Protection by hydroxychloroquine prevents placental injury in obstetric antiphospholipid syndrome

Abstract Obstetric antiphospholipid syndrome (OAPS) is mediated by antiphospholipid antibodies (aPLs, and anti‐β2 glycoprotein I antibody is the main pathogenic antibody), and recurrent abortion, preeclampsia, foetal growth restriction and other placental diseases are the main clinical characteristics of placental pathological pregnancy. It is a disease that seriously threatens the health of pregnant women. Hydroxychloroquine (HCQ) was originally used as an anti‐malaria drug and has now shown benefit in refractory OAPS where conventional treatment has failed, with the expectation of providing protective clinical benefits for both the mother and foetus. However, its efficacy and mechanism of action are still unclear. After clinical data were collected to determine the therapeutic effect, human trophoblast cells in early pregnancy were prepared and treated with aPL. After the addition of HCQ, the proliferation, invasion, migration and tubule formation of the trophoblast cells were observed so that the therapeutic mechanism of HCQ on trophoblast cells could be determined. By establishing an obstetric APS mouse model similar to the clinical situation, we were able to detect the therapeutic effect of HCQ on pathological pregnancy. The normal function of trophoblast cells is affected by aPL. Antibodies reduce the ability of trophoblast cells to invade and migrate and can impair tubule formation, which are closely related to placental insufficiency. HCQ can partially reverse these side effects. In the OAPS mouse model, we found that HCQ prevented foetal death and reduced the incidence of pathological pregnancy. Therefore, HCQ can improve pregnancy outcomes and reverse the aPL inhibition of trophoblast disease. In OAPS, the use of HCQ needs to be seriously considered.


| INTRODUC TI ON
Antiphospholipid syndrome (APS) is a systemic autoimmune disease, with thrombosis and/or pathological pregnancy as the main clinical features. APS can occur alone or can coexist with other autoimmune diseases. 1 The main clinical feature of pathological pregnancy, such as stillbirth, severe preeclampsia, spontaneous abortion, and foetal growth restriction, is obstetric APS (obstetric antiphospholipid syndrome, OAPS). 2 The pregnancy complications of OAPS are believed to affect the function of the placenta, damaging placental decidua cells and trophoblast cells, affecting placental implantation, material transport, and in severe cases, causing thrombosis and infarction in the placenta. It is a systemic autoimmune disease with or without thrombosis that seriously threatens the health of pregnant women and their foetuses. The pathogenesis of OAPS is not yet clear, the clinical heterogeneity is high, and there are many bottlenecks in diagnosis and treatment. [3][4][5] Anti-phospholipid antibodies (aPLs) are pathogenic antibodies to APS and are a heterogeneous group of antibodies against phospholipids and phospholipid-binding proteins. 6,7 For patients with suspected OAPS, it is recommended to determine lupus anticoagulant (LA), anticardiolipin antibody (aCL) and anti-β2 glycoprotein I antibody (anti-β2GP1 Ab). 1 In the normal population, the detection rate of aPLs is only 0%-0.5%, while the detection rate in patients with preeclampsia is 7%-17%. The detection rate can reach 5%-20% in patients with recurrent miscarriage. For positive aPLs among women, 15%-30% have foetal growth restriction during pregnancy, and approximately 50% have experienced recurrent miscarriages. 8,9 In OAPS, aPLs are an important cause of pathological pregnancy.
Among these pathogenic antibodies, β2 glycoprotein I (β2GPI)dependent antiphospholipid antibodies are considered to be the main pathogenic autoantibodies in OAPS. The target antigen β2GPI has a different tissue distribution and is highly expressed in trophoblast cells. Anti-β2GP1 Ab could bind to trophoblasts and decrease hCG secretion. 10,11 Some experiments have shown that antibodies affect the cell invasion function of trophoblast cells. 12,13 In in vivo experiments in mice, the antibody entered the blood circulation and rapidly deposited in the foetal sacs, affecting mouse foetal cortical neuron cytoarchitecture. 14 The antibodies isolated from patients could cause foetal loss. 15 Polyclonal antibodies, including high concentrations of anti-β2GP1 Ab infusion in pregnant CD39-or CD73-knockout mice, trigger an increase in miscarriages. This study concludes the possibility that perturbation of CD39 (NTPDase1) and CD73 (nucleotidase), could act as the "second hit" in the manifestation of APS. 16 In OAPS patients, placental inflammation and thrombosis or infarction are important pathogenic factors of pathological pregnancy. The anti-β2GP1 Ab can directly or indirectly cause these pathological pregnancy injuries.
Hydroxychloroquine (HCQ) was originally used as an anti-malaria drug and has now become popular in treating autoimmune rheumatic diseases. It has been shown to be beneficial to mothers and foetuses in some autoimmune diseases. 17,18 In primary or secondary antiphospholipid syndrome, HCQ has been shown to reduce aPLs titers in the plasma. 19,20 Prenatal exposure to HCQ in patients with anti-SSA/SSB seropositivity, could decrease the risk of foetal heart manifestations of neonatal lupus. 21,22 However, there are few clinical data about the use of HCQ in obstetrical primary APS. HCQ can be used for refractory OAPS where conventional treatment has failed (low dose aspirin and heparin are the first choice for the treatment of OAPS) and is often administered empirically by obstetricians. 23,24 Due to the lack of clinical randomized clinical trials and the unclear mechanism of action, this cheap and safe drug has not been widely used in OAPS patients.
HCQ was able to mitigate the features of endothelial dysfunction in preeclampsia induced by recombinant TNF-a in vitro. 25 HCQ reduced aPL binding to syncytiotrophoblasts and reversed aPLmediated disruptions of annexin A5. 26,27 Some of these in vitro studies demonstrated that HCQ can partially enhance the invasion function of trophoblasts and affect the release of inflammatory factors. It did not alter the release of hCG, sFlt-1, PlGF or hPL, the levels of IL-10 were increased, and the secretion of IL-1β, TNFα, IL-6 or MCP-1 did not change. 28,29 An in vivo study have shown that HCQ protects pregnancy in APS patients by inhibiting complement activation in placentas and foetal brains. 14 Therefore, how HCQ plays a role in OAPS has not yet been clearly studied.
Accordingly, we systematically evaluated the effect of anti-β2GP1 antibodies on trophoblast cell function and investigated the effects of HCQ in vivo and in vitro. We aimed to determine whether HCQ can reverse the adverse effects of antibodies on multiple functions of trophoblasts and whether HCQ plays a crucial role in the pathogenesis of adverse pregnancy and foetal outcomes.

| HCQ treatment on pregnancy outcome
We conducted a retrospective study. The records of all outpatients and inpatients diagnosed with antiphospholipid syndrome in the obstetrics department from July 2019 to December 2020 were searched. All patients met the Sidney Laboratory criteria for APS that was updated in 2006. 30  The patients were divided into two groups: one group received HCQ treatment in this pregnancy, and the other group was not exposed to HCQ. The basic data of the two groups of patients, including the patient's general condition, disease classification, pregnancy history and antibody titers, were recorded. The occurrence of adverse events was recorded and compared between the two groups of patients for this pregnancy.

| Cell culture
HTR-8/SVneo is a first-trimester human extravillous trophoblast cell line purchased from the American Type Culture Collection (ATCC; Manassas, VA). The cell culture medium included RPMI 1640 medium supplemented with 10% foetal bovine serum (FBS, Gibco) and 100 nmol/L penicillin/streptomycin (Gibco). The cells were all maintained in a cell incubator at 37°C and 5% CO 2 .

| The preparation of anti-β2GP1 antibody
We developed an anti-β2GP1 Ab in collaboration with Affinity Biosciences. The gene polypeptide sequence was determined using AbDesigner software as Trairak Pisitkun et al. previously described. 31 The polypeptide sequence is GRTCPKPDDLPF, and the required polypeptide chain was synthesized. Then, we coupled it with the immune antigen. Immunization of Balb/c mice was enhanced, and blood was collected to screen out immunized mice.
The mouse spleen was collected and crushed and was mixed with DMEM to form a cell suspension. The SP20 cells and splenic cells were mixed, and then preheated PEG was added to fuse the cells.
These cells were resuspended, and the fused cells were screened by ELISA to establish a monoclonal cell line. The fluid from the ascites in the mouse abdominal cavity was taken to identify and purify the antibody and to measure the concentration. Technical route of antibody preparation is shown in Figure S1. Before using the monoclonal antibody, its reactivity was investigated compared with polyclonal antibodies purified from APS patients. There was no significant difference between the two antibodies ( Figure S2). All aPL used in the experiment were prepared with an anti-β2GP1 Ab.
After dilution, 50 μl was added to each well of the 96-well plate and the plates were incubated at 37°C for 30 min to solidify the Matrigel.

| Scratch wound-healing assay
HTR-8/SVneo cells were evenly seeded in a 6-well plate (5 × 10 5 ) in an incubation room at 37°C and 5% CO 2 overnight. A 200 μl sterile pipette tip was scratched straight across the confluent monolayer to make a clear and equal-width scratch in the middle of each well. PBS was used to gently wash off the exfoliated cells. HCQ and aPL were added to the groups, and the dosages were the same as those mentioned above.
After 24 h, the cell migration area was observed by inverted microscopy (4 × objective) and was analysed using ImageJ software.

| Cell invasion
To assess the cell invasion, a two-chamber Transwell migration assay was used as previously described. Matrigel (BD Bioscience) was diluted 1:8 and was precoated with 60 μl in a chamber membrane at 37°C overnight. HTR-8/SVneo trophoblast cells (5 × 10 4 / well in 200 μl RPMI 1640 medium supplemented with 2% FBS and 100 nmol/L penicillin/streptomycin) were seeded in an 8μm pore size cell culture insert that served as the top chamber. The lower chambers (24-well culture plate) were filled with 600 μl of complete medium (RPMI 1640 medium supplemented with 10% FBS and 100 nmol/L penicillin/streptomycin). After a 24 h incubation, the upper chambers were removed, and cotton swabs were used to wipe the noninvaded cells. Trophoblast invasion across the membrane was fixed with 4% paraformaldehyde and was stained with crystal violet dye. The invaded cells were observed by an inverted microscope (Olympus, Tokyo, Japan) at a magnification of 100 ×, and five fields of view were randomly selected. The cell invasion number was estimated using ImageJ software.

| Mouse model construction of OAPS
The mouse model of OAPS was established by intravenous injection of antibodies. 100 μg/mouse of aPL antibodies were administered on Day 0 and Day 7 of pregnancy and at a dose of 50 μg/ mouse on Day 8. This model could maintain a stable antibody concentration in pregnant mice, which is closely consistent with the clinical condition. HCQ was dissolved in sterile distilled water and was administered by a sterile insulin syringe (100 mg/mouse/day) on Day 8-Day 14 in a group of pregnant mice. Based on the clinical medication guidelines, in which the recommended dosage of HCQ is 6.5 mg/kg/day, we used a dose of 100 mg HCQ/mouse/ day. 32 The control groups were treated with sterile saline instead of antibodies and HCQ.
Day15 was the day that pregnant mice were euthanized for dissection. The foetal resorption frequency was (number of foetal loss)/ (total number of foetuses and resorptions) × 100%. The uterus was dissected, and then the foetuses and placentas were removed and weighed. The required specimen was removed, washed in sterile PBS to remove the remaining blood, soaked in formalin or stored at −80°C.

| Pathological studies
After euthanasia, the placentas were removed and fixed in 4% paraformaldehyde. Low-concentration to high-concentration alcohol was used as a dehydrating agent to gradually remove the water in the tissue mass. Then, the alcohol was replaced in the tissue block with xylene before embedding the tissues in paraffin. The embedded paraffin blocks were cut into slices (5-8 μm), were placed on a glass slide, and were dried at 45°C. Before dying, xylene and different concentrations of alcohol were used for deparaffinization.
The slides were stained with haematoxylin and eosin for histological analysis.

| Statistical analyses
Data are expressed as the mean ± SEM. All data came from three independent experiments. Statistical differences between the groups were assessed using one-way anova with a subsequent twotailed Student's t-test. p Values < .05 were considered statistically significant.

| HCQ might ameliorate clinical-pathological pregnancies of OAPS
We conducted a retrospective study involving 96 patients to analyse whether HCQ has a positive effect on the pregnancy outcome of APS. After analysing the situation of OAPS patients from June 2019 to December 2020, HCQ can be considered to have a positive effect on the prevention of some obstetric complications. Table 1, HCQ was more likely to be used to treat patients with APS + SLE and patients with a history of recurrent miscarriage. No correlation was found between the positive rate of the two antibodies and whether to use HCQ treatment. Table 2 shows the pregnancy outcome with antiphospholipid antibodies in patients who were treated with or without HCQ. In the cases we counted, without using HCQ, the proportion of spontaneous abortions, foetal distress and amniotic fluid muddy increased. On the contrary, in the HCQ group, the incidence of preterm birth increased, and the incidence of miscarriage was low. Although some of them were not statistically significant, we observed that HCQ appeared to be able to reduce some placental complications. Foetal distress in particular, treatment of HCQ remained a protective factor after the removal of confounders. Does this mean that HCQ may play a role in placental pathological pregnancy?

As indicated in
From these data, we found that antiphospholipid antibodies can cause many placental pathological pregnancies, and HCQ treatment may be effective. To study whether HCQ reverses the harm of antibodies and protects the placenta to reduce the occurrence of abortion, we carried out the following in vitro and in vivo experiments.

| The safety profile of HCQ in trophoblast cell proliferation
The issue of the potential for toxicity by HCQ has been of particular concern for patients with OAPS. 32 The HTR-8/SVneo cell viability assay was the first objective to determine the appropriate concentration of HCQ to determine the therapeutic dose to prevent cytotoxicity. Figure 1B shows that 100 μg/ml HCQ significantly reduced cell viability, and 10 μg/ml, 1 μg/ml and 0.1 μg/ml were noncytotoxic doses.
Therefore, a concentration of 10 μg/ml were undertaken in all subsequent experiments. The viability of cells was not time-dependent.
Using the same method, the experiments of trophoblast cell proliferation were determined with a range of aPL concentrations ( Figure 1A). When compared to the untreated control, 50 μg/ml was a cytotoxic dose. In the following experiments, we used a concentration of 20 μg/ml, which is consistent with previous studies. There was no significant effect on cell viability at the concentrations of 2 kinds of experimental doses-APS 20 μg/ml + HCQ 10 μg/ml, APS 20 μg/ml + HCQ 10 μg/ml ( Figure 1C). Therefore, at therapeutic doses, HCQ does not affect the growth and proliferation of placental trophoblast cells. We found that the proliferation and activity of placental trophoblast cells were not targets of HCQ treatment of OAPS.

| HCQ reverses the adverse effects of aPL on tubule formation
OAPS is mediated by antiphospholipid antibodies and affects the placental structure and function. 8 To assess the effect of trophoblast  (Figure 2A). Treatment with HCQ (10 μg/ml, not 1 μg/ml) improved this effect. We found that the decrease in these indicators (junction, mesh, mesh area) caused by aPL was improved by the addition of HCQ therapy. For the junction and mesh area, HCQ can improve the adverse effects of APS and can have a significant effect but cannot completely reverse its effect and restore it to the normal control level. Interestingly, HCQ was particularly effective in improving the mesh area, reaching 99.1% of the control group ( Figure 2B).

| HCQ reverses the adverse effects of aPL on trophoblast invasion and migration
The

| The OAPS mouse model was successfully constructed by means of the aPL injection
An OAPS model was constructed to determine the pathogenicity of antiphospholipid antibodies in vivo. APS mouse models were established by an intravenous injection of aPL antibodies ( Figure 4A,B). It is important to note that OAPS mainly presents as pathological pregnancy rather than vascular thrombosis. 13 Compared with passive immunization, a tail vein injection allows antibodies to be present in the circulatory system and directly affects the maternal-foetal interface, which directly affects placenta development.
In our APS model, approximately 31% of the embryos did not survive 15 days after the start of pregnancy ( Figure 5A, Table 3).  The foetal resorption frequency increased compared with other APS models in which antibodies were given by an intraperitoneal injection or by a subcutaneous injection. 16,33 Additionally, compared with the patient's serum antibodies, which are difficult to purify due to their wide variety, the model formed by prepared antibodies had a higher abortion rate and was more stable. It is worth noting that despite the significant difference in the embryo absorption rate, the weight difference between P15 and P0 of each group of pregnant mice was not statistically significant.
However, there were significant differences between the groups of mice that were not pregnant after the vaginal plug. Compared with the controls, aPL significantly reduced the ratio of successful pregnancy after the vaginal plug, while HCQ treatment failed to restore the normal level ( Table 3). Whether this means that HCQ has a longer onset time than antiphospholipid antibodies or that HCQ mainly targets the developmental stage after placenta formation is worthy of further study.

| HCQ alleviates the pathological reactions in the placentas of OAPS mice
To judge the effect of APS on placental cell tissue and the thera-

| DISCUSS ION
Antiphospholipid antibodies cause placental-mediated pregnancy complications, leading to recurrent abortion, preeclampsia or stillbirth. 8,34 Current research suggests that aPLs, especially anti-β2GPI Ab, affect the invasion of trophoblast cells, causing superficial implantation of the placenta and placental insufficiency. 35 HCQ has been applied as a clinical treatment in pregnant women with APS, but the therapeutic mechanism is unclear. 43 Some studies have shown that HCQ can inhibit aPLs binding to trophoblast cells. 27,29 On the basis of previous research, our experiment improved the protective mechanism of HCQ on placental trophoblasts in many aspects ( Figure 6). Our study confirmed that HCQ could improve the invasion function of trophoblast cells, which was consistent with other studies. At the same time, we found that HCQ can also improve the damage of aPL to trophoblast cell formation, protect the formation of the placenta and whether PMPs is one of its therapeutic targets. This is our next research direction.
There are studies that suggest a ApoER2 is associated with the pathogenesis of APS. 57,58 By binding apoER2 to β2GPI, aPL plays a role in inducing foetal loss and IUGR. This is consistent with our conclusion that aPL damages the placenta and causes adverse pregnancy outcomes. Also, the study suggest apoER2 is abundantly expressed in the labyrinth of the mouse placenta. Could this be a potential therapeutic target for HCQ? We will continue to study it.
Recently, there has been interest in the potential efficacy of  project administration (supporting); resources (lead); supervision (equal).

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.