Anti-HPA-1a standard sera (03/152) were obtained from the National Institute for Biological Standards and Control (Potters Bar, Hertfordshire, UK). Depending on the stock solution, final dilutions of 1:16 (lot #324735, #331314) or 1:4 (lot #332285) were used. High-titer anti-HLA class I sera were collected from five female donors and characterized in our laboratory. A pool of AB sera was also collected from six healthy male blood donors. This study was conducted in accordance with the declaration of Helsinki and was approved by the Ethics Committee of the Medical Faculty, Justus Liebig University, Giessen, Germany (file no. 82/09 and file no. 05/00).

Mouse monoclonal antibodies (mAbs) against HLA-ABC (clone w6/32; IgG2a), HLA-DR (clone L243; IgG2a), and HLA-DQ (clone Tü169; IgG2a), along with an isotype control (IgG2a), were purchased from Biolegend (Hamburg, Germany). The F(ab′)₂ IgG fragment was generated by pepsin digestion using the FragIT Z Micro Preparation Kit and purified as recommended by the manufacturer (Genovis, Lund, Sweden). The purity of the F(ab′)₂ fragment was verified by JESS protein capillary electrophoresis (Bio-Techne, Minneapolis, MN, USA) ( Supplementary Figure 1 ). The fluorescently labeled mAbs used for flow cytometry analysis are listed in Supplementary Table 1 . A mAb against MHC-I (34-1-2S; IgG2a) was purchased from ATCC (VA, USA) and a mAb targeting αIIbβ3 integrin (Leo.F2; IgG2a) was obtained from Emfret Analytics (Shanghai, China).

In this study, platelet phagocytosis by monocytes was performed in the following three different experimental settings: a) untreated platelet-free whole blood, b) antibody-pretreated platelet-free whole blood, and c) isolated monocytes.

a) Phagocytosis of anti-HPA-1a antibody-opsonized platelets by monocytes in untreated platelet-free whole blood was performed using the WHOPPA, as previously described (14). Briefly, 200 µL of pHrodo-labeled platelets (1.5 × 10⁷ platelets) was opsonized with diluted anti-HPA-1a standard sera for 30 min in the dark. After two washes with Dulbecco’s Phosphate Buffered Saline (D-PBS; Anprotec, Bruckberg, Germany) containing 10 mM EDTA and 0.3 mM prostaglandin (Sigma-Aldrich, St. Louis, MI, USA) (PEP, pH 7.4) (10 min, 800 × g), platelets were resuspended in 200 µL of Hank’s Balanced Salt Solution (HBSS) containing calcium and magnesium (Anprotec) and 10 mM HEPES (Gibco, Life Technologies, Darmstadt, Germany). Then, 200 µL of pHrodo-labeled, sensitized platelets was mixed with 200 µL of platelet-free whole blood containing 5 × 10⁴ monocytes and incubated for 60 min at 37°C. After erythrocyte lysis, viable monocytes were double-labeled with allophycocyanin (APC)-conjugated anti-CD14 and fluorescein isothiocyanate (FITC)-conjugated anti-CD45 mAbs. These were then counterstained with 1 mM SYTOX Blue (Invitrogen, Paisley, UK) for 30 min in the dark and measured by flow cytometry (FACSCanto II, Becton Dickinson, Franklin Lakes, NJ, USA). The percentage of pHrodo-positive CD45⁺CD14⁺ cells (monocytes that phagocytose opsonized platelets) was analyzed using FlowJo v.10.6.1_CL software (Becton and Dickinson). A minimum of 5,000 CD45⁺CD14⁺ cells were examined. b) Platelet-free whole blood (see above; 200 µL) was treated with mAbs against HLA-ABC, HLA-DR, or HLA-DQ IgG or F(ab′)₂ fragments (7.5 or 5.0 mg/mL, respectively), or with an anti-HLA serum sample (dilution 1:4 or 1:16) for 30 min. The corresponding isotypes and diluted AB sera served as negative controls. After washing, anti-HPA-1a antibody-opsonized platelets were added and the rate of platelet phagocytosis by monocytes was measured by flow cytometry as described above. c) Monocytes were isolated from whole blood using the MojoSort™ Human Pan Monocyte Isolation Kit (Biolegend, San Diego, CA, USA). The monocyte suspension was adjusted to a concentration of 0.25 × 10⁶/mL in HBSS buffer containing 1 M HEPES. A 200 µL volume of the monocyte suspension (5 × 10⁴ monocytes) was incubated with mAbs against HLA-ABC and HLA-DR. After washing, anti-HPA-1a antibody-opsonized platelets were added, and the rate of platelet phagocytosis by monocytes was measured by flow cytometry as described above.

EDTA-anticoagulated blood (100 µL) was incubated with FITC-labeled anti-HLA-ABC antibody (W6/32; 10 µg/mL) together with BV510-labeled anti-CD45 (HI30; 5 µg/mL), phycoerythrin (PE)-conjugated anti-CD61 (VI-PL2; 3.0 mg/mL), or PB-labeled anti-CD41 (HIPS; 10.0 µg/mL) antibodies for 20 min in the dark following the manufacturer’s protocol. The corresponding fluorescently labeled isotypes served as controls ( Supplementary Table 1 ). After spontaneous sedimentation (approximately 20 min), platelet-rich plasma (PRP) was collected and washed with D-PBS (Anprotec) containing 10 mM EDTA and 0.3 µm PGE (Sigma-Aldrich). The remaining platelet-poor blood was subsequently treated with erythrocyte lysis buffer for 5 min and then washed with D-PBS. Fluorescently labeled cells were measured by flow cytometry as described above. Leukocytes were gated via the CD45 marker, following which lymphocytes, monocytes, and granulocytes were identified based on forward scatter (FSC) and sideward scatter (SSC) properties. Platelets were identified by the presence of CD41 and CD61 markers. The normalized mean fluorescence intensity (MFI) was calculated as the ratio of the antibody signal to that of the corresponding isotype control.

MAb 34-1-2S was isolated from Protein-Free Hybridoma Medium (PFHM-II, Life Technologies Corporation, NY, USA) and purified by affinity chromatography as previously described (15). Purified mAb 34-1-2S and purified mAb Leo. F2 (0.25, 1 and 2 mg/kg) were administered to Balb/c (H-2K^d haplotype) mice via the tail vein. Mouse IgG2a (eBM2a, eBioscience, San Diego, CA, USA) was used as negative control. A small amount of blood (10 µL) was obtained from the tail vein and diluted with 240 µL of buffer for blood cell analysis. Platelet counts were determined using the Veterinary Auto Hematology Analyzer (V-52 series, Mindray, Shenzhen, China) before and at different time points after antibody administration. Furthermore, Balb/c female mice were sequentially injected with mAb 34-1-2S (2 and 0.25 mg/kg), followed 30 min later by Leo.F2 (1 and 0.25 mg/kg). Platelet counts were subsequently measured as described above.

EDTA-anticoagulated blood (50 µL) was absorbed from mice before and after antibody administration (see above). Antibody bound to platelets and monocytes was then quantified by flow cytometry. In brief, PRP was obtained by centrifugation at 100 × g for 15 min. Peripheral blood mononuclear cells (PBMCs) were isolated from the remaining whole blood using density gradient centrifugation (1.084; Percoll; GE Healthcare, Uppsala, Sweden). PRP (50 µL; 1 × 10⁵ platelets/µL) in D-PBS (Gibco, Suzhou, China) was stained with FITC-conjugated goat anti-mouse IgG (JacksonImmunoResearch, West Grove, PA, USA) and PE-labeled anti-mouse CD41a antibody (MWReg30, Life Technologies Corp, Carlsbad, CA, USA) at room temperature in the dark. Monocytes were labeled with FITC-conjugated goat anti-mouse IgG (JacksonImmunoResearch) and PE-labeled anti-mouse CD11b antibody (M1/70, Life Technologies Corp). After two D-PBS washes, the cells were resuspended and quantified by flow cytometry (FACS Canto II; BD Biosciences, San Jose, CA, USA).

Data were analyzed using GraphPad Prism Software (Version 10.0.2, GraphPad Software, Inc., La Jolla, CA, USA) and SAS OnDemand for Academics (SAS Institute Inc., Cary, NC, USA). Exploratory analysis was based on descriptive statistics. As the data did not meet the assumptions of parametric tests, non-parametric methods were used. For non-normally distributed percentage data, the Friedman test was applied for group sizes greater than two, while the Wilcoxon test with pairwise comparisons was used for group sizes of two. Power analysis was also performed. To compare data across different time points relative to 0 h, one-way ANOVA was used. For repeated measures involving multiple comparisons, two-way ANOVA was applied.

It remains unclear which cells in whole blood are preferentially targeted by anti-HLA-ABC antibodies. To address this, we first used double staining approaches with flow cytometry to measure the binding of anti-HLA-ABC antibodies on different cells in whole blood. As shown in Figure 1 , HLA-ABC expression on the platelet surface was low compared to that in other blood cells. Notably, monocytes showed the highest HLA-ABC labeling, indicating that anti-HLA-ABC antibodies may preferentially bind to them in whole blood. This result aligns with previous proteasome expression data (13). It is therefore conceivable that anti-HLA-ABC antibodies bound to monocytes can interfere with platelet phagocytosis induced by other platelet-reactive antibodies, such as anti-HPA-1a antibodies.

To test this hypothesis, we performed a WHOPPA using whole blood pretreated with anti-HLA-ABC antibody. As shown in Figure 2 , the pre-incubation of whole blood with mAb W6/32, which targets HLA-ABC, or pooled human sera containing anti-HLA-ABC antibodies, inhibited anti-HPA-1a antibody-opsonized platelet phagocytosis in a concentration-dependent manner. Furthermore, the phagocytosis of anti-HPA-1a antibody-opsonized platelets was inhibited not only by anti-HLA-ABC IgG (mAb W6/32) but also by anti-HLA-DR IgG (mAb L243). However, like the isotype control, the anti-HLA-DQ antibody, mAb Tü169, did not affect platelet phagocytosis ( Figures 3A ). Conversely, pretreatment of whole blood with the F(ab′)₂ fragment of the anti-HLA-ABC and anti-HLA-DR antibodies did not inhibit the engulfment of anti-HPA-1a antibody-opsonized platelets ( Figure 3B ). To verify that this inhibition was mediated through antibody binding to monocytes, these cells were isolated and subsequently treated with anti-HLA-ABC antibody or anti-HLA-DR IgG as described above ( Figure 3C ). Similar results were obtained, namely, the phagocytosis of anti-HPA-1a antibody-opsonized platelets was inhibited when monocytes were pretreated with anti-HLA-ABC or HLA-DR antibodies. In the control experiment, isotype-treated monocytes could still engulf anti-HPA-1a antibody-opsonized platelets. Although monocytes also express HLA-DQ, the binding of anti-HLA-DQ antibodies did not prevent platelet phagocytosis, unlike that observed with the anti-HLA-DR antibody ( Figure 3A ).

Our results demonstrated that anti-HLA-ABC antibodies not only bound to platelets, triggering their phagocytosis but also strongly interacted with monocytes, consequently inhibiting the phagocytosis of anti-HPA-1a antibody-opsonized platelets. It remains unclear whether a similar mechanism operates in vivo.

To test our hypothesis, we investigated in vivo platelet clearance in female Balb/c mice following administration of mAbs against MHC-I (clone 34-1-2S) and αIIbβ3 (clone Leo.F2) ( Figure 4 ). The injection of anti-MHC-I antibodies (1 mg/kg) resulted in a significant decrease in the platelet count within 24 h (0 h: 1261.8 ± 59.85 × 10⁹/L vs. 755.6 ± 107.93 × 10⁹/L at 0.5 h; 741.0 ± 89.52 × 10⁹/L at 1 h; 815.4 ± 63.28 × 10⁹/L at 2 h; 793.2 ± 168.27 × 10⁹/L at 4 h; 850.6 ± 133.15 × 10⁹/L at 6 h; and 969.8 ± 77.43 × 10⁹/L at 24 h; p < 0.05). Platelet counts gradually returned to baseline levels by 72 h post-injection. A similar clearance profile was observed at a higher antibody dose (2 mg/kg). However, the administration of anti-αIIbβ3 antibodies (1 mg/kg) led to significant higher platelet clearance over the different time points (0 h: 1299.8 ± 208.38 × 10⁹/L vs. 408.6 ± 137.53 × 10⁹/L at 0.5 h; 319.2 ± 119.30 × 10⁹/L at 1 h; 305.8 ± 100.31 at 2 h; 211.6 ± 70.97 at 4 h; 183.8 ± 69.84 at 6 h; and 280.6 ± 70.77 at 24 h; p < 0.01). After 72 h, the platelet counts also returned to a normal value. A comparable result was achieved with a higher antibody dose (2 mg/kg). In the control experiment, the administration of the isotype control did not influence the platelet count over time (p > 0.05). When compared with mouse IgG, the platelet counts of Balb/c mice treated with 34-1-2S (1 mg/kg) were significantly decreased over time. However, the platelet counts of Leo.F2-treated mice were continuously very low for 24 h ( Supplementary Table 2 ). Next, we sought to define the minimal dose at which anti-MHC-I antibodies could still induce platelet clearance. As shown in Figure 4 , treatment with 0.25 mg/kg anti-MHC-I antibodies did not induce thrombocytopenia in our mice, as also observed with normal mouse IgG ( Supplementary Table 2 ). In contrast, anti-αIIbβ3 antibodies still induced thrombocytopenia at the same dose.

These results indicated that anti-MHC-I and anti-αIIbβ3 antibodies mediate different kinetics of platelet clearance in mice, with anti-MHC-I antibodies inducing mild platelet clearance and anti-αIIbβ3 antibodies promoting strong platelet clearance.

Our in vitro results indicated that anti-MHC-I antibodies can bind to monocytes and inhibit the phagocytosis of anti-HPA-1a antibody-opsonized platelets. This phenomenon, however, might not apply under in vivo conditions. To examine antibody distribution in whole blood, we isolated platelets and monocytes from Balb/c female mice before and after the administration of anti-MHC-I or anti-αIIbβ3 antibodies (dose 1mg/kg). Bound antibodies were analyzed by flow cytometry using FITC-labeled secondary antibodies. Platelets were identified with PE-labeled anti-CD41a antibodies while monocytes were differentiated with PE-labeled anti-CD11b antibodies ( Supplementary Figure 2 ). As shown in Figure 5A , strong platelet binding by the anti-αIIbβ3 antibody was found at the 0.5 and 2 h time points. As expected, the binding of platelet-specific anti-αIIbβ3 antibodies on monocytes was not detected ( Figure 5A ). In contrast, strong anti-MHC-I antibody binding to monocytes was evident after 0.5 h, whereas its binding to platelets was significantly weaker ( Figure 5B ). After 2 h, however, monocytes showed heterogeneous staining and some monocytes were negative for MHC-I, indicating antigen membrane mobility triggered by patching, capping, and endocytosis (16). However, at the low dose (0.25 mg/kg), anti-MHC-I antibodies were only detected on monocytes and were not detectable on platelets ( Figure 5C ). In accordance with our earlier observations ( Figure 4 ), platelet clearance was not detected at this antibody dose.

These findings suggested that anti-MHC-I antibodies preferentially interact with monocytes rather than platelets in vivo, resulting in significantly weaker reactivity compared to the platelet-specific anti-αIIbβ3 antibodies. This may explain the observed lower clearance of anti-MHC-I-opsonized platelets compared to platelet clearance triggered by anti-αIIbβ3 antibodies.

Next, we examined whether anti-MHC-I antibodies bound to monocytes could modify anti-αIIbβ3 antibody-induced platelet clearance in vivo. At a low anti-αIIbβ3 antibody dose (0.25 mg/kg), co-administration of a low dose of anti-MHC-I antibodies (0.25 mg/kg) did not significantly alter platelet clearance (p > 0.05; Figure 6A ; Supplementary Table 3 ). A similar result was observed when a high dose of anti-MHC-I antibodies (2 mg/kg) was administered 30 min before either the low-dose (0.25 mg/kg) or high-dose (1 mg/kg) anti-αIIbβ3 antibodies ( Figures 6B, C ; Supplementary Table 3 ).

These results suggested that the presence of anti-MHC-I antibodies does not significantly affect the platelet clearance induced by anti-αIIbβ3 antibodies in this murine model.