Decreased level of TREM like Transcript 1 (TLT‐1) is associated with prematurity and promotes the in‐utero inflammatory response to maternal lipopolysaccharide (LPS) exposure

The occurrence of preterm birth is associated with multiple factors including bleeding, infection and inflammation. Platelets are mediators of hemostasis and can modulate inflammation through interactions with leukocytes. TREM like Transcript 1 (TLT‐1) is a type 1 single Ig domain receptor on activated platelets. In adults, it plays a protective role by dampening the inflammatory response and facilitating platelet aggregation at sites of vascular injury. TLT‐1 is expressed in human placenta and found in cord blood. We thus hypothesized that TLT‐1 deficiency is associated with prematurity and fetal inflammation.


INTRODUCTION
Preterm birth (PTB), defined in humans as birth of the conceptus before 37 weeks of gestational age, is the leading cause of neonatal morbidity and mortality worldwide. 1Premature infants are at increased risk of respiratory diseases, severe infections, and intraventricular hemorrhage and while it is known that the inflammatory response is often a trigger for PTB, its role is poorly understood.PTB can be the outcome of medical intervention in the case of severe fetal or maternal disease.
It can also be the result of spontaneous labor leading to premature rupture of membranes, or of premature rupture of membranes followed by uterine contractions.These later, spontaneous, elements are preceded by cervical ripening and dilation, the loss of uterine smooth muscle quiescence and weakening of the fetal membranes.The underlying mechanisms include oxidative stress, 2 hormonal, 3,4 or metabolic dysregulation, 5 activation of the hypothalamic-pituitary-adrenal (HPA) axis, 6 mechanical overload, 7 infection, 8,9 and bleeding.These all contribute to a final common pathway of inflammation. 10Maternal (e.g., decidua) or fetal (e.g., membranes, trophoblast) tissue inflammation transmits further inflammatory signals to mother and fetus.On the one hand, inflammation in these tissues leads to further inflammation and uterine contractions and cervical dilation.On the other, this process leads to the fetal inflammatory response syndrome (FIRS). 11is syndrome also includes the elevation of cytokines such as Interleukin (IL)−6 and inflammatory mediators such as C reactive protein, IL-1species or IL-17.Importantly, in humans, FIRS has been associated with bacterial, viral, fungal, and protozoan infection, and with noninfectious inflammation (reviewed in 11 ).FIRS is also associated with the dysfunction or abnormal development of several fetal organs including lung, brain, gut, and thymus.
In humans, bleeding has been associated with a higher risk of PTB. 12,13However, the relationship is likely complex 14 and dependent on the presence of both anti-and pro-coagulant molecules which contribute to the cross talk between bleeding, tissue damage, and inflammation.For example, structural damage to decidual vascular endothelium can lead to increased thrombin generation due to the expression of tissue factor by the decidua. 15High levels of coagulant molecules, such as thrombin, have been found associated with PTB. 16Thrombin can act via G-protein-coupled protease-activated receptors and, as observed in a mouse model, it activates the phosphatidylinositolsignaling pathway which results in the generation of cytosolic calcium oscillations and phasic myometrial contractions. 17Thrombin generation may also alter local (e.g., decidual) progesterone responsiveness, metalloproteinase expression and increase production of inflammatory cytokines. 15,18,19Moreover, this molecule plays an essential role in the clotting cascade and in platelet activation.
Platelets are a-nucleate cells from the megakaryocyte lineage that are classically slated as mediators of hemostasis.However, platelets not only release cytokines from their granules upon activation, but once activated, they adhere to neighboring leukocytes and modulate their immune function which highlights a role in the inflammatory process.Platelets have demonstrated roles in the initiation and exacerbation of cardiovascular disease, cancer and sepsis.Additionally, platelet dysfunction has been associated with intraventricular hemorrhage in preterm infants, 20 with severe hypertension-related disease in pregnancy, 21 and with functional progesterone withdrawal. 22Given the dual role of platelets in clotting and inflammation, modulation of platelet activation is an attractive target that may lower the risk of spontaneous PTB. 23e Triggering Receptor Expressed on Myeloid cells (TREM) like Transcript-1 (TLT-1) 24 is expressed exclusively in the platelet and megakaryocyte lineage. 25TLT-1 is stored inside the alpha granules of platelets and, upon activation, it is translocated to the surface of the cell where it binds the plasma clotting protein fibrinogen, a glycoprotein crucial for platelet-platelet interactions 26 While mice deficient in TLT-1 (treml1 −/− , KO) have longer bleeding times, reduced platelet aggregation and increased inflammation associated bleeding compared to controls, 26 TLT-1′s role in coagulation remains undefined.
TLT-1 mediates transmigration of neutrophils during inflammation and absence of the receptor impedes neutrophil transmigration into tissue.
This makes TLT-1 an attractive target for the treatment of syndromes such as lung injury where neutrophil infiltration is the hallmark of the disease.Because TLT-1 deficiency has been shown to exacerbate bleeding and inflammation, we hypothesized that TLT-1 deficiency is associated with prematurity and fetal inflammation.In this study, we used a pregnant mouse model of LPS exposure 27 to evaluate the role of TLT-1 in inflammation-induced PTB.

Human samples
This study of stored de-identified cord blood serum from human preterm and term singleton gestations was designated as not requiring a human subject protocol approval by the University of Vermont Institutional Review Board.Birthing persons were otherwise healthy.
Samples were thawed and assayed for soluble TLT (s-TLT).Limited clinical data on the samples assayed is presented in Supplemental Table 1.

TLT-1 ELISAs
Soluble TLT-1 levels for mice and humans were measured by ELISA (R&D Systems cat# DY2424 and DY2394, respectively) according to manufacturer specifications.ELISAs were developed using reagents from R&D Systems and read on a BioTek Synergy HTX multi-mode plate reader.

Mouse housing and study approval
Mice deficient in TLT-1 (treml1 −/− ) have been reported elsewhere. 26All mice were between 8 and 10 weeks of age and weighed 18 to 21 g.Animal care was provided in accordance with the procedures outlined in Guide for the Care and Use of Laboratory Animals. 28C57BL/6J and treml −1/− mice were housed and bred at an ICUAC approved animal care facility at the University of Puerto Rico, Rio Piedras Campus.

Breeding and induction of PTB
Timed mating occurred as has been done previously. 27Age-matched (8−10-week-old) wild-type or trem1l −/− male mice were placed in individual cages for one day before being removed and replaced with individual female mice of the same genotype.Synchronized breeding was conducted after the female mice had been in the cage for three days by placing the same male mouse that had been previously removed with the female for 24 h.The male mouse was subsequently

Tissue and sample harvesting
Pregnant females were euthanized, and maternal blood was collected and stored for assay of serum.Uterine draining lymph nodes (Paraaortic, inguinal) were removed for assay.Maternal uteri were removed and opened to reveal the existing implantation sites.The number of total implantation sites, the number of sites with a small discreet area of necrotic tissue (resorption), the sites with a placenta, but not a fetus (delivered) and the implantation sites with intact placenta and associated amniotic sacs with fetuses were counted.Using a 30-guage needle, amniotic sacs were entered and 50-100 μL of amniotic fluid was removed and stored at -80 • C. Then the amniotic sac was broken, and the fetal membranes were harvested.The placenta and decidua were removed and the white disc representing the decidua was separated from the placenta.In some cases, implantation (e.g., under the decidua) and non-implantation uterus was also removed.Fetal tissues were isolated from normal-appearing pups and included fetal liver and brain.Tissues were stored in 4% paraformaldehyde, RNA later, or frozen using liquid nitrogen.Decidua, uterine draining lymph nodes, and fetal liver were also used to generate single-cell suspensions for flow cytometry.

RNA extraction and quantitative PCR
Two protocols were used to determine RNA expression in harvested tissues.In one, total RNA was extracted from .5 to 1 mg of tissue using the PrepEase RNA spin kit from USB.The iScript cDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA)] was used to synthesize cDNA from 250 ng of RNA template using a mix of random hexamers and oligo dTs.From each sample 1 μL cDNA was used to amplify the genes of interest.QRT-PCR was performed on an ABI Prism 7000 (Applied biosystems-CA) using Power Sybrgreen master mix.Each sample was run in triplicate and the CTs were averaged.The primers used for amplification are presented in Table 1.
In alternative protocols the Qiagen RNeasy Mini kit (PN 74104) or the Qiagen RNeasy Plus (Qiagen 74134) kit were utilized as recommended by the manufacturer.RNA was quantified using Qubit HS DNA (Thermofisher Q32851).cDNA was synthesized as above.Relative

Slide preparation and staining of mouse placenta, decidua and fetal brain
Isolated mouse mid-gestation (∼day 16) tissues were fixed in 4% paraformaldehyde for 24 h and sent to Reveal Biosciences (San Diego, California) for paraffin embedding, sectioning and staining with Hematoxylin and Eosin.Heat induced antigen retrieval was performed using the Leica Bond Epitope Retrieval system.Non-specific antibody binding was blocked using Novolink Protein Block (Leica, cat#RE7280-CE, lot#6071120).Tissue sections were incubated with an antibody to fibrinogen (Abcam, ab92572, lot# GR29760-10) at a concentration of 1:250 versus a no-antibody control.Endogenous peroxidase activity was prevented using Vector Labs BloxAll (cat#SP-6000, lot#ZF1007).Voldemort-polyclonal rabbit antibody developed by our laboratory 29 followed by Alexa 647 anti-rabbit secondary (Jackson Immunoresearch, West Grove, PA).Antibody incubations were for 1 h in dark at room temperature.Slides were mounted in Fluoromount-G™ Mounting Medium (Invitrogen™ Cat # 00-4958-02).Confocal analysis was TA B L E 1 Primer sequences used in these studies.

Statistical analyses
The data was analyzed by unpaired two-tailed t-tests, one-way or twoway ANOVA and Bonferroni multiple comparisons.All analyses were done on GraphPad PRISM 9.0.0 for Windows and Mac OS.A p-value smaller than .05was considered statistically significant for t-tests, oneway or two-way ANOVA and multiple comparisons.

Analysis of human samples suggests a potential role of TLT in human PTB
To understand if TLT-1 may play a role in PTB we measured sTLT-1 levels by gestational age from human cord blood samples.Figure 1 suggests that plasma sTLT-1 increases with gestation and presents the possibility that TLT-1 may play a role in normal or abnormal gestation. 30

Analysis of TLT-1 deficiency in maternal tissue response to LPS in mice
The exact source of sTLT-1 in human cord blood is unknown.However, to begin to examine this issue, immunohistochemical analysis of the mouse placenta (Supplemental Figure 1) suggested that TLT-1 is related to trophoblast lining the intervillous space or cells (e.g., platelets) associated with that layer.
F I G U R E 1 Soluble TLT in human cord blood by gestational age.Concentrations present in cord blood at delivery were assayed by ELISA and analyzed by linear regression.A significant relationship between concentration and gestational age (p = .0027)was observed.
To begin to evaluate if TLT-1 plays a role in PTB we used the treml1 −/− mouse model to compare birth at 12 or 16 h after LPS treatment with wild type mice.On average, unmanipulated treml1 −/− mice have slightly larger litters than wild type mice (Figure 2A; 7.5 vs 6 pups) but this is not statistically significant (p = .8).At sixteen hours after injection of PBS, no mice were delivered (data not shown).Sixteen hours after injection of LPS, 2 of 4 treml1 −/-mothers delivered at least one pup (proportion or non-resorbed implantation sites delivered .6 and 1, Figure 2B) while no WT mothers delivered.No mice delivered by 12 h after LPS injection (data not shown).Histological examination of the placenta in response to LPS in treml1 −/-pregnancies as opposed to wild type pregnancies is shown in Figure 2C (top panels), where the absence of TLT-1 is associated with increased bleeding as potentially consistent with other studies, 26 and a tendency for increased deposition of fibrinogen as seen in Figure 2C (bottom).
Because inflammation is a driver of poor birth outcomes, and because of the role played by maternal decidua in pregnancy success, 31,32 we next evaluated the decidua for the infiltration of lymphocyte and granulocytic population by flow cytometry.Our gating scheme is shown in Supplemental Figure 2.While there were no significant differences in the overall size of the lymphocyte pool in this tissue (data not shown) there were proportionately less Gr1+(p = .008)and Gr1+/platelet conjugates (p = .01)in the decidua of treml1 −/-mice (Figure 3A).Platelet or neutrophil conjugates with T cells was similar in WT and KO decidua (Supplemental Figure 5).This may suggest trafficking to the myometrium, which would be consistent with an increase in PTB 33 or to the fetal membranes. 34,35Alternatively, it may represent a delay in granulocyte arrival that would have been observed at a later time point.Because inflammation has been implicated as a risk factor in PTB, we evaluated RNA expression levels of inflammatory cytokines at 12 and 16 h post exposure in the decidua Figure 3B.We observed that in the decidua of KO mothers exposed to LPS, there was a relative decrease in IL-10 at 12 h, suggesting a possible significant early deregulation of anti-inflammatory responses in the absence of TLT-1 as compared to that present in WT decidua.This difference however was no longer present by 16 h, suggesting other mechanisms might be important.
Fetal membranes promote entry into the parturition pathway and dysregulation of this tissue drives premature rupture, labor, and PTB. 36 thus examined the effect of early induction in both KO and wild type mouse fetal membranes (Figure 4A and B).We observed that between 12 and 16 h post LPS exposure there was an increase in IL-6, IFN-γ and TNF which trended towards significance in IFN-γ and TNF.This suggests that the fetal membranes in trem1 −/− mice may have a great capacity for an inflammatory response to LPS.
The placenta is anchored in the maternal decidua basalis and placental trophoblast is bathed in maternal blood.Therefore, this tissue could be a conduit for both infectious and inflammatory responses initiated in the uterus and the maternal systemic circulation.To understand if TLT1 plays a role in the inflammatory processes generated in response to intraperitoneal injection of LPS, we isolated this tissue at 12 h post injection (Figure 4C).Placental tissue showed higher IL-6 expression (p = .0365)and higher expression of TNF-α (p = .011)in KO as compared to WT pregnancies.This early difference resolved, however, by 24 h post injection when the remaining fetal placental units had similar levels of expression of several pro-inflammatory cytokines (Supplemental Figure 3).
In human studies, amniotic fluid inflammation and or the presence of inflammatory cytokines is a biomarker of PTB 37,38 and in mouse studies, injection of "alarmins" generate PTB. 39We hypothesized that there would be a TLT-1-related response to LPS in amniotic fluid.We observed that soluble TLT-1 was increased in the amniotic fluid of pups whose mothers were injected with LPS as compared to pups of PBSinjected mothers (Figure 5A).In addition, we observed that both IFNγ and IL-6 were elevated in the amniotic fluid of treml1 −/-pups whose mothers were exposed to LPS (Figure 5B).However, the cytokines IL-10, MCP-1, TNF and IL-12p70 demonstrated no significant differences between the two groups (Figure 5B).
Inflammatory mediators in the amniotic cavity can trigger signaling pathways which extend outward, but they can also be taken up by the developing fetus.We next examined the bodies of pups exposed to LPS via maternal injection.Firstly, because of the potential 40 but complex 41 role of TNF in dysregulation of the fetal brain, and because of data suggesting that amniotic inflammation rapidly generates a signal in this tissue, 40,42,43 we isolated fetal brain and measured RNA expression of TNF.We observed that exposure to LPS, as opposed to PBS, during pregnancy generated increased expression of TNF in the fetal brain of both WT and KO pups by 16 h post injection (Figure 6).However, the level of expression was not different in pups of either type.Because of the potential role of inflammatory cytokines IFN-γ 44,45 and IL-6 46,47 in shaping the developing immune system, we also examined the expression of these cytokines in the fetal liver, but we did not observe a difference between the pups of WT and KO pregnancies (Supplemental Figure 4).

DISCUSSION
In this study we began an investigation of the potential role of TLT-1 in inflammation-induced adverse pregnancy outcomes.First, we used Here we found an inverse correlation between the presence of TLT-1 and IL-6 and surprisingly, γ-interferon levels in the amniotic fluid.
While we have seen increased levels of IL-6 in various TLT-1 studies, this is the first association we have seen with IFN-γ.Increases in IL-6 may represent perturbation of the endothelium not protected by TLT-1 mediated platelet function.IFN−γ, however, represents a potentially new direction in TLT-1 regulated aspects of PTB.
In the same time frame as increased amniotic fluid cytokines, we observed a trend to increased inflammatory cytokines in the fetal membranes of LPS-exposed KO as compared to WT pregnancies.
Inflammation in the fetal membranes is a driver of PTB, 38 and in animal models, direct injection of inflammatory molecules into the amniotic cavity leads to PTB, 39 It is interesting that deficiency in TLT-1 may uncover a maternal-to-fetal-to-maternal inflammatory feed forward loop that could augment inflammation-induced PTB.Our data (Figure 2) does not argue against this hypothesis.Moreover, recent data raises speculation that the immune regulatory effects of TLT-1 may extend to adaptive immunity at the maternal-fetal interface. 48e finding that amniotic fluid demonstrated increased sTLT-1 levels after LPS treatment suggested that in addition to its activationregulatory role, 49 TLT-1 may help to maintain vascular integrity in the face of inflammation.Histological analysis of the placenta showed a significant amount of extravascular red blood cells in the treml1 −/− mice compared to controls.This is consistent with what has been seen in the lungs of treml1 −/− mice.This bleeding may drive PTB if it is associated with thrombin generation which itself may drive uterine contractions.However, the biology of hemostasis at the maternal interface is complex.
We were not able to detect differences in the pups of LPS-exposed WT versus LPS-exposed KO pregnancies.Further, we could not detect cytokine expression differences in fetal liver.However, we cannot rule out developmental differences which may be revealed by examination of surviving pups. 50,51 our knowledge, this is the first study evaluating TLT-1 function the context of inflammation-induced PTB.The study highlights the potential interaction between bleeding, hemostasis, and inflammation in normal and abnormal pregnancy and underlines the potential importance of platelets as well as immune and inflammatory cells.
The definitive impact and relationship of both platelet and TLT-1 function in the context of PTB remains unclear.We however speculate that these studies suggest the potential to assay of the presence of sTLT-1 in the peripheral blood or in amniotic fluid of human pregnancies to utilize as a diagnostic tool to assess the risk of PTB.This may also shed light on the biology underlying normal and preterm parturition.

2 F I G U R E 3 4 F I G R 5 6
Figure2B), in KO pregnancies as compared to WT there was an accelerated maternal-to-fetal inflammatory signature that began with alteration of the maternal decidual immune cell pool and decreased anti-inflammatory cytokines such as IL-10.This was accompanied by an increased expression of inflammatory cytokines, such as IL-6 and IFN-γ in the amniotic fluid of treml1 −/− pups when compared to control pups, consistent with a role for TLT-1 in regulating inflammation.