/ml), h128-3 (10 μg/ml) or Staphylococcus enterotoxin B (2 μg/ml) for 15 min at 37 • C. Expanded T cells were then seeded in co-culture with the monocytic AML cells at E:T ratios of 1:1, 4:1 and 8:1 for 24 h at 37 • C. The co-cultured cells were washed with 2% BSA/PBS and live/dead-stained with DAPI before analysis by flow cytometry.Live target THP-1 cells (GFP + DAPI − ) were gated and antibody-or toxin-mediated T-cell cytotoxicity of NT CTRL or Fcγ RI KO THP-1 cells was calculated relative to T-cell cytotoxicity of untreated NT CTRL or Fcγ RI KO THP-1 cells, respectively.