In vitro analysis of anti-HPA-1a dependent platelet phagocytosis and its inhibition using a new whole blood phagocytosis assay (WHOPPA)

Fetal and neonatal alloimmune thrombocytopenia (FNAIT) is a serious bleeding condition mostly caused by the reaction between maternal anti-HPA-1a antibodies and fetal platelets. This reaction leads to Fc-dependent platelet phagocytosis. Although several serological methods have been developed to identify maternal antibodies, a reliable laboratory parameter as a prognostic tool for FNAIT severity is still lacking. In this study, we developed whole blood platelet phagocytosis assay (WHOPPA), a flow cytometry-based phagocytosis assay that uses a pH-sensitive fluorescent dye (pHrodo-SE) to analyze anti-HPA-1a-dependent platelet phagocytosis in whole blood. WHOPPA revealed a high phagocytosis rate for the anti-HPA-1a opsonized platelets by monocytes but not by neutrophils. Analysis of different monocyte populations showed that all monocyte subsets, including classical (CD14++CD16−), intermediate (CD14++CD16+), and nonclassical (CD14+CD16++) monocytes, were able to engulf opsonized platelets. A unique monocyte subset, termed shifted monocytes (CD14+CD16−), showed the highest phagocytosis rate and was detected after platelet engulfment. FcγR inhibition tests revealed that except for FcγRIIa, FcγRI and FcγRIII on monocytes were responsible for the phagocytosis of anti-HPA-1a opsonized platelets. Analysis of anti-HPA-1a antibodies from FNAIT cases (n = 7) showed the phagocytosis of HPA-1aa but not of HPA-1bb platelets by monocytes. The phagocytosis rate was highly correlated with bound antibodies measured by flow cytometry (p < 0001; r = 0.9214) and MAIPA assay (p < 0.001; r = 0.7692). The phagocytosis rates were equal for type I and II anti-HPA-1a antibodies recognizing the plexin–semaphoring–integrin (PSI) domain and PSI/epidermal growth factor 1 domain of β3 integrin, respectively. By contrast, type III anti-HPA-1a antibodies reacting with αvβ3 integrin did not induce platelet phagocytosis. Furthermore, effector-silenced mAbs against HPA-1a inhibited the phagocytosis of anti-HPA-1a opsonized platelets. In conclusion, WHOPPA is a reliable in vitro platelet phagocytosis assay that mimics the phagocytosis of anti-HPA-1a opsonized platelets in whole blood. This assay allows to prove platelet phagocytosis ex vivo and evaluate the inhibitory capacity of different inhibitors as therapeutically strategies for the prevention of fetal thrombocytopenia in FNAIT in the future.


Introduction
Fetal and neonatal alloimmune thrombocytopenia (FNAIT) is a serious bleeding condition caused by the transplacental transfer of maternal IgG alloantibodies reacting with platelet antigens expressed on fetal platelets.This antigen-antibody reaction leads to platelet phagocytosis, which is predominantly mediated by the interaction between the Fc part of bound antibodies and the Fcg receptors (FcgRs) expressed on macrophages (1).In Caucasian population, anti-HPA-1a alloantibodies are responsible for about 85% of all FNAIT cases, with an incidence of approximately 1/1800 pregnancies (2).The clinical presentation of FNAIT varies from asymptomatic thrombocytopenia to severe clinical complications, including intrauterine growth restriction, intrauterine fetal demise, and intracranial hemorrhage (ICH).Among which, ICH is the most serious complication that occurs in around 10%-20% of FNAIT cases (3), leading to fetal death or persistent neurological sequelae in neonates (4).However, reliable laboratory parameters as a prognostic tool to identify fetuses/newborns at risk of severe FNAIT are still lacking (5).
HPA-1a is formed by a single amino acid substitution (Leu33Pro) located on the flexible plexin-semaphorin-integrin (PSI) domain of integrin b3 chain (6).On the platelet surface, the b3 chain can form heterodimers with either aIIb or av subunit, functioning as the fibrinogen or vitronectin receptor, respectively (7).Recent studies demonstrated that immunized mothers could develop three different subtypes of anti-HPA-1a antibodies.The anti-HPA-1a subtypes I and II react with epitopes formed by the polymorphic PSI domain of the b3 chain alone and the epidermal growth factor 1 (EGF1) domain, respectively (8).Subtype III recognizes the antigenic determinants formed by avb3 heterodimer and could impair endothelial function associated with ICH in severe FNAIT cases (9).Whether subtype III antibody could also induce platelet phagocytosis is unclear.
From the current general view, maternal anti-HPA-1a antibodies bound to fetal platelets lead to fetal platelet clearance by monocytes/macrophages in the spleen (1).Recent evidence indicated that platelet antibodies could also cause platelet clearance via an Fc-independent mechanism, leading to platelet phagocytosis via the Ashwell-Morell receptor in the liver (10).In humans, three FcgRs, namely, FcgRI, FcgRII, and FcgRIII, are present during fetal development (11).The most abundant blood monocytes (classical subset) mainly express FcgRI and FcgRIIa, and a small population of monocyte subsets (nonclassical and intermediate subsets) express FcgRIIIa.Meanwhile, neutrophils express abundant FcgRIIIb (12).Determining which blood phagocyte subsets and which FcgRs are responsible for anti-HPA-1a-mediated platelet phagocytosis is necessary.
Our previous studies demonstrated that deglycosylated monoclonal antibody (mAb) against HPA-1a could abrogate platelet phagocytosis without interfering with FcRn-mediated placental transport (13).Whether such effector-silenced mAb could represent a novel inhibitor to prevent fetal platelet destruction by maternal anti-HPA-1a antibodies in FNAIT is worthy of exploration.Xu and colleagues (14) demonstrated that deglycosylated mAb GZ1 against CD36 could effectively prevent fetal death in an anti-CD36-mediated FNAIT mouse model.However, the efficacy of such approach in preventing anti-HPA-1a platelet clearance remains unclear.
In the last decades, a number of flow cytometry monocyte phagocytic assays have been developed (15).However, these methods may overestimate phagocytosis due to the nonspecific binding of fluorescence-labeled platelets to monocyte surface (16,17).This problem could be overcome by using a pH-sensitive fluorescent dye (pHrodo) that only generates fluorescence signal when pHrodo-labeled platelets have entered lysosomes (acidic milieu) but not while they are in the cytoplasm or outside monocytes (18-20).However, this current technique is limited to the use of isolated phagocytic cells and therefore may not mimic the mechanism of antibody-mediated platelet phagocytosis in whole blood.
In this study, we developed a pHrodo-based platelet phagocytosis assay (PPA) using whole blood.Termed whole blood platelet phagocytosis assay (WHOPPA), this technique identifies the subset of phagocytic blood cells responsible for the phagocytosis of anti-HPA-1a opsonized platelets, to correlate the antibody titer with phagocytosis rate and to evaluate various inhibitors including anti-FcgRs and effector-silenced mAbs against HPA-1a.
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).Patient consent was waived due to the deidentified analysis of retrospective data or blood samples.

WHOPPA
Isolation, labeling of platelets with pHrodo, and sensitization by antibodies were described in the supplementary data.Aliquots (200 µL) of sensitized, pHrodo-stained platelets (1.5 x 10 7 platelets) were incubated with 200 µL of whole blood containing 5 x 10 4 monocytes for 60 min at 37°C and 5% CO 2 .After the erythrocytes in whole blood were lysed with lysis buffer (155 mM ammonium chloride, 10 mM potassium hydrogen carbonate, and 100 µM EDTA) for 5 min, the monocytes were labeled with 3 µL of APC-conjugated anti-CD14 mAb for 30 min in the dark.Viable monocytes were identified by flow cytometry after staining with 1 µL of SYTOX Blue (Invitrogen, Paisley, UK).All CD14 + and pHrodo + cells were defined as viable monocytes that phagocytosed sensitized platelets.Approximately 5,000-20,000 cells were evaluated.The phagocytosis rate was defined as the percentage of monocytes that engulfed platelets.In some experiments, the neutrophils were labeled with 3 µL of Pacific Blue-conjugated anti-CD16 mAb to examine platelet engulfment by neutrophils.Whole blood was incubated with 20 mg of pHrodo-stained E. coli bacteria as control (Life Technologies Corporation, Eugene, USA) at 4°C and 37°C and analyzed by flow cytometry using BD FACSCanto ™ II Flow Cytometer and BD FACSAria ™ III Cell Sorter and BD FACSDiva ™ software v6.1.3(Becton Dickinson Biosciences).To exclude false negative gating of neutrophils due to shedding of CD16 antigen from neutrophil surface, additional gating strategy independently from CD16 marker was performed to identify neutrophils and monocytes (see below).

Analysis of neutrophil and monocyte subsets
The neutrophils and monocytes were analyzed by flow cytometry using gating strategy as previously described (22).Leukocyte subsets derived from whole blood were separated by side scatter (SSC) and CD45 marker.The monocytes and neutrophils were identified as SSC low and SSC high, respectively.Cell aggregates ("doublets") and other nontarget subsets (T-, B-, dendritic, and dead cells) were excluded by forward scatter area (FSC-A) against forward scatter height and by negative selection using CD1c, CD3, CD19, and live/death markers (lineage cocktail), respectively.Finally, the viable neutrophils (CD66b + ) and monocytes (CD66b − ) were analyzed for pHrodo signal, FcgRI, FcgRII, FcgRIII, and HLA-DR expression.The monocyte subsets were gated with anti-CD14 (clone M5E2) and anti-CD16 (clone LNK16) as described (23).

t-SNE analysis of monocyte subsets
T-distributed stochastic neighbor embedding (t-SNE) analysis was performed as previously described (24).Individual fcs files were analyzed by FlowJo (FlowJo ™ v10.9 Software; BD Life Sciences).The viable monocytes were gated as described before (Supplementary Figure S3) and cleaned of time artifacts before the population size was normalized to 50.000 cells using the FlowJo plugin DownSample v3.3.1 (BD life Science).The fcs files were then concatenated into a single fcs file with a total of 100.000 cells to create a t-SNE map.Barnes-Hut implementation of t-SNE was used to apply the dimensionality reduction algorithm (25).Best separation of monocyte subsets was achieved by generating a global t-SNE map using CD64, CD32, and CD16 markers under the following running conditions (iterations: 5.000 iterations; perplexity parameter: 90; and learning rate: 9.000).t-SNE maps were generated by plotting t-SNE (x-and y-axis) in a dot plot.Color mapping was then applied by overlaying the fluorescence intensity for additional markers (CD14, HLA-DR, and pHrodo) to assess unknown map clusters.Within the heatmap overlays, the scaling of all markers was optimized uniformly, ranging from low (−1635.8786,blue) to high signal (84005.4883;red).In addition to the pHrodo heatmap, the overlays of pHrodo positive and high positive subsets were performed to evaluate map regions positive (compared with negative control) or highly positive (top 1% with highest fluorescence intensity) for pHrodo signals.Previously gated monocyte subsets were overlaid onto the t-SNE dot plot for subset phenotype comparison based on t-SNE map position.

Analysis of platelet antibody binding by flow cytometry
A 10 µL aliquot of sensitized pHrodo-labeled platelets in HBSS buffer was incubated with 1 µL of FITC-conjugated rabbit F(ab) 2 antihuman IgG (dilution 1:10; Dako, Glostrup, Denmark) for 15 min in the dark and analyzed by flow cytometry.The platelets were gated in forward/side scatter (FSC/SSC).Antibody binding was calculated based on the median fluorescence intensity (MFI).

Inhibition of platelet phagocytosis by effector silencing anti-HPA-1a and anti-FcgR antibodies
PHrodo-stained platelets were incubated with 10 µL of effector silencing anti-HPA-1a antibodies (8 µg) at 37°C before incubation with anti-HPA-1a antibodies.Deglycosylated IgG from isotype control or AB serum was run in parallel.For FcgR inhibition, whole blood or isolated monocytes were incubated with anti-CD64, anti-CD32, anti-CD16, isotype control (10 µg/ml) IgG, or an equal amount of F(ab')2 fragment for 30 min at room temperature before phagocytosis.
Since neutrophils could shed CD16 from the cell surface after platelet phagocytosis leading to false negative determination of engulfed monocytes, additional gating strategy independently from CD16 marker (Figure 1B; left panel)) was conducted.Analysis of different blood donors (n=13) showed the engulfment of opsonized platelets by monocytes but not by neutrophils in whole blood (Figure 1B, right panel).
Furthermore, the specificity of platelet phagocytosis by monocytes was supported by confocal and electron microscopy analysis (Figure S2).
Subsequently, monocytes were gated (Figure S3), and their subsets were analyzed in detail to study the role of different monocyte subsets on anti-HPA-1a-mediated platelet phagocytosis (Figure 2A; see also Figure S4).Three well-known monocyte subsets, namely, classical (CD14 ++ CD16 − ), intermediate (CD14 ++ CD16 + ), and nonclassical (CD14 + CD16 ++ ) monocytes, were detected when the platelets were opsonized with AB serum (control).A small unique monocyte subset (CD14 + CD16 − ) termed "shifted" monocytes was found solely after the engulfment of anti-HPA-1a opsonized platelets.This result could be confirmed by repeated experiments with different donors (Figure 2B).As shown in Figure 2C, all three monocyte subsets were capable of mediating platelet phagocytosis.Although classical monocytes (CD14 ++ CD16 − ) were the most abundant monocyte subset in whole blood, they showed significant lower phagocytosis capability than the intermediate (CD14 ++ CD16 + ).On the basis of the FcgR expression pattern, the shifted monocytes were most likely derived from all three monocyte nsubsets (Figure 2D).
Furthermore, we analyzed the expression of FcgRs and HLA-DR on different monocyte subsets (Figure 3).FcgRI (CD64) expression was found on classical and intermediate subsets but not on nonclassical and shifted subsets.The significantly reduced CD64 expression after the phagocytosis of anti-HPA-1a-opsonized platelets indicated the role of ligand-induced internalization mediated by CD64 receptor.A similar phenomenon was observed for FcgRIIIa (26).By contrast, FcgRII was detected in all monocyte subsets and was upregulated after platelet phagocytosis.Similarly, HLA-DR was upregulated in all monocyte subsets that had engulfed opsonized platelets, and this phenomenon was most probably associated with monocyte activation (27).Overall, these results suggested that FcgRI and FcgRIIIa are involved in platelet phagocytosis.

Anti-FcgRI and anti-FcgRIII antibodies inhibited the phagocytosis of anti-HPA-1a opsonized platelets in whole blood
Monocytes in whole blood were incubated with mAbs specific for anti-FcgRI (CD64), FcgRII (CD32), and anti-FcgRIII (CD16) and analyzed by WHOPPA to further evaluate the role of different FcgRs on anti-HPA-1a mediated phagocytosis (Figure 4).In accordance to the previous observation, the suppression of FcgRI and FcgRIII with anti-CD16 and anti-CD64 IgG antibodies significantly inhibited platelet phagocytosis when compared to isotype control.By contrast, anti-CD32 IgG did not block platelet phagocytosis.Analysis with F (ab')2 fragments showed that anti-CD16 F(ab')2 fragment did not inhibit platelet phagocytosis (IgG versus F(ab')2 p<0.001), whereas anti-CD64 F(ab') 2 still did (IgG versus F(ab')2 not significant).Although CD16 is expressed only less than 20% of monocytes, anti-CD16 IgG antibodies were able to reduce the phagocytosis rate by around half.This observation suggested that anti-CD16 could not only directly impair monocytes via Fab part (cis-effect), but also indirectly (trans-effect) via Fc part.

WHOPPA revealed a high correlation between bound antibodies and phagocytosis
Homozygous HPA-1a (+) and HPA-1a (−) typed platelets were sensitized with different dilutions of anti-HPA-1a standard serum (NIBSC 03/152; 1:16-1:512 dilutions) to further evaluate the specificity and sensitivity of anti-HPA-1a-mediated platelet phagocytosis by monocytes.Significant platelet phagocytosis was observed only with sensitized HPA-1a (+) but not with HPA-1a (−) platelets (Figure 5A), which remained with 1:32 diluted serum (Figure 5B).Bound anti-HPA-1a was still observed with 1:128 diluted sera, indicating that a certain threshold of antibody level is necessary to trigger platelet phagocytosis.Nonetheless, a significant correlation (r = 0.9256; p < 0.0001) was observed between antibody binding (MFI) measured by flow cytometry and phagocytosis rate (Figure 5C).All tested anti-HPA-1a samples (n = 7) induced allele-specific phagocytosis, but only HPA-1a (+), not HPA-1a (−), 01; Cutoff of phagocytosis is 2.5%; ns, not significant, (D) t-SNE-guided phenotype comparison of monocyte subsets.After the engulfment of platelets opsonized with AB serum or anti-HPA-1a, monocytes from one representative donor were gated and merged to create a single t-SNE map of 100.000 events.Dimensionality reduction analysis by t-SNE was performed as described in Materials and Methods.(A) Signal intensity of CD64 (FcgRI), CD32 (FcgRII), CD16 (FcgRIIIa), CD14, HLA-DR, and pHrodo projected onto the t-SNE space according to a heatmap color scale ranging from low signal (blue) to (high signal, red), (B) positive and high positive for pHrodo signal (red), and (C) Manually gated monocyte subsets (black) were overlaid onto the t-SNE map and highlighted by specific colors.Ungated subsets (light blue or gray) refer to all events.opsonized the platelets (Figure 5D).The phagocytosis rate was correlated with the quantity of platelet-bound antibodies (r = 0.9214, p < 0.0001) (Figures 5E, F).However, a low correlation was observed when antigen capture assay, MAIPA, was used to measure bound antibodies (Figure 5G).

Anti-HPA-1a antibody subtypes I and II, not III, caused platelet clearance
Two different subtypes (I and II) of anti-HPA-1a antibodies reacting solely with the polymorphic b3 subunit have been reported.Subtype I anti-HPA-1a antibodies recognize HPA-1a epitopes residing on the PSI domain, and subtype II antibodies react with complex epitopes formed by the PSI together with the EGF1 domain of b3 integrin (8,28).As shown in Figure 6A, subtypes I (mAb SZ21) and II (mAbs B2G1 and 26.4) induced the phagocytosis of HPA-1a (+) but not HPA-1a (−) platelets.However, subtype I mAb SZ21 at a high concentration (10 mg/ mL) also induced the phagocytosis of HPA-1a (−) platelets.This phenomenon is in accordance with our previous real-time antibody binding analysis using surface plasmon resonance technology, demonstrating that mAb SZ21 could also interact with immobilized HPA-1b antigen.However, this interaction was 100fold weaker than that with HPA-1a antigen (13).
Preliminary experiment showed that type III anti-HPA-1a antibodies did not inhibit platelet phagocytosis, most probably due to low expression of avb3 on platelet surface (data not shown).

Silenced mAbs against HPA-1a inhibited anti-HPA-1a-mediated platelet phagocytosis
Removing N-glycan (linked to asparagine 297) on the Fc part significantly weakens the binding of IgG to FcgRs and its ability to activate complement factor C1q (29).IgG antibodies were digested by EndoS to test the phagocytosis capability of deglycosylated mAbs against HPA-1a, and only a completely digested IgG fraction was used in this study (Figure S5).In contrast to naïve IgG, all deglycosylated mAbs (SZ21, 26.4, and B2G1) samples against HPA-1a were unable to induce platelet phagocytosis (Figure 6B).In the control experiment, Fc-silenced B2G1 delta nab did not induce platelet phagocytosis.Inhibition experiments were then performed to investigate whether Fc-silenced anti-HPA-1a mAbs could block platelet phagocytosis induced by anti-HPA-1a antibodies.Deglycosylated mAbs SZ21, 26.4, and B2G1 and B2G1 delta nab significantly inhibited anti-HPA-1a-mediated platelet  phagocytosis.A similar result was obtained with another Fcsilenced anti-HPA-1a mAb (26.4 IgG1 LALAP) (Figure 6C).

Discussion
Fetal bleeding in FNAIT cases is mostly attributed to maternal platelet alloantibodies, which opsonize fetal platelets and thereby enhance platelet clearance by FcgR-mediated phagocytosis in the reticuloendothelial system of the spleen (1).Different alloantibodies against HPAs associated with FNAIT are reliably detected by serological methods.However, the ability of these antibodies to trigger platelet clearance appears to be heterogeneous, and the mechanism underlying this phenomenon is not completely understood.Therefore, an in vitro PPA that mimics in vivo conditions must be developed to improve our understanding regarding the clinical relevance of platelet antibodyinduced thrombocytopenia.
In the last decades, a number of monocyte monolayer fluorescence-based PPAs have been developed (16,17).Among them, the most reliable method is PPA using pHrodo-labeled platelets that only generates fluorescence signal in acidic lysosomes milieu of phagocytosing cells (18-20).However, such monocyte monolayer approach is difficult to standardize (15) and does not allow for the simultaneous analysis of blood phagocytes.Here, we developed WHOPPA that abrogates these drawbacks, allowing for the specific analysis of engulfed platelets under standardized conditions that nearly mimic in vivo environment.
Given that WHOPPA uses whole blood, opsonized platelets could theoretically be phagocytosed not only by monocytes but also by other blood phagocytes, such as neutrophils.Here, we found that only monocytes, but not neutrophils, could engulf anti-HPA-1asensitized platelets.Although monocytes and neutrophils share many properties, they have distinctive morphologic and functional characteristics, including granules protein production, chemotactic response, and metabolic burst activity during phagocytosis (30).Furthermore, monocytes constitutively express FcgRI in contrast to neutrophils, which only express this receptor in response to inflammatory stimuli (30).The fact that neutrophils were capable to engulf platelets in both antibody and non-antibody mediated manner has been documented.Previous studies, however, were performed with either isolated monocytes (18, 19) or with activated platelets (31,32), but not with whole blood and nonactivated platelets as in the experimental setting of WHOPPA.Actually, the reason for this discrepancy is not yet clear.Monocytes are currently divided in three subsets according to the relative surface expression of CD14 and CD16.The classical subset represents the most abundant portion (80%) of the total monocyte population, and the remaining 20% express CD16 and are further classified into nonclassical and intermediate subsets (33).Classical monocytes are phagocytic without inflammatory attributes, and nonclassical monocytes display inflammatory characteristics and exhibit properties for antigen presentation.Intermediate monocytes are a minor transitional subset that displays phagocytic and inflammatory functions (34,35).In this study, we found that all three monocyte subsets can engulf anti-HPA-1a opsonized Although classical monocytes are the most abundant monocyte population in whole blood, they showed significant lower phagocytosis capability than intermediate and nonclassical subsets, indicating the importance of inflammatory response attributes in anti-HPA-1a-mediated platelet phagocytosis.A new monocyte subset (termed "shifted" subset) appeared after platelet phagocytosis most probably due to the loss of CD14 and CD16 expression after engulfment of platelets.
Furthermore, we detected the significant inhibition of platelet phagocytosis with anti-FcgRI and anti-FcgRIII but not with anti-FcgRII antibodies.The inhibitory effect of anti-FcgRIII F(ab')2 disappeared when F(ab')2 fragment was applied.Meanwhile, anti-FcgRI F(ab')2 remained inhibitory.Only anti-FcgRIII as intact IgG strongly inhibited platelet phagocytosis in whole blood, indicating the involvement of other FcgR-carrying blood cells.Wiener et al. (17,36) demonstrated the significant inhibition of anti-HPA-1a opsonized platelet phagocytosis by isolated monocytes with humanized mAb H22 specific for FcgRI but not with anti-FcgRIII (mAb 3G8).Furthermore, previous clinical studies showed that treatment with mouse and humanized anti-FcgRIII (mAb 3G8 and GMA161) for a patient with refractory immune thrombocytopenic purpura led to increased platelet counts, accompanied by neutropenia (37,38).These phenomena might be a result of apoptotic neutrophil clearance by macrophages due to the binding of anti-CD16 antibodies to FcgRIIIb expressed on neutrophils (39).
Our result is in accordance with a recent study demonstrating the importance of FcgRI and FcgRIII on the phagocytosis of platelets opsonized with anti-aIIbb3 autoantibodies from patient with immune thrombocytopenia by splenic macrophages (40).The less important role of FcgRII on anti-HPA-1a-mediated platelet phagocytosis was also indicated in an earlier work (17).Rijkers and coworkers showed that the cross-linking of FcgRIIa on platelets by antibodies (anti-HLA antibodies) could trigger platelet activation and subsequent platelet phagocytosis by macrophages (41).Whether certain anti-HPA-1a antibodies could also induce phagocytosis via this pathway remains unknown.When maternal anti-HPA-1a from different FNAIT cases were tested by WHOPPA, a strong correlation between phagocytosis rate and bound anti-HPA-1a was observed when antibody binding was measured by flow cytometry and MAIPA.However, a relatively low correlation was observed with MAIPA.This difference is attributed to the different detection systems: fluorescence-(linear reaction) versus enzyme-(nonlinear reaction) based assay.Titration analysis of standard anti-HPA-1a serum showed that platelet phagocytosis was no longer present at an antibody titer of 1:64, although the binding of anti-HPA-1a antibody was still measurable by flow cytometry.This observation revealed that a certain minimum number of antibody-antigen interaction is required to initiate platelet phagocytosis and served as an indication why platelet phagocytosis may not occur for platelets carrying low copy number of antigen, such as HPA-5 (42).Previous clinical studies, however, showed that the titer of anti-HPA-1a antibodies was not strictly correlated with fetal thrombocytopenia (43), indicating that other confounding factors such as antibody fucosylation (18) and CRP (19) play also important role.Besides this aspect, the question of whether the quantification of in vitro phagocytosis of anti-HPA-1a opsonized platelets by WHOPPA correlates with clinical outcomes, such as fetal or neonatal platelet count in anti-HPA-immunized mothers is another limitation of this study, that should be addressed in the near future.
Our previous study showed that the effector silencing of deglycosylated mAb SZ21 against HPA-1a (deg-SZ21) could prevent platelet phagocytosis by macrophages in the NOD/SCID mouse model (13).However, recent investigations showed that anti-HPA-1a epitopes are heterogeneous, which is expected for a polyclonal immune response (8,9).Therefore, a panel of anti-HPA-1a mAbs recognizing different epitopes is required to displace the binding of maternal anti-HPA-1a antibodies to prevent fetal platelet phagocytosis from the circulation.Using WHOPPA, we showed that both types (type I, mAb SZ21 and type II, mAbs 26.4; B2G1) of effector-silenced mAbs against HPA-1a did not induce platelet phagocytosis but could prevent platelet anti-HPA-1a-mediated platelet phagocytosis.However, complete inhibition could not be achieved by each of these antibodies alone or in combination, indicating that a selected panel of mAbs is necessary to inhibit the binding of maternal anti-HPA-1a antibodies.
Lastly, whether IgG-mediated platelet phagocytosis solely occurrs in the spleen remains unclear.Biburger and colleagues reported that splenectomized mice were still capable of depleting platelets, suggesting that splenic-resident phagocytic cell populations are not required for platelet depletion (44).Through this mouse model, the authors also demonstrated the nonrequirement of neutrophils and underlined the importance of monocyte subsets for IgG-dependent platelet depletion.
We established a simple, reliable, and specific in vitro assay that mimics platelet phagocytosis by monocytes caused by anti-HPA-1a antibodies in whole blood.This assay allows us to study the mechanism of antibody-mediated platelet phagocytosis under ex vivo conditions and evaluate the inhibitory capacity of different inhibitors as therapeutically strategies for the prevention of fetal thrombocytopenia in FNAIT in the future.

Sensitization of pHrodo-labeled platelets by antibodies
In brief, 500 µL aliquots of pHrodo-labeled platelets (4.5 × 10 7 platelets) were incubated with mAbs at different concentrations (0.01, 0.1, 1, or 10 µg/mL) or serum samples (final dilution 1:10) in polypropylene tubes (Sarstedt, Nümbrecht, Germany) for 30 min in the dark.IgG isotypes were run in parallel as a control.After being washed twice with 1 mL of PEP buffer (10 min, 800 g), the platelets were resuspended in 200 µL of Hank's Balanced Salt Solution (HBSS) containing calcium and magnesium (Anprotec) and 10 mM HEPES.A 10 µL aliquot of sensitized platelets was tested by flow cytometry to determine antibody binding.In some experiments, 250 µL aliquots of pHrodo-labeled platelets (3.75 × 10 7 platelets) were incubated with anti-HPA-1a (NIBSC 03/152) or AB serum (final dilution 1:16) for 30 min at room temperature in the dark and processed as described above.

Adjustment of monocyte concentration in whole blood
Leukocyte count in whole blood was measured using a cell counter (Sysmex KX-21N, Norderstedt, Germany), and monocyte percentage in this blood fraction was determined by flow cytometry.In brief, 100 µL aliquot of whole blood fraction was incubated with 15 mL of erythrocyte lysis buffer (155 mM NH 4 Cl, 10 mM KHCO 3 , and 10 mM EDTA, pH 7.4) for 5 min under gentle shaking.After centrifugation (5 min, 450 g), the cells were washed with PBS (5 min, 450 g), resuspended in 100 µL of PBS, and incubated with 5 µL of APC-conjugated anti-CD14 mAb for 15 min in the dark.The percentage of CD14-positive monocytes in the total leukocytes was analyzed by flow cytometry (FACS Canto II, Becton Dickinson, San Jose, USA), and the absolute number of monocytes was calculated by referring to leukocyte counts (see above).Afterward, 200 µL aliquots of whole blood fraction containing 5 x 10 4 CD14-positive monocytes were prepared for the phagocytosis of 1.5 x 10 7 platelets (ratio: 1:300 (monocytes: platelets) to mimic physiological conditions.

Microscopic analysis of platelet phagocytosis
Microscopic analysis of platelet phagocytosis was performed as previously described with minor modifications. 19In brief, 200 µL aliquots of PBMC (2 × 10 7 /mL in RPMI containing 10% FCS) were incubated in a microtiter well for 1 h at 37°C and 5% CO 2 .After 30 min, the platelets were added with 5 µL of APC-conjugated anti-CD14 antibody (see above) and 5 µL of SYTOX Blue (dilution 1:100) and incubated again for 30 min.The platelets were washed with 200 µL of warm HBSS buffer (Aprotec) for 10 min at 800 g.The adherent monocytes were added with 50 µL aliquots of HBSS buffer before being incubated with 50 µL of sensitized pHrodo-labeled platelets for 1 h at 37°C and 5% CO 2 .Platelet phagocytosis was then examined by THUNDER DMi8 Live Cell Imager (Leica, Wetzlar, Germany) at 100× magnification.

2
FIGURE 1(A) Anti-HPA-1a mediated platelet phagocytosis by monocytes and neutrophils.PHrodo-labeled platelets were sensitized with anti-HPA-1a standard serum (dilution 1:16) or AB serum (as control) and subjected to WHOPPA.Peripheral blood mononuclear cells (PBMCs) were gated by forward scatter (FSC-A) and side scatter (SSC-A).Viable cells were identified using SYTOX Blue staining.Total monocytes (blue) and neutrophils (green) were identified by staining with APC-conjugated anti-CD14 and Brilliant Violet 510-conjugated anti-CD16, respectively.The percentages of monocytes that engulfed platelets (pHrodo + ) are indicated.In the control experiments, pHrodo-labeled E. coli bacteria were subjected to phagocytosis by monocytes and neutrophils at 37°C and 0°C (negative control).(B) Anti-HPA-1a mediated platelet phagocytosis by monocytes and neutrophils Left panel (Gating strategy): Cell aggregates ("doublets") and other nontarget subsets (T-, B-, dendritic, and dead cells) were excluded by forward scatter area (FSC-A) against forward scatter height (FSC-H) and by negative selection using CD1c, CD3, CD19, and live/death markers (lineage cocktail), respectively.The viable neutrophils (CD66b + ) and monocytes (CD66b − ) were analyzed for pHrodo signal and FcgRI, FcgRII, FcgRIII and HLA-DR expression.Right panel: PHrodo-labeled platelets were sensitized with anti-HPA-1a standard serum (dilution 1:16) or AB serum (as control) and subjected to WHOPPA.The percentage of phagocytic monocytes (Mono) and neutrophils (Neu) derived from 13 biological replicates is presented.Bars indicate means ± SEM.Statistical significance was assessed by one-way ANOVA with Tukey posthoc test.***p < 0.001.

FIGURE 3 FcgR
FIGURE 3 FcgR and HLA class II expression of monocyte subsets.After the engulfment of anti-HPA-1a opsonized platelets, monocyte subsets that engulfed anti-HPA-1a opsonized platelets were gated for pHrodo − (open circle) and pHrodo + populations (closed circle) and analyzed for FcgRs and HLA-DR as indicated.The results of five biological replicates are presented.Median fluorescence intensity (MFI) was normalized by isotype or negative controls.Data distribution was analyzed using Shapiro-Wilk and Kolmogorov-Smirnov tests.Paired two-tailed two-sample t-tests were performed, and significance levels were indicated as * (p ≤ 0.05), ** (p ≤ 0.01), and *** (p ≤ 0.001).