Phosphatidylserine (PS)-targeting monoclonal Abs (mAbs) that directly target PS and target PS via β2-gp1 (β2GP1) have been in preclinical and clinical development for over 10 y for the treatment of infectious diseases and cancer. Although the intended targets of PS-binding mAbs have traditionally included pathogens as well as stressed tumor cells and its associated vasculature in oncology, the effects of PS-targeting mAbs on activated immune cells, notably T cells, which externalize PS upon Ag stimulation, is not well understood. Using human T cells from healthy donor PBMCs activated with an anti-CD3 + anti-CD28 Ab mixture (anti-CD3/CD28) as a model for TCR-mediated PS externalization and T cell stimulation, we investigated effects of two different PS-targeting mAbs, 11.31 and bavituximab (Bavi), on TCR activation and TCR-mediated cytokine production in an ex vivo paradigm. Although 11.31 and Bavi bind selectivity to anti-CD3/28 activated T cells in a PS-dependent manner, surprisingly, they display distinct functional activities in their effect on IFN-γ and TNF-ɑ production, whereby 11.31, but not Bavi, suppressed cytokine production. This inhibitory effect on anti-CD3/28 activated T cells was observed on both CD4+ and CD8+ cells and independently of monocytes, suggesting the effects of 11.31 were directly mediated by binding to externalized PS on activated T cells. Imaging showed 11.31 and Bavi bind at distinct focal depots on the cell membrane. Collectively, our findings indicate that PS-targeting mAb 11.31 suppresses cytokine production by anti-CD3/28 activated T cells.

The asymmetric distribution of lipids across the inner and outer leaflets of the plasma membrane, characterized by the distribution of anionic phospholipids, such as phosphatidylserine (PS), phosphatidylethanolamine (PE), and phosphatidylinositol on the inner leaflet and phosphatidylcholine and sphingomyelin on the outer leaflet, is important for membrane homeostasis and for the proper functioning of proteins embedded in the membrane or bound to the extracellular or intracellular leaflets (1, 2). Membrane asymmetry and homeostasis are irreversibly breached during apoptosis as a result of caspase-dependent cleavage events, notably the activation of the PS scramblase Xkr8 (3) as well as the inactivation of P4 subfamily ATPase PS flippases, such as ATP11A and ATP11C, the latter of which incapacitates the relocalization of PS to the inner surface (47). The net irreversible relocalization of PS to the outer membrane surface provides an “eat me” signal for efficient efferocytosis that prevents the release of intracellular contents that can serve as inflammatory danger signals and a source for self-autoantigens (8). Externalized PS on apoptotic cells is also associated with immune modulatory events associated with the production of anti-inflammatory cytokines and resolving factors that maintain tissue tolerance.

PS is also reversibly externalized on activated healthy viable cells that serve important biochemical and physiological functions. For example, on thrombin-activated platelets, PS is externalized via the TMEM16 family of calcium-activated phospholipid scramblases to facilitate recruitment of coagulation factors (9). On activated CD8+ T lymphocytes, PS is externalized following TCR stimulation, which is important for both productive TCR signaling as well as downregulation of T cell activation (2, 914). In support of this latter idea, targeting externalized PS on Ag-stimulated CD8+ T cells with Annexin V diminished TCR-mediated induction of cytokines (15), and studies by Rothlin and colleagues (16) showed that PS externalization on activated T cells recruits a receptor-bridging molecule, Protein S, that engages a PS-receptor (Tyro-3) (17), on APC to control the magnitude of Ag activation of T cells (16).

Pathophysiologically, PS is constitutively externalized on both metabolically stressed cancer cells and in cells following chronic viral infections (8, 18, 19). In solid cancers, often called a “wound that never heals,” chronic PS exposure has been shown to occur constitutively in the tumor mass on viable tumor cells, stroma, and vascular endothelial cells, presenting a strong focal immune suppressive event and likely contributing to immune escape mechanisms and tumor progression (8, 20). As a result of this constitutively externalized PS in tumors, PS-targeting mAbs, and more recently PS-targeting biologicals, have been in development to localize to tumor microenvironments (2123). Such approaches led to the identification and characterization of a β2-glycoprotein (β2GPI)–specific mAb 3G4 (mouse IgG3) (24) that recognizes PS and specifically localizes to a variety of tumors and their associated vasculatures (2528). The variable fragment of 3G4 was later synthesized into a human IgG1 backbone yielding “bavituximab” (Bavi), a mouse-human chimeric IgG1 Ab, which became the keystone of Peregrine Pharmaceuticals and a variety of clinical trials for the treatment of solid cancers (2933), often producing a more-inflamed tumor microenvironment including the polarization of suppressive M2 macrophages toward an M1 phenotype (34) and promoting a T cell–mediated host antitumor response.

In addition to Bavi, direct binding types of PS-targeting mAbs have been identified that have potential antiviral activity (35), expanding on the idea that distinct repertoires and types of natively occurring PS-targeting mAbs can exist (36). One such mAb, PGN632 or “11.31,” isolated from a healthy donor phage library, was characterized as a direct PS-binding (β2GPI-independent) mAb with differing properties from Bavi described above. Although it was hypothesized that 11.31 mAb might opsonize the PS+ HIV-1 envelope and prevent viral entry by directly blocking PS, additional studies showed that 11.31 bound to viable PS-positive monocytes and elicited a unique chemokine secretion that competed with HIV-1 for binding to the chemokine receptor CCR5, the main receptor implicated in CCR5 T-tropic HIV-1 infection (37). Such observations suggest diversification in the biological activities and functions of PS-targeting mAbs, implying potential unique functions in oncology or in antiviral applications. In this study, we compared effects of two distinct PS-targeting mAb (11.31 and Bavi) on the modulation of T cell responses using ex vivo stimulation of human PBMCs from healthy donors via anti-CD3/28 to address whether PS-targeting mAbs may have direct checkpoint-like activity on T cells. Results indicate distinct binding profiles and distinct abilities to downregulate T cell responses following anti-CD3/28 stimulation. These studies affirm that different functional repertoires of PS-targeting mAbs may exist endogenously and that PS-targeting mAbs, depending on their mechanism of PS binding, will have distinct clinical utilities.

l-PS (brain, porcine) was purchased from Avanti Polar Lipids (catalog no. 840032P and 840012) and dissolved in methanol. β2 glycoprotein was purchased from Hematological Technologies (catalog no. B2GI-0001), and lyophilized protein was resuspended in PBS. Recombinant β2GPI was produced by transfecting Expi293T cells with APOH-Bio-HIS (no. 52177; Addgene), and proteins in the supernatant were purified via HisPur resin. The 96-well medium binding ELISA plates (Greiner Bio-One) were coated with PS (12.5 μg/ml) at room temperature and placed in a sterile fume hood until evaporated. β2GPI plates were coated at a concentration of 2 μg/ml in NaHCO3 and placed overnight at 4°C. PS ELISA plates were then blocked with 5% BSA in PBS overnight at 4°C while β2GPI plates were blocked for 2 h at 37°C. All PS/β2GPI-coated plates were made in parallel with their solvent buffer (blank) coated plates. Plates coated with PS were washed three times with lipid-wash buffer (10 mM HEPES, 150 mM NaCl, 2.5 mM CaCl2) while plates coated with β2GPI were washed with PBS + 0.1% Tween 20 and incubated with PS-targeting mAbs in 5% BSA for 2 h at 37°C. Plates were washed three times and incubated with secondary Abs (anti-human H chain and L chain) conjugated to alkaline phosphatase for 1 h at 37°C. Plates were washed three times in corresponding buffers and incubated in alkaline phosphatase substrate (catalog no. S0942; Sigma) dissolved (1 mg/ml) in diethanolamine (DEA) buffer. Absorbance was measured at an OD of 405 nm over the course of 2 h.

Human PBMCs were obtained from healthy donor Leukopaks from the New York Blood Center. PBMCs were isolated via Ficoll density gradient and frozen in 30 × 106 cells per milliliter aliquots in liquid nitrogen. For T cell activation assays, cells were thawed and placed into wells containing 1.5 × 106 cells per milliliter in RPMI 1640, supplemented by 10% FBS (heat inactivated) and 2% l-glutamine. Cells were activated with a mixture of Abs specific for CD3 (catalog no. BDB555339; BD) and CD28/49d Fastimmune (catalog no. 347690; BD) for 6 h, unless specified otherwise. Aforementioned activation assays in certain instances included treatments with PHA (Invivogen) or PMA and ionomycin (eBioscience)). During activation studies in which intracellular cytokines were analyzed, GolgiPlug (catalog no. 555029; BD) was incubated in the media to prevent the secretion of cytokines, and staining protocol was used per BD Biosciences protocol. Prior to staining cells, Human TruStain FcX (catalog no. 422301; BioLegend) Fc block reagent was used. Human Abs used for flow cytometric studies are as follows: anti-CD3 (catalog no. BDB555339; BD), anti-CD4 (catalog no. BDB562429; BD), anti-CD8 (catalog no. BDB562316; BD), anti–IFN-γ, anti–TNF-ɑ (catalog no. IC9677S100; R&D Systems), and anti-CD14 (catalog no. 301805-BL; BioLegend). A total of 1 × 105 PBMCs were acquired and gated on according to specific aims of the experiment. Experiments were performed in triplicate, unless otherwise specified, and data were normalized to stimulated controls. Staining of PS on the surface of activated T cells was performed via Annexin V (catalog no. 559763; BD) or fluorescently labeled PS-targeting mAbs using Alexa Fluor 647 Ab Labeling kit (catalog no. A20186; Thermo Fisher Scientific). Monocyte depletion studies were performed by thawing PBMCs and leaving them overnight at 37°C. Depletion was confirmed following staining with anti-CD14 and analyzation via flow cytometry, and any remaining monocytes were removed by FACS. T cell purification was performed by negative magnetic separation (catalog no. 130-096-535; Miltenyi), and in certain cases PBMCs were stained with anti-CD3 and FACS sorted for imaging purposes. Cells were placed at 37°C overnight to allow recovery, and stimulations were subsequently followed with IL-2 (catalog no. 200-02; Peprotech) supplemented media. mRNA flow cytometry was performed as described in Vir et al. and Bushkin et al. (38, 39), and cells were stained 4 h after anti-CD3/CD28 stimulation. For proliferation studies, PBMCs were stained with Cell Trace Far Red (catalog no. C34564; Thermo Fisher Scientific) and stimulated with anti-CD3/28 in the presence or absence of PS mAbs, as well as PHA as a positive control, for 72 h. Proliferation was directly measured via flow cytometry by gating on CD3+/7AAD viable T cells, and loss membrane dye fluorescence was quantified as increased proliferation. For the experiments involving human Jurkat cell lines, cells were cultured in DMEM containing 5% FBS and stimulated with either anti-CD3/28 or calcium ionophore as described in Supplemental Fig. 3.

Human PBMCs were stimulated with CD3/CD28 for 4 h at 37°C. Fc block was added for 5 min, and the cells were then stained for 30 min at 4°C with anti-CD3. The cells were washed twice in flow staining buffer (PBS + 1% FBS) and incubated with either 1) Recombinant Annexin V + PE-conjugated 11.31, 2) PE-conjugated Annexin V + unconjugated 11.31, 3) Recombinant Annexin V + PE-conjugated Bavi, or 4) PE-conjugated Annexin V + unconjugated Bavi. Binding of each Ab/Annexin V to the activated CD3 cells and the amount of competition were assayed for using flow cytometry.

Human PBMCs were activated with anti-CD3/CD28 for multiple time points mentioned in Results. AF-647–conjugated 11.31, biotinylated 11.31, and/or AF647-conjuagted Bavi were added to the cells prior to fixation. After incubations with PS mAbs and staining with anti-CD3, the cells were washed twice in flow staining buffer, and then total T cells were sorted via FACS on total CD3. CD3+ cells were then fixed in 4% paraformaldehyde for 20 min at room temperature, and cells were quickly washed in dH2O. Cells were then resuspended in dH2O, smeared onto a gelatin-coated slide, and placed on a hot plate until all liquid had evaporated. Cells were blocked in 10% donkey serum, 0.3% Triton X-100. Cells were washed twice (1× PBS, 0.1% BSA) and counterstained with DAPI (1 μl of 14.5 mM stock for every 5 ml of PBS). Slides were rinsed once with PBS and then once with water, antifade mounting solution was applied, and cells were visualized via confocal microscopy.

All statistical analysis was performed using GraphPad Prism version 7.00 for Windows (GraphPad Software, La Jolla, CA). The comparison of all parameters was performed using an unpaired two-tailed Student t test, unless stated otherwise. A p value ≤0.05 was deemed to be significant for the above studies.

Previous preclinical studies using PS-targeting mAbs 11.31 and Bavi have shown promising therapeutic effects (20, 37). Bavi has been shown to bind constitutively externalized PS on metabolically stressed tumor cells and PS-positive tumor vascular endothelial cells (24, 29). In contrast, 11.31 exhibits an antiviral effect indirectly by interacting with monocytes (37). PS is also externalized on the surface of activated T cells, but it is presently unclear whether PS-targeting mAbs might block inhibitory signals on T cells possibly akin to checkpoint inhibitors like anti–PD-1 and anti–CTLA-4 mAbs (40). To investigate the effects of 11.31 versus Bavi in functional side-by-side studies on activated T cells in vitro, we first synthesized the unique H and L chain variable sequences obtained from the published patent sequences of 11.31 (VH, Vλ; US20110318360A1) and Bavi (VH, Vκ; WO2018064013A1) and expressed in a human IgG1 backbone containing the corresponding constant H and L chain regions (Fig. 1A). Resulting recombinant mAbs were expressed in the Expi293 system and purified using protein A affinity chromatography, with >95% purity assessed by gel electrophoresis and Coomassie blue staining. Subsequently, the purified 11.31 and Bavi recombinant mAbs were assayed for reactivity to PS via direct binding ELISA assays in the presence or absence of FBS (containing β2GPI). As indicated in (Fig. 1B, although both 11.31 and Bavi bound microtiter plates in a PS-dependent manner, β2GPI was required for Bavi, but not 11.31 binding, consistent with previous studies (24, 37). Moreover, we confirmed that Bavi interactions were indeed mediated by β2GPI by recombinantly producing and purifying human β2GPI (no. 52177; Addgene) (41) and performed PS ELISA (Supplemental Fig. 1A, 1B).

FIGURE 1.

Differential recognition of direct binding PS Ab, 11.31, versus indirect binding (via β2GPI) PS Ab Bavi. (A) Schematic representation of 11.31 and Bavi binding properties; 11.31 binds directly to PS, whereas Bavi requires β2GPI as a bridging molecule. (B) Sandwich ELISA on PS-coated plates with serially diluted PS-binding mAbs 11.31 and Bavi in the presence or absence of β2GPI (1% FBS); values represented at 30 min postsubstrate addition. (C) Flow cytometric analysis of PS mAbs binding cell line CDC50AED29 in the presence or absence of β2GPI (1% FBS). (D) Flow cytometric gating scheme of viable and apoptotic T cells using a Caspase 3/7 probe, CellEvent Green. Cells that are CellEvent Green negative harbor caspase in its inactive form and are nonapoptotic. (E) Four-hour stimulation of T cells with anti-CD3/CD28 ligation of TCR. Caspase inactive cells were gated, and PS externalization was measured by Annexin V-PE. (F) Comparison of PS externalization on Caspase 3/7 positive (apoptotic) T cells versus viable T cells stimulated with anti-CD3/CD28. PS externalization measured via Annexin V-PE. (G) Percentage of CD3+ cells in PBMCs externalizing PS and (H) relative MFI of PS-staining following a 2-h stimulation of T cells in PBMCs using anti-CD3/CD28. PS externalization measured by Annexin V-PE and labeled PS-targeting mAbs 11.31-PE and Bavi-AF647. Significance determined between treatment group and stimulation alone by Student t test. *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005.

FIGURE 1.

Differential recognition of direct binding PS Ab, 11.31, versus indirect binding (via β2GPI) PS Ab Bavi. (A) Schematic representation of 11.31 and Bavi binding properties; 11.31 binds directly to PS, whereas Bavi requires β2GPI as a bridging molecule. (B) Sandwich ELISA on PS-coated plates with serially diluted PS-binding mAbs 11.31 and Bavi in the presence or absence of β2GPI (1% FBS); values represented at 30 min postsubstrate addition. (C) Flow cytometric analysis of PS mAbs binding cell line CDC50AED29 in the presence or absence of β2GPI (1% FBS). (D) Flow cytometric gating scheme of viable and apoptotic T cells using a Caspase 3/7 probe, CellEvent Green. Cells that are CellEvent Green negative harbor caspase in its inactive form and are nonapoptotic. (E) Four-hour stimulation of T cells with anti-CD3/CD28 ligation of TCR. Caspase inactive cells were gated, and PS externalization was measured by Annexin V-PE. (F) Comparison of PS externalization on Caspase 3/7 positive (apoptotic) T cells versus viable T cells stimulated with anti-CD3/CD28. PS externalization measured via Annexin V-PE. (G) Percentage of CD3+ cells in PBMCs externalizing PS and (H) relative MFI of PS-staining following a 2-h stimulation of T cells in PBMCs using anti-CD3/CD28. PS externalization measured by Annexin V-PE and labeled PS-targeting mAbs 11.31-PE and Bavi-AF647. Significance determined between treatment group and stimulation alone by Student t test. *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005.

Close modal

Furthermore, to assess whether 11.31 and Bavi retained binding characteristics in a cell-based assay, we examined binding of 11.31 and Bavi to viable CDC50AED29 cells, a transformed T cell line that constitutively exposes PS because of a mutation in the catalytic subunit of ATP11C, rendering the flippase activity of ATP11C defective, thereby allowing for constitutive externalization of PS (42). Similar to the results of the solid-phase ELISA, when 11.31 and Bavi were incubated in the presence or absence of β2GPI (using RPMI media with and without 1% FBS as a source of β2GPI), affirmative staining by flow cytometry showed that although both mAbs bound to CDC50AED29 cells, only Bavi required the supplementation of β2GPI via FBS (Fig. 1C). These data suggest that cloned 11.31 and Bavi, in a human chimeric IgG1 vector, retain their biochemical and specificity characteristics as “direct” and “indirect” PS-targeting agents, respectively (24).

Next, we assessed whether 11.31 and Bavi bound to externalized PS on activated T cells following Ag-independent TCR stimulation. Primary human T cells from healthy donor PBMCs were stimulated with anti-CD3/CD28 mAb mixture, an Ab combination that crosslinks TCRs on CD3+ T cells. To distinguish live versus apoptotic T lymphocytes in the binding assays and assure that PS-targeting mAbs bound live T cells, anti-CD3/CD28–stimulated cells were concomitantly incubated with CellEvent Green probe, which measures real-time caspase activity. As shown in (Fig. 1D, CD3-allophycocyanin positive/CellEvent Green negative (gate G2, caspase population)–stimulated T cells externalized PS, as assessed by Annexin V staining, in a caspase-independent manner. In addition to binding to activated, viable cells, Annexin V also bound to PS-positive dying T cells, albeit more intensely to the live cell population, indicating that the level of PS externalized on dying (apoptotic) T cells is higher than on activated viable T cells. PS externalization via anti-CD3/CD28 stimulation peaks at 2 h and exhibits a reduced PS signature by 4 h (Fig. 1E). As shown in (Fig. 1F and 1H, Annexin V, 11.31, and Bavi all preferentially bound to activated T cells externalizing PS in a similar quantitative fashion. Thus, TCR stimulation results in the caspase-independent transient externalization of PS on T cells that can be targeted by PS mAbs.

We followed up binding assays by testing whether 11.31 and Bavi binding affects TCR-dependent signaling. A previous study indicated that binding of Annexin V on the surface of activated T cells could partially attenuate T cell activation (15), suggesting that PS externalization and targeting by PS binding factors have precedence to impinge on TCR signaling. PBMCs were pretreated with either 11.31 or Bavi, after which cells were stimulated with anti-CD3/CD28 to activate TCR (Fig. 2A–H,). Subsequently, levels of intracellular IFN-γ and TNF-ɑ were assessed cytofluorimetrically. Interestingly, pretreatment with 11.31, but not Bavi, suppressed T cell activation (measured by IFN-γ and TNF-ɑ expression in CD3+ T cells) and cytokine production (measured by IFN-γ and TNF-ɑ mean fluorescence intensity [MFI]) (Fig. 2A–C). Similar results were observed when PBMCs were stimulated by anti-CD3/CD28 conjugated beads (as distinct from soluble anti-CD3/CD28 Ab mixture; (Fig. 2D, 2E) and with the PBMCs of four healthy donors (Fig. 2F). Moreover, the inhibitory function of 11.31 was not T cell subtype specific, whereby both CD4+ Th cells and CD8+ cytotoxic T cells were equally affected in this assay (Fig. 2G, 2H). The inhibition of cytokine expression by 11.31 was dependent on the dose (Fig. 3A, 3B) and the time of addition (Fig. 3C, 3D) of Ab. Furthermore, when we stimulated the T cells for 20 h in the presence or absence of 11.31, we found that the number of T cells producing IFN-γ and TNF-ɑ (Fig. 3E–G) as well as the level of cytokine production, and inhibition thereof by 11.31 (Fig. 3H, 3I), were similar to the 6-h timepoints. That the level of 11.31-mediated inhibition was decreased with increasing delay of Ab addition relative to the initiation of anti-CD3/CD28 treatment suggests that 11.31 binding may affect an early step of T cell activation mediated by anti-CD3/28. Accordingly, we analyzed the production of IFN-γ mRNA transcripts in T cells using RNA flow cytometry (FISH-Flow) (38, 39) following anti-CD3/CD28 stimulation and observed that 11.31 inhibition of cytokine transcription was consistent with translated cytokine protein detected intracellularly (Fig. 3J–L).

FIGURE 2.

Incubation with direct binding PS mAb 11.31 results in inhibition of anti-CD3/28 stimulation of T cells. (A) T cells in PBMCs were stimulated with soluble anti-CD3/CD28 in the presence or absence of PS-targeting mAbs (100 μg/ml) or human IgG1 isotype control (100 μg/ml), and production of IFN-γ and TNF-ɑ was analyzed via flow cytometry. In each case, samples labeled 11.31 and Bavi describe samples stimulated with anti-CD3/CD28 in the presence of the corresponding mAb. (B and C) Quantification of flow cytometric assays were represented in CD3+ cells producing cytokines (B) and MFI of cytokine expression (C) relative to anti-CD3/CD28–stimulated control, which was normalized to 1. (D and E) T cell stimulations were repeated with anti-CD3/CD28 conjugated beads. (F) T cell stimulation assays were repeated in four separate healthy donor PBMCs. (G and H) Effects of PS mAb treatment on CD4 and CD8 T cell subset activations and expression of IFN-γ and TNF-ɑ quantified and represented by CD3+/CD4+ or CD3+/CD8+ cells producing cytokines (G) and MFI of cytokine expression (H) relative to stimulated controls. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005, ****p ≤ 0.00005.

FIGURE 2.

Incubation with direct binding PS mAb 11.31 results in inhibition of anti-CD3/28 stimulation of T cells. (A) T cells in PBMCs were stimulated with soluble anti-CD3/CD28 in the presence or absence of PS-targeting mAbs (100 μg/ml) or human IgG1 isotype control (100 μg/ml), and production of IFN-γ and TNF-ɑ was analyzed via flow cytometry. In each case, samples labeled 11.31 and Bavi describe samples stimulated with anti-CD3/CD28 in the presence of the corresponding mAb. (B and C) Quantification of flow cytometric assays were represented in CD3+ cells producing cytokines (B) and MFI of cytokine expression (C) relative to anti-CD3/CD28–stimulated control, which was normalized to 1. (D and E) T cell stimulations were repeated with anti-CD3/CD28 conjugated beads. (F) T cell stimulation assays were repeated in four separate healthy donor PBMCs. (G and H) Effects of PS mAb treatment on CD4 and CD8 T cell subset activations and expression of IFN-γ and TNF-ɑ quantified and represented by CD3+/CD4+ or CD3+/CD8+ cells producing cytokines (G) and MFI of cytokine expression (H) relative to stimulated controls. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005, ****p ≤ 0.00005.

Close modal
FIGURE 3.

11.31 inhibition of anti-CD3/28 T cell stimulation occurs dose and time dependently. Dose- and time-dependent experiments of PS mAbs on T cells were performed to further characterize the effects of PS mAbs on T cell activation. T cells in PBMCs were stimulated in the absence or presence of 100 μg/ml, 10 μg/ml, or 1.0 μg/ml 11.31. (A and B) Quantification of flow cytometric assays were represented in CD3+ cells producing cytokines (A) and MFI of cytokine expression (B) relative to anti-CD3/28–stimulated control normalized to 1. Time-dependent studies performed at the following time points: 1 h prior to stimulation (−1 h), at the moment of stimulation (0 h), and 1 h, 2 h, and 4 h poststimulation. (C and D) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (C) and MFI of cytokine expression (D) relative to anti-CD3/28–stimulated control normalized to 1. (E) PBMCs were stimulated with soluble anti-CD3/CD28 for 20 h in the presence or absence of PS-targeting mAb 11.31 (100 μg/ml), and production of IFN-γ and TNF-ɑ was analyzed via flow cytometry. In each case, samples labeled 11.31 describe samples stimulated with anti-CD3/CD28 in the presence of the corresponding mAb. Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (F and G) and MFI of cytokine expression (H and I) relative to anti-CD3/CD28–stimulated control, which was normalized to 1. (J) T cells in PBMCs were stimulated via anti-CD3/28 in the absence or presence of 100 μg/ml 11.31 or Bavi for 4 h and assayed for cytokine mRNA expression via FISH-Flow. Quantification of flow cytometric assays was represented in CD3+ cells producing cytokine mRNA transcripts (K) and MFI of cytokine expression (L) relative to anti-CD3/28–stimulated control normalized to 1. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005, ****p ≤ 0.00005.

FIGURE 3.

11.31 inhibition of anti-CD3/28 T cell stimulation occurs dose and time dependently. Dose- and time-dependent experiments of PS mAbs on T cells were performed to further characterize the effects of PS mAbs on T cell activation. T cells in PBMCs were stimulated in the absence or presence of 100 μg/ml, 10 μg/ml, or 1.0 μg/ml 11.31. (A and B) Quantification of flow cytometric assays were represented in CD3+ cells producing cytokines (A) and MFI of cytokine expression (B) relative to anti-CD3/28–stimulated control normalized to 1. Time-dependent studies performed at the following time points: 1 h prior to stimulation (−1 h), at the moment of stimulation (0 h), and 1 h, 2 h, and 4 h poststimulation. (C and D) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (C) and MFI of cytokine expression (D) relative to anti-CD3/28–stimulated control normalized to 1. (E) PBMCs were stimulated with soluble anti-CD3/CD28 for 20 h in the presence or absence of PS-targeting mAb 11.31 (100 μg/ml), and production of IFN-γ and TNF-ɑ was analyzed via flow cytometry. In each case, samples labeled 11.31 describe samples stimulated with anti-CD3/CD28 in the presence of the corresponding mAb. Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (F and G) and MFI of cytokine expression (H and I) relative to anti-CD3/CD28–stimulated control, which was normalized to 1. (J) T cells in PBMCs were stimulated via anti-CD3/28 in the absence or presence of 100 μg/ml 11.31 or Bavi for 4 h and assayed for cytokine mRNA expression via FISH-Flow. Quantification of flow cytometric assays was represented in CD3+ cells producing cytokine mRNA transcripts (K) and MFI of cytokine expression (L) relative to anti-CD3/28–stimulated control normalized to 1. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005, ****p ≤ 0.00005.

Close modal

To analyze the effect of PS mAb treatments on activities downstream TCR stimulation, PBMCs were labeled with cell trace and stimulated with anti-CD3/28 in the presence or absence of PS mAbs 11.31 and Bavi. Consistent with cytokine production data, we found that 11.31 reduced the percentage of proliferated T cells in comparison with untreated or Bavi-treated PBMCs during a 72-h period (Supplemental Fig. 2A–C). These data indicate that 11.31-mediated inhibition of anti-CD3/28 stimulation of T cells reduces the amounts of protein and mRNA production of stimulatory cytokines as well as downstream activities such as proliferation.

Additionally, we explored whether the inhibitory effects of 11.31 involved crosstalk with other cell types, especially including monocytes. We assessed cytokine production from T cells after depleting monocytes from the PBMCs using forward scatter (FSC)Hi/side scatter (SSC)Hi and CD14+ FACS (Fig. 4A). Indeed, previous studies have shown that 11.31 can modulate T cell infection by HIV-1 virions in a monocyte-dependent fashion (37), possibly by altering paracrine factors that can influence T cells. As shown in (Fig. 4B–D, depletion of monocytes had no effect on 11.31-mediated T cell suppression. To further assess whether 11.31 inhibition was mediated through direct binding to activated T cells, CD3+ T cells were isolated from PBMCs using T cell purification negative selection beads per Miltenyi protocols (catalog no. 130-096-535) (Fig. 4E). Complementarily, 11.31-mediated suppression also was evident with positively enriched T cells (Fig. 4E–G), indicating that 11.31 binding directly to T cells abrogates T cell activation.

FIGURE 4.

Inhibitory effects of PS mAb 11.31 are T cell specific and monocyte independent. (A) Gating scheme of FACS-mediated monocyte depletion and confirmation of successful depletion. (B) Monocyte-containing and depleted PBMCs were stimulated with anti-CD3/CD28 in the presence or absence of 11.31, and the production of IFN-γ and TNF-α were measured in CD3+ T cells. (C and D) Quantification of flow cytometric assays is represented in CD3+ cells producing cytokines (C) and MFI of cytokine expression (D) relative to anti-CD3/28–stimulated control normalized to 1. (E) T cells were purified from PBMCs using negative selection and magnetic cell separation, and >95% purify was confirmed via flow cytometry per Miltenyi protocols (catalog no. 130-096-535). (F) Purified T cells were stimulated in the presence or absence of PS mAbs, and the production of IFN-γ and TNF-ɑ was measured in CD3+ T cells. (G and H) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (G) and MFI of cytokine expression (H) relative to anti-CD3/28–stimulated control normalized to 1. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005.

FIGURE 4.

Inhibitory effects of PS mAb 11.31 are T cell specific and monocyte independent. (A) Gating scheme of FACS-mediated monocyte depletion and confirmation of successful depletion. (B) Monocyte-containing and depleted PBMCs were stimulated with anti-CD3/CD28 in the presence or absence of 11.31, and the production of IFN-γ and TNF-α were measured in CD3+ T cells. (C and D) Quantification of flow cytometric assays is represented in CD3+ cells producing cytokines (C) and MFI of cytokine expression (D) relative to anti-CD3/28–stimulated control normalized to 1. (E) T cells were purified from PBMCs using negative selection and magnetic cell separation, and >95% purify was confirmed via flow cytometry per Miltenyi protocols (catalog no. 130-096-535). (F) Purified T cells were stimulated in the presence or absence of PS mAbs, and the production of IFN-γ and TNF-ɑ was measured in CD3+ T cells. (G and H) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (G) and MFI of cytokine expression (H) relative to anti-CD3/28–stimulated control normalized to 1. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005.

Close modal

To address mechanistically whether 11.31 acts directly on TCR signaling, or conversely downstream of receptor activation, we compared the effects of 11.31 on cellular responses following extracellular engagement with PHA, a lectin that binds and crosslinks glycans on the TCR (43), or with pharmacologic agents (PMA and ionomycin) that activate intracellular targets (protein kinase C and calmodulin, respectively). As shown in (Fig. 5A–C, when PBMCs were stimulated with PHA-P for 6 h, intracellular cytokine staining for IFN-γ and TNF-ɑ was significantly increased, and 11.31 treatment, like with anti-CD3/CD28, suppressed cytokine expression. In further support of the idea that 11.31 acts on a downstream target independent of immediate TCR activation, we examined tyrosine phosphorylation of Lck and ZAP-70 in oncogenic Jurkat cells, an immortalized line of human T cells that retain TCR activation capacity and externalize PS upon cellular stimulation (Supplemental Fig. 3A). As shown in Supplemental Fig. 3B and 3C, when Jurkat cells were stimulated with anti-CD3/CD28, immediate postreceptor-mediated tyrosine phosphorylation was not notably compromised. These data suggest that 11.31 acts downstream of PLC-γ and calcium mediated PS externalization. Indeed, when we performed activation assays using PMA-ionomycin, we observed that 11.31 treatment had minimal effects on the cytokine production in stimulated PBMCs (Fig. 5D–F). In these scenarios, effects of 11.31 on cytokine expression was unaltered after PMA-ionomycin treatment, suggesting that 11.31, and binding to externalized PS on Ag-independent, CD3-activated T cells, represents an inhibitory node that impairs subsequent cytokine production.

FIGURE 5.

Suppressive effects of 11.31 consistent in PHA-P and not PMA-ionomycin–stimulated cells, suggesting TCR interference. (A) T cells from PBMCs were stimulated with PHA-P in the presence or absence of PS mAbs. (B and C) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (B) and MFI of cytokine expression (C) relative to PHA-P–stimulated control normalized to 1. (D) T cells from PBMCs were stimulated with PMA-ionomycin in the presence or absence of PS mAbs. (E and F) Quantification of flow cytometric assays were represented in CD3+ cells producing cytokines (E) and MFI of cytokine expression (F) relative to PMA-ionomycin–stimulated control normalized to 1. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005, ****p ≤ 0.00005.

FIGURE 5.

Suppressive effects of 11.31 consistent in PHA-P and not PMA-ionomycin–stimulated cells, suggesting TCR interference. (A) T cells from PBMCs were stimulated with PHA-P in the presence or absence of PS mAbs. (B and C) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (B) and MFI of cytokine expression (C) relative to PHA-P–stimulated control normalized to 1. (D) T cells from PBMCs were stimulated with PMA-ionomycin in the presence or absence of PS mAbs. (E and F) Quantification of flow cytometric assays were represented in CD3+ cells producing cytokines (E) and MFI of cytokine expression (F) relative to PMA-ionomycin–stimulated control normalized to 1. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005, ****p ≤ 0.00005.

Close modal

Given that 11.31 and Bavi had distinct effects with respect to cytokine expression after CD3/28 stimulation, as well as distinct requirements for β2GPI (Fig. 6), we queried whether these PS-targeting mAbs may recognize different sources of externalized PS and/or at different subcellular and membrane locations. To address this experimentally, human PBMCs were activated by anti-CD3/CD28, stained with either 11.31 or Bavi (both conjugated to AF647), and assessed by AMNIS ImageStream (Fig. 6A). Under these conditions, the binding profiles of 11.31 and Bavi to activated T cells occurred in membrane puncta and were not uniformly distributed on the cell surface. Brightfield morphologies did not yield notable blebbing that would be indicative of any form of cell death, nor was the integrity of the membranes compromised, as was evident from the lack of DAPI staining.

FIGURE 6.

PS-targeting mAbs bind to externalized PS in punctate locales on the surface of activated T cells. (A) T cells in PBMCs were activated with anti-CD3/CD28 for 3 h and subjected to high-throughput visualization of 11.31 and Bavi binding to activated T cells using AMNIS ImageStream MKII. CD3-FITC and 11.31-AF647 or Bavi-AF647 was used for visualizing activated T cells bound by PS-targeting mAbs. (B) T cells in PBMCs were activated for 3 h in the presence of 11.31-AF647 and enriched for CD3+ cells using FACS. Representative confocal images of a resting T cell (left) and an activated T cell (right) that exhibit punctate staining of 11.31. (C) T cells, treated with camptothecin for 4 h, were stained for PS using 11.31-AF647 and Bavi-AF647 and analyzed via confocal microscopy. Scale bar, 5 µm. (D) Confocal images depicting 11.31 and Bavi costaining on activated T cells using Bavi-AF647 and 11.31-Biotin with a streptavidin-TRITC secondary stain. (E) Confocal images depicting subcellular localization of 11.31 using surface staining anti-CD3ε and wheat germ agglutinin (WGA). (F) 11.31 detection following permeabilization accomplished using anti-human IgG (H+L)-AF647 secondary Ab.

FIGURE 6.

PS-targeting mAbs bind to externalized PS in punctate locales on the surface of activated T cells. (A) T cells in PBMCs were activated with anti-CD3/CD28 for 3 h and subjected to high-throughput visualization of 11.31 and Bavi binding to activated T cells using AMNIS ImageStream MKII. CD3-FITC and 11.31-AF647 or Bavi-AF647 was used for visualizing activated T cells bound by PS-targeting mAbs. (B) T cells in PBMCs were activated for 3 h in the presence of 11.31-AF647 and enriched for CD3+ cells using FACS. Representative confocal images of a resting T cell (left) and an activated T cell (right) that exhibit punctate staining of 11.31. (C) T cells, treated with camptothecin for 4 h, were stained for PS using 11.31-AF647 and Bavi-AF647 and analyzed via confocal microscopy. Scale bar, 5 µm. (D) Confocal images depicting 11.31 and Bavi costaining on activated T cells using Bavi-AF647 and 11.31-Biotin with a streptavidin-TRITC secondary stain. (E) Confocal images depicting subcellular localization of 11.31 using surface staining anti-CD3ε and wheat germ agglutinin (WGA). (F) 11.31 detection following permeabilization accomplished using anti-human IgG (H+L)-AF647 secondary Ab.

Close modal

To visualize 11.31 versus Bavi binding localizations at higher resolution, T cells were stimulated with anti-CD3/CD28 and stained as above, followed by fixation with paraformaldehyde without detergent membrane solubilization to prevent PS internalization, and assessed by confocal microscopy (Fig. 6B–E). Notably, fixation was performed after staining to prevent any artifactual staining. Representative cell images revealed again that both 11.31 and Bavi preferentially bound to activated T cells in a punctate fashion (Fig. 6B, right), similar to results found in (Fig. 6A. In contrast, little, if any, PS-targeting mAb staining was observed on unstimulated cells, suggesting that assay conditions reflect TCR-mediated PS externalization and not binding to intracellular pools of PS (Fig. 6B, left). Interestingly, the patterns of PS staining (on the externalized surface) of activated cells were unique from the homogeneous and intense staining of apoptotic cells treated with camptothecin for 4 h (Fig. 6C). To visualize binding of 11.31 and Bavi on the same activated T cells, and query whether these mAbs bind distinct or nonoverlapping sites, PBMCs were stimulated and incubated simultaneously with AF647-conjugated Bavi or biotinylated 11.31 for 1 h. Cells were then fixed and stained with streptavidin–rhodamine B to visualize biotinylated 11.31. Although both 11.31 and Bavi bound to activated T cells in a nonuniform punctate fashion, they appeared as distinct depots (Fig. 6D). These microscopic analyses suggest that PS mAbs 11.31 and Bavi recognize nonoverlapping and distinct PS subsets on the surface of activated T cells, suggesting that there may exist specific membrane portals of PS exit. In recent years, it has been reported that anti-PE Abs recognize externalized PE on stressed endothelial cells and are ultimately endocytosed (44). To investigate whether 11.31 was internalized, PBMCs were stimulated with anti-CD3/CD28 and incubated with 11.31 for 3 h. Cells were stained with membrane staining wheat germ agglutinin, washed with PBS, and subsequently fixed/permeabilized (following a similar preparation for poly-l-lysine slide staining) to assess intracellular PS mAb staining. Anti-human IgG (H+L) secondary Abs were used to detect 11.31 bound to or internalized by T cells. We observed that 11.31 puncta were present and, in many cases, found in subcellular compartments (Fig. 6E, 6F), indicating that these mAbs are likely internalized by T cells over time; however, we cannot definitively conclude that this process is involved in 11.31 mechanism of inhibition.

To further explore the apparent distinction in PS bound by 11.31 and Bavi, we performed flow cytometry–based competition assays with Annexin V (which binds directly to PS in a calcium-dependent manner). Toward this goal, anti-CD3/CD28–stimulated T cells were incubated with fluorescently labeled 11.31 or Bavi in the presence of unconjugated Annexin V in molar excess (Fig. 7A). Additionally, these cells were incubated with fluorescently labeled Annexin V in the presence of either unconjugated 11.31 or Bavi in molar excess. Interestingly, whereas 11.31 was able to compete with Annexin V for binding to activated T cells (Fig. 7B), Bavi was ineffective in the presence of β2GPI (Fig. 7C).

FIGURE 7.

Epitope recognized by 11.31 and Annexin V on activated T cells distinct from that of Bavi. (A) Flow cytometric–based competition assays performed between 11.31 and Annexin V. T cells in PBMCs were stimulated with anti-CD3/CD28, and PS externalization was detected using fluorescently labeled 11.31 (top panel) or Annexin V (bottom panel). (B) Unconjugated competitors were incubated in molar excess, and percent competition was quantified and represented relative to binding to stimulated control. (C) Annexin V and Bavi competition assays were performed, and percent competition was quantified and represented relative to binding to stimulated control. (D) T cells in PBMCs were stimulated with anti-CD3/CD28 in the presence or absence of 11.31, Bavi, or Annexin V. (E and F) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (E) and MFI of cytokine expression (F) relative to anti-CD3/28–stimulated control normalized to 1. (G) SDS-PAGE followed by Coomassie blue staining displaying the m.w. differences between monomeric Annexin V and dimeric Annexin D19. (H) Sandwich ELISA on PS-coated plates analyzing difference in binding of monomeric Annexin V and Annexin D19. (I) T cells in PBMCs were stimulated with anti-CD3/CD28 in the presence or absence of Annexin D19. (J) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines relative to anti-CD3/28–stimulated control normalized to 1. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005, ****p ≤ 0.00005.

FIGURE 7.

Epitope recognized by 11.31 and Annexin V on activated T cells distinct from that of Bavi. (A) Flow cytometric–based competition assays performed between 11.31 and Annexin V. T cells in PBMCs were stimulated with anti-CD3/CD28, and PS externalization was detected using fluorescently labeled 11.31 (top panel) or Annexin V (bottom panel). (B) Unconjugated competitors were incubated in molar excess, and percent competition was quantified and represented relative to binding to stimulated control. (C) Annexin V and Bavi competition assays were performed, and percent competition was quantified and represented relative to binding to stimulated control. (D) T cells in PBMCs were stimulated with anti-CD3/CD28 in the presence or absence of 11.31, Bavi, or Annexin V. (E and F) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines (E) and MFI of cytokine expression (F) relative to anti-CD3/28–stimulated control normalized to 1. (G) SDS-PAGE followed by Coomassie blue staining displaying the m.w. differences between monomeric Annexin V and dimeric Annexin D19. (H) Sandwich ELISA on PS-coated plates analyzing difference in binding of monomeric Annexin V and Annexin D19. (I) T cells in PBMCs were stimulated with anti-CD3/CD28 in the presence or absence of Annexin D19. (J) Quantification of flow cytometric assays was represented in CD3+ cells producing cytokines relative to anti-CD3/28–stimulated control normalized to 1. Significance determined between treatment group and stimulation alone by Student t test. *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005, ****p ≤ 0.00005.

Close modal

Because Annexin V and 11.31 appear to share a functional PS subpopulation specificities, we also queried whether pretreatment of cells with Annexin V also blocked cytokine expression. We tested monomeric Annexin V and a dimeric form of Annexin V (D19; Advanced Proteomic Research) (Fig. 7F). As shown in (Fig. 7D, 7E, monomeric Annexin V did not inhibit anti-CD3/CD28–mediated T cell signaling. In contrast, D19, which recognizes PS with a higher affinity (Fig. 7G), exhibited a level of inhibition comparable to 11.31 (Fig. 7H, 7I). Collectively, our data suggest that 11.31 and Bavi recognize unique epitopes on activated T cells, accounting for the functional differences in modulating downstream activation, and that 11.31 and a dimeric Annexin protein divalently recognize externalized PS and alter T cell cytokine production following TCR stimulation.

The irreversible externalization of PS, from the inner to the outer surface of the plasma membrane, during caspase-mediated apoptosis is a well-characterized signal for efferocytosis. However, PS is also externalized reversibly on viable cells under a variety of physiological conditions that fail to induce efferocytosis, for example, during Ag- and/or ligand-mediated activation of T cells, B cells, and monocytes as well as on collagen-activated platelets, the latter serving as a platform for the recruitment and activation of plasma-borne coagulation factors to form the prothrombinase complex (45). On activated T cells, recent studies suggest that externalized PS is functionally significant (15). Studies by Fischer et al. (15) observed that Ag stimulation of CD4+ T cells led to the transient externalization of PS and that treatment with Annexin V diminished proinflammatory cytokine secretion. Moreover, Bollinger et al. (46) demonstrated that CD4+/Annexin V+ T cells (membrane-bound Annexin V) inhibited the proliferation of CD4+/Annexin V T cells and phosphorylation of mammalian target of rapamycin, a function that is partially dependent on cell–cell contact.

In this study, to investigate effects of therapeutic PS-targeting mAbs on T cell function, we investigated 11.31 (originally isolated from a phage display library from a healthy patient) and Bavi (human-engineered chimeric mAb from Peregrine Pharmaceuticals) previously developed for potential clinical use in cancer and infectious disease models. Similar to Ag-mediated T cell activation, Ag-independent anti-CD3/CD28 Ab-mediated activation of T cells results in rapid but reversible externalization of PS, presumably as a mechanism to control TCR overactivation and promote TCR resolution and downregulation (14, 15). We observed that direct PS mAb 11.31, but not Bavi (mAb requiring β2GPI), strongly suppressed anti-CD3/CD28–mediated and PHA-P–mediated induction of TNF-ɑ and IFN-γ cytokine production. These studies suggest that PS-targeting mAbs, including those produced naively or induced transiently during viral or bacterial infections or chronically present in various autoimmune diseases, may be contextual in activity and function, but there may exist natively a repertoire of PS-targeting mAbs (some activating, some inhibitory) that impinge on immune activation.

The results presented in this article also raise general queries concerning the physiological role of PS externalization in T cell biology. First, the fact that 11.31, a direct PS-binding mAb, and Annexin V dimers effectively blocked anti-CD3/28–mediated signaling implies that transmembrane phospholipid scrambling, and/or PS externalization a priori, may be important for functional TCR activation and signaling. For example, it is possible that 11.31 may trap externalized PS and prevent normal PS scrambling to prevent TCR signaling, thereby preventing critical signaling proteins from interacting dynamically with either intracellular or externalized PS. With respect to immediate TCR-mediated postreceptor signaling, it has been reported that during T cell activation, immunoisolated TCR active domains are enriched with phospholipids, such as plasmaenyl PS and PE (47). Additionally, several postreceptor kinases important for TCR signaling, including Lck kinase, are localized at the inner surface of the plasma membrane via N-terminal myristoylated and palmitoylated motifs that require the presence of anionic phospholipids for normal function (48). Additionally, it has been demonstrated that ITAM phosphorylation of Lck can be enhanced in membranes containing 10% of PS versus vesicles devoid of PS (49).

However, at the same time, the observations that TCR-mediated PS externalization temporally precedes IFN-γ and TNF-ɑ production, combined with the observation that 11.31 pretreatment did not inhibit immediate postreceptor tyrosine phosphorylation of ZAP-70 and Lck in Jurkat cells, suggest that 11.31 blocks cytosolic signal transduction between immediate receptor activation and changes in cytokine gene expression. Consistent with this idea, whereby 11.31 suppressed anti-CD3/CD28–inducible (activation of TCR) and PHA-inducible (crosslinks TCR by binding to glycosylated residues on cell surface proteins) cytokine expression, 11.31 did not block cytokine production following PMA/ionomycin treatment (downstream of PS scrambling), suggesting that 11.31 inhibitory signals are likely upstream to PLC-γ activation and the subsequent calcium-dependent PS scrambling and PS externalization. Interesting studies by Hu et al. (14) showed that TMEM16F was not only the dominant scramblase expressed in T cells and also that TMEM16F deficiency promoted T cell overactivation, further supporting the idea that PS externalization has a signaling function in the regulation T cell biology.

A second important observation made in this study identified that the externalization and localization of PS in T cells under conditions of cell stimulation (anti-CD3/CD28) and apoptosis (camptothecin) appear to be functionally and quantitatively distinct. Indeed, using immunofluorescence with fluorescently labeled 11.31 or Bavi, we observed externalized PS in punctate, focal topologies in TCR-stimulated cells, in contrast to what is observed in dying cells, whereby both the MFI of total PS is enhanced but the localization is more diffuse and uniform. These data are consistent with proposed distinct mechanisms of PS externalization via caspase-mediated apoptosis (namely, Xkr8, a caspase-activated scramblase and ATP11C, a caspase-inactivated flippase) and calcium-activated scramblases (such as the TMEM16 family) during TCR-mediated activation but also consistent with previous observations that not all PS externalizations are sufficient to induce efferocytosis. Interesting studies by Nagata and colleagues (50) have shown that viable cells that externalize PS by constitutively active TMEM16F are not recognized for phagocytic removal by macrophages. More-recent studies have implicated the chaperone CDC50A (flippase chaperone), whereby viable B cells that are CDC50A null are engulfed by phagocytes (51), indicating that flippase activity is important to prevent engulfment of viable cells. Such differences in PS arrangements on viable versus dying cells may suggest that there are distinct “ports of export” that influence interactions with neighboring cells and or modulate downstream intracellular signaling cascades.

Additionally, the implications that Bavi and 11.31 bind potentially distinct pools of PS on the surface of activated T cells, combined with their distinct effects on TCR signaling, might also imply there are specific local domains of PS that are critical for signaling and regulating cytokine expression. This is supported by several observations in the article that include 1) 11.31 and Bavi bind to a pool of nonoverlapping membrane depots when simultaneously incubated with activated T cells; 2) 11.31, but not Bavi, could be quantitatively competed with Annexin V for its ability to bind activated T cells; and 3) 11.31 binds directly to PS, whereas Bavi binds PS via β2GPI. Such direct binding of 11.31 to externalized PS may prevent physiological phospholipid scrambling as noted above or conversely may itself become internalized to the cytosol during reversible scrambling. For example, it has been reported that anti-PE Abs employed PE as a pathogenic target on endothelial cells (44) via Ab endocytosis and sequestration to Rab5 endosomes. Finally, 11.31 might also be expected to alter cytokine expression that indirectly impinges on TCR signaling. Previous studies by Moody et al. (37) showed that 11.31 modulated CCR5 trophic HIV infection of CD4+ T cells in a monocyte-dependent manner. Although the effects observed in the current study appeared to be directly attributed to T cells (i.e., depleting monocytes from PBMCs using CD14+FSCHI/SSCHI FACS and using negatively selected, purified T cells), we cannot discount the possibility that 11.31 alters autocrine cytokines that results in the suppression of IFN-γ and TNF-ɑ.

The fact that Bavi binds PS via β2GPI also has implications toward protein and lipid interactions. The mechanism of β2GPI mAb binding has been previously identified, in which the anionic phospholipids of PS are required to release serum β2GPI from a closed conformation into an open configuration that allows for more stable interactions with the lipid (52, 53). Once the β2GPI-PS-Ab complex is stably formed, it can remain bound to phospholipids or bind to other targets, such as TLR (5456), Annexin A2 (57), Ibɑ (58), and low-density lipoprotein receptor ApoER2 (59). These may explain in part the immune-activating capacity of Bavi and why Bavi can bind to M2 macrophages and myeloid-derived suppressor cells, stimulate a chemokine response, and differentiate cells toward an M1 macrophage phenotype (20).

In conclusion, we present evidence that PS mAbs can differentially recognize stimulated T cells and modulate TCR-mediated cytokine production. Through these studies, we show that direct binding mAb 11.31 can inhibit the production of IFN-γ and TNF-ɑ following TCR cross-linking on CD3+ T cells, whereas the β2GPI-mediated PS mAb does not. We also believe this study provides evidence to suggest that there is a diverse array of PS mAbs that exist clinically and can be used for different therapeutic purposes. For example, the sequence of 11.31 was derived from a phage library obtained from a healthy human donor (37), indicating that PS mAbs occur naturally and are produced in the absence of any apparent pathology. Much of the research in the field of antiphospholipid Ab production has been in the context of thrombotic disease such as antiphospholipid syndrome, and the immunomodulatory functions of these mAbs are modestly commented on. We anticipate that our results will highlight a novel role of antiphospholipid Abs, and future directions of this project will be to identify additional PS mAbs that exist clinically and investigate the potential immunomodulatory roles that they may possess.

We thank Shigekazu Nagata for helpful comments on the manuscript and for providing the CDC50 mutant cell lines. We also thank Bryan Ciccarelli and former Birge Lab members for helpful discussions in experimental analyses. We thank Allen Krantz for providing Annexin D19 (Advanced Proteomics Research).

This work was supported by The New Jersey Commission on Cancer Research (to D.C.).

The online version of this article contains supplemental material.

Abbreviations used in this article

Bavi

bavituximab

MFI

mean fluorescence intensity

PE

phosphatidylethanolamine

PS

phosphatidylserine

1.
Leventis
P. A.
,
S.
Grinstein
.
2010
.
The distribution and function of phosphatidylserine in cellular membranes.
Annu. Rev. Biophys.
39
:
407
427
.
2.
Balasubramanian
K.
,
A. J.
Schroit
.
2003
.
Aminophospholipid asymmetry: a matter of life and death.
Annu. Rev. Physiol.
65
:
701
734
.
3.
Suzuki
J.
,
D. P.
Denning
,
E.
Imanishi
,
H. R.
Horvitz
,
S.
Nagata
.
2013
.
Xk-related protein 8 and CED-8 promote phosphatidylserine exposure in apoptotic cells.
Science
341
:
403
406
.
4.
Segawa
K.
,
S.
Kurata
,
S.
Nagata
.
2016
.
Human type IV P-type ATPases that work as plasma membrane phospholipid flippases and their regulation by caspase and calcium.
J. Biol. Chem.
291
:
762
772
.
5.
Davra
V.
,
S. G.
Kimani
,
D.
Calianese
,
R. B.
Birge
.
2016
.
Ligand activation of TAM family receptors-implications for tumor biology and therapeutic response.
Cancers (Basel)
8
:
107
.
6.
Kumar
S.
,
R. B.
Birge
.
2016
.
Efferocytosis.
Curr. Biol.
26
:
R558
R559
.
7.
Kumar
S.
,
D.
Calianese
,
R. B.
Birge
.
2017
.
Efferocytosis of dying cells differentially modulate immunological outcomes in tumor microenvironment.
Immunol. Rev.
280
:
149
164
.
8.
Birge
R. B.
,
S.
Boeltz
,
S.
Kumar
,
J.
Carlson
,
J.
Wanderley
,
D.
Calianese
,
M.
Barcinski
,
R. A.
Brekken
,
X.
Huang
,
J. T.
Hutchins
, et al
2016
.
Phosphatidylserine is a global immunosuppressive signal in efferocytosis, infectious disease, and cancer.
Cell Death Differ.
23
:
962
978
.
9.
Suzuki
J.
,
M.
Umeda
,
P. J.
Sims
,
S.
Nagata
.
2010
.
Calcium-dependent phospholipid scrambling by TMEM16F.
Nature
468
:
834
838
.
10.
Falzone
M. E.
,
M.
Malvezzi
,
B. C.
Lee
,
A.
Accardi
.
2018
.
Known structures and unknown mechanisms of TMEM16 scramblases and channels.
J. Gen. Physiol.
150
:
933
947
.
11.
Pomorski
T.
,
A. K.
Menon
.
2006
.
Lipid flippases and their biological functions.
Cell. Mol. Life Sci.
63
:
2908
2921
.
12.
Bevers
E. M.
,
P. L.
Williamson
.
2011
.
Phospholipid scramblase: when phospholipid asymmetry goes away.
In
Transmembrane Dynamics of Lipids.
P. F.
Devaux
,
A.
Herrmann
.
John Wiley & Sons
,
Hoboken, NJ
, p.
119
146
.
13.
Ousingsawat
J.
,
P.
Wanitchakool
,
A.
Kmit
,
A. M.
Romao
,
W.
Jantarajit
,
R.
Schreiber
,
K.
Kunzelmann
.
2015
.
Anoctamin 6 mediates effects essential for innate immunity downstream of P2X7 receptors in macrophages.
Nat. Commun.
6
:
6245
.
14.
Hu
Y.
,
J. H.
Kim
,
K.
He
,
Q.
Wan
,
J.
Kim
,
M.
Flach
,
T.
Kirchhausen
,
A.
Vortkamp
,
F.
Winau
.
2016
.
Scramblase TMEM16F terminates T cell receptor signaling to restrict T cell exhaustion.
J. Exp. Med.
213
:
2759
2772
.
15.
Fischer
K.
,
S.
Voelkl
,
J.
Berger
,
R.
Andreesen
,
T.
Pomorski
,
A.
Mackensen
.
2006
.
Antigen recognition induces phosphatidylserine exposure on the cell surface of human CD8+ T cells.
Blood
108
:
4094
4101
.
16.
Chan
P. Y.
,
E. A.
Carrera Silva
,
D.
De Kouchkovsky
,
L. D.
Joannas
,
L.
Hao
,
D.
Hu
,
S.
Huntsman
,
C.
Eng
,
P.
Licona-Limón
,
J. S.
Weinstein
, et al
2016
.
The TAM family receptor tyrosine kinase TYRO3 is a negative regulator of type 2 immunity.
Science
352
:
99
103
.
17.
Stitt
T. N.
,
G.
Conn
,
M.
Gore
,
C.
Lai
,
J.
Bruno
,
C.
Radziejewski
,
K.
Mattsson
,
J.
Fisher
,
D. R.
Gies
,
P. F.
Jones
, et al
1995
.
The anticoagulation factor protein S and its relative, Gas6, are ligands for the Tyro 3/Axl family of receptor tyrosine kinases.
Cell
80
:
661
670
.
18.
Kim
R.
,
M.
Emi
,
K.
Tanabe
.
2005
.
Cancer cell immune escape and tumor progression by exploitation of anti-inflammatory and pro-inflammatory responses.
Cancer Biol. Ther.
4
:
924
933
.
19.
Amara
A.
,
J.
Mercer
.
2015
.
Viral apoptotic mimicry.
Nat. Rev. Microbiol.
13
:
461
469
.
20.
Belzile
O.
,
X.
Huang
,
J.
Gong
,
J.
Carlson
,
A. J.
Schroit
,
R. A.
Brekken
,
B. D.
Freimark
.
2018
.
Antibody targeting of phosphatidylserine for the detection and immunotherapy of cancer.
ImmunoTargets Ther.
7
:
1
14
.
21.
Ran
S.
,
A.
Downes
,
P. E.
Thorpe
.
2002
.
Increased exposure of anionic phospholipids on the surface of tumor blood vessels.
Cancer Res.
62
:
6132
6140
.
22.
Taylor
D. D.
,
C.
Gercel-Taylor
.
2011
.
Exosomes/microvesicles: mediators of cancer-associated immunosuppressive microenvironments.
Semin. Immunopathol.
33
:
441
454
.
23.
Utsugi
T.
,
A. J.
Schroit
,
J.
Connor
,
C. D.
Bucana
,
I. J.
Fidler
.
1991
.
Elevated expression of phosphatidylserine in the outer membrane leaflet of human tumor cells and recognition by activated human blood monocytes.
Cancer Res.
51
:
3062
3066
.
24.
Luster
T. A.
,
J.
He
,
X.
Huang
,
S. N.
Maiti
,
A. J.
Schroit
,
P. G.
de Groot
,
P. E.
Thorpe
.
2006
.
Plasma protein β-2-glycoprotein 1 mediates interaction between the anti-tumor monoclonal antibody 3G4 and anionic phospholipids on endothelial cells.
J. Biol. Chem.
281
:
29863
29871
.
25.
Thorpe
P. E.
2004
.
Vascular targeting agents as cancer therapeutics.
Clin. Cancer Res.
10
:
415
427
.
26.
Huang
X.
,
M.
Bennett
,
P. E.
Thorpe
.
2005
.
A monoclonal antibody that binds anionic phospholipids on tumor blood vessels enhances the antitumor effect of docetaxel on human breast tumors in mice.
Cancer Res.
65
:
4408
4416
.
27.
Soares
M. M.
,
S. W.
King
,
P. E.
Thorpe
.
2008
.
Targeting inside-out phosphatidylserine as a therapeutic strategy for viral diseases.
Nat. Med.
14
:
1357
1362
.
28.
Zhao
D.
,
J. H.
Stafford
,
H.
Zhou
,
P. E.
Thorpe
.
2011
.
Near-infrared optical imaging of exposed phosphatidylserine in a mouse glioma model.
Transl. Oncol.
4
:
355
364
.
29.
DeRose
P.
,
P. E.
Thorpe
,
D. E.
Gerber
.
2011
.
Development of bavituximab, a vascular targeting agent with immune-modulating properties, for lung cancer treatment.
Immunotherapy
3
:
933
944
.
30.
Chalasani
P.
,
M.
Marron
,
D.
Roe
,
K.
Clarke
,
M.
Iannone
,
R. B.
Livingston
,
J. S.
Shan
,
A. T.
Stopeck
.
2015
.
A phase I clinical trial of bavituximab and paclitaxel in patients with HER2 negative metastatic breast cancer.
Cancer Med.
4
:
1051
1059
.
31.
Digumarti
R.
,
P. P.
Bapsy
,
A. V.
Suresh
,
G. S.
Bhattacharyya
,
L.
Dasappa
,
J. S.
Shan
,
D. E.
Gerber
.
2014
.
Bavituximab plus paclitaxel and carboplatin for the treatment of advanced non-small-cell lung cancer.
Lung Cancer
86
:
231
236
.
32.
Gerber
D. E.
,
A. T.
Stopeck
,
L.
Wong
,
L. S.
Rosen
,
P. E.
Thorpe
,
J. S.
Shan
,
N. K.
Ibrahim
.
2011
.
Phase I safety and pharmacokinetic study of bavituximab, a chimeric phosphatidylserine-targeting monoclonal antibody, in patients with advanced solid tumors.
Clin. Cancer Res.
17
:
6888
6896
.
33.
Gerber
D. E.
,
L.
Horn
,
M.
Boyer
,
R.
Sanborn
,
R.
Natale
,
R.
Palmero
,
P.
Bidoli
,
I.
Bondarenko
,
P.
Germonpre
,
D.
Ghizdavescu
, et al
2018
.
Randomized phase III study of docetaxel plus bavituximab in previously treated advanced non-squamous non-small-cell lung cancer.
Ann. Oncol.
29
:
1548
1553
.
34.
Yin
Y.
,
X.
Huang
,
K. D.
Lynn
,
P. E.
Thorpe
.
2013
.
Phosphatidylserine-targeting antibody induces M1 macrophage polarization and promotes myeloid-derived suppressor cell differentiation.
Cancer Immunol. Res.
1
:
256
268
.
35.
Callahan
M. K.
,
P. M.
Popernack
,
S.
Tsutsui
,
L.
Truong
,
R. A.
Schlegel
,
A. J.
Henderson
.
2003
.
Phosphatidylserine on HIV envelope is a cofactor for infection of monocytic cells.
J. Immunol.
170
:
4840
4845
.
36.
Dowall
S. D.
,
V. A.
Graham
,
K.
Corbin-Lickfett
,
C.
Empig
,
K.
Schlunegger
,
C. B.
Bruce
,
L.
Easterbrook
,
R.
Hewson
.
2015
.
Effective binding of a phosphatidylserine-targeting antibody to Ebola virus infected cells and purified virions.
J. Immunol. Res.
2015
:
347903
.
37.
Moody
M. A.
,
H.-X.
Liao
,
S. M.
Alam
,
R. M.
Scearce
,
M. K.
Plonk
,
D. M.
Kozink
,
M. S.
Drinker
,
R.
Zhang
,
S.-M.
Xia
,
L. L.
Sutherland
, et al
2010
.
Anti-phospholipid human monoclonal antibodies inhibit CCR5-tropic HIV-1 and induce β-chemokines.
J. Exp. Med.
207
:
763
776
.
38.
Vir
P.
,
R.
Arrigucci
,
K.
Lakehal
,
A. L.
Davidow
,
R.
Pine
,
S.
Tyagi
,
Y.
Bushkin
,
A.
Lardizabal
,
M. L.
Gennaro
.
2015
.
Single-cell cytokine gene expression in peripheral blood cells correlates with latent tuberculosis status.
PLoS One
10
:
e0144904
.
39.
Bushkin
Y.
,
F.
Radford
,
R.
Pine
,
A.
Lardizabal
,
B. T.
Mangura
,
M. L.
Gennaro
,
S.
Tyagi
.
2015
.
Profiling T cell activation using single-molecule fluorescence in situ hybridization and flow cytometry.
J. Immunol.
194
:
836
841
.
40.
Seidel
J. A.
,
A.
Otsuka
,
K.
Kabashima
.
2018
.
Anti-PD-1 and anti-CTLA-4 therapies in cancer: mechanisms of action, efficacy, and limitations.
Front. Oncol.
8
:
86
.
41.
Sun
Y.
,
C.
Vandenbriele
,
A.
Kauskot
,
P.
Verhamme
,
M. F.
Hoylaerts
,
G. J.
Wright
.
2015
.
A human platelet receptor protein microarray identifies the high affinity immunoglobulin E receptor subunit α (FcεR1α) as an activating platelet endothelium aggregation receptor 1 (PEAR1) ligand.
Mol. Cell. Proteomics
14
:
1265
1274
.
42.
Segawa
K.
,
S.
Kurata
,
S.
Nagata
.
2018
.
The CDC50A extracellular domain is required for forming a functional complex with and chaperoning phospholipid flippases to the plasma membrane.
J. Biol. Chem.
293
:
2172
2182
.
43.
Kinoshita
S.
,
L.
Su
,
M.
Amano
,
L. A.
Timmerman
,
H.
Kaneshima
,
G. P.
Nolan
.
1997
.
The T cell activation factor NF-ATc positively regulates HIV-1 replication and gene expression in T cells.
Immunity
6
:
235
244
.
44.
Hou
S.
,
H.
Fölsch
,
K.
Ke.
,
J.
Cook Mills
,
R.
Ramsey-Goldman
,
M.
Zhao
.
2017
.
Early endosome as a pathogenic target for antiphosphatidylethanolamine antibodies.
Proceedings of the National Academy of Sciences
114
:
13798
13803
.
45.
Lentz
B. R.
2003
.
Exposure of platelet membrane phosphatidylserine regulates blood coagulation.
Prog. Lipid Res.
42
:
423
438
.
46.
Bollinger
A. L.
,
T.
Bollinger
,
J.
Rupp
,
K.
Shima
,
N.
Gross
,
L.
Padayachy
,
R.
Chicheportiche
,
G. L.
Puga Yung
,
J. D.
Seebach
.
2020
.
Annexin V expression on CD4+ T cells with regulatory function.
Immunology
159
:
205
220
.
47.
Zech
T.
,
C. S.
Ejsing
,
K.
Gaus
,
B.
de Wet
,
A.
Shevchenko
,
K.
Simons
,
T.
Harder
.
2009
.
Accumulation of raft lipids in T-cell plasma membrane domains engaged in TCR signalling.
EMBO J.
28
:
466
476
.
48.
Kim
P. W.
,
Z. Y.
Sun
,
S. C.
Blacklow
,
G.
Wagner
,
M. J.
Eck
.
2003
.
A zinc clasp structure tethers Lck to T cell coreceptors CD4 and CD8.
Science
301
:
1725
1728
.
49.
Hui
E.
,
R. D.
Vale
.
2014
.
In vitro membrane reconstitution of the T-cell receptor proximal signaling network.
Nat. Struct. Mol. Biol.
21
:
133
142
.
50.
Segawa
K.
,
J.
Suzuki
,
S.
Nagata
.
2011
.
Constitutive exposure of phosphatidylserine on viable cells. [Published erratum appears in 2012 Proc. Natl. Acad. Sci. USA. 109: 995.]
Proc. Natl. Acad. Sci. USA
108
:
19246
19251
.
51.
Segawa
K.
,
Y.
Yanagihashi
,
K.
Yamada
,
C.
Suzuki
,
Y.
Uchiyama
,
S.
Nagata
.
2018
.
Phospholipid flippases enable precursor B cells to flee engulfment by macrophages.
Proc. Natl. Acad. Sci. USA
115
:
12212
12217
.
52.
de Groot
P. G.
2011
.
Mechanisms of anti-phospholipid antibody formation and action.
Thromb. Res.
127
(
Suppl 3
):
S40
S42
.
53.
Blank
M.
,
I.
Krause
,
M.
Fridkin
,
N.
Keller
,
J.
Kopolovic
,
I.
Goldberg
,
A.
Tobar
,
Y.
Shoenfeld
.
2002
.
Bacterial induction of autoantibodies to β2-glycoprotein-I accounts for the infectious etiology of antiphospholipid syndrome.
J. Clin. Invest.
109
:
797
804
.
54.
Satta
N.
,
S.
Dunoyer-Geindre
,
G.
Reber
,
R. J.
Fish
,
F.
Boehlen
,
E. K. O.
Kruithof
,
P.
de Moerloose
.
2007
.
The role of TLR2 in the inflammatory activation of mouse fibroblasts by human antiphospholipid antibodies.
Blood
109
:
1507
1514
.
55.
Allen
K. L.
,
F. V.
Fonseca
,
V.
Betapudi
,
B.
Willard
,
J.
Zhang
,
K. R.
McCrae
.
2012
.
A novel pathway for human endothelial cell activation by antiphospholipid/anti-β2 glycoprotein I antibodies.
Blood
119
:
884
893
.
56.
Raschi
E.
,
C.
Testoni
,
D.
Bosisio
,
M. O.
Borghi
,
T.
Koike
,
A.
Mantovani
,
P. L.
Meroni
.
2003
.
Role of the MyD88 transduction signaling pathway in endothelial activation by antiphospholipid antibodies.
Blood
101
:
3495
3500
.
57.
Ma
K.
,
R.
Simantov
,
J.-C.
Zhang
,
R.
Silverstein
,
K. A.
Hajjar
,
K. R.
McCrae
.
2000
.
High affinity binding of β 2-glycoprotein I to human endothelial cells is mediated by annexin II.
J. Biol. Chem.
275
:
15541
15548
.
58.
Shi
T.
,
B.
Giannakopoulos
,
X.
Yan
,
P.
Yu
,
M. C.
Berndt
,
R. K.
Andrews
,
J.
Rivera
,
G. M.
Iverson
,
K. A.
Cockerill
,
M. D.
Linnik
,
S. A.
Krilis
.
2006
.
Anti-β2-glycoprotein I antibodies in complex with β2-glycoprotein I can activate platelets in a dysregulated manner via glycoprotein Ib-IX-V.
Arthritis Rheum.
54
:
2558
2567
.
59.
Lutters
B. C. H.
,
R. H. W. M.
Derksen
,
W. L.
Tekelenburg
,
P. J.
Lenting
,
J.
Arnout
,
P. G.
de Groot
.
2003
.
Dimers of β 2-glycoprotein I increase platelet deposition to collagen via interaction with phospholipids and the apolipoprotein E receptor 2'.
J. Biol. Chem.
278
:
33831
33838
.

R.A.B. is supported for work distinct from the current study by a sponsored research agreement from Oncologie Inc., a biopharmaceutical company developing Bavi. The other authors have no financial conflicts of interest.

Supplementary data