Previously, we have shown that pSTBM, but not mSTBM, induced PBMC from non-pregnant donors to produce inflammatory cytokines (TNF-α, IL-18, and minimal amounts of IL-12p70 and IFN-γ) 
. Here, we have used cytokine arrays to extend this study and have identified that a wider range of pro-inflammatory cytokines are induced by pSTBM as well as one (IP-10) which is inhibited. Furthermore, for some of the cytokines, there were significant differences in their production in pregnant compared to non-pregnant women. IL-18 is not included on this array, therefore previous observations of IL-18 induction were not confirmed. Also, neither IL-12p70 nor IFN-γ were altered on the cytokine arrays, which may be due to the sensitivity of the array, however on the basis of this result we decided not to include them in the further experiments in favour of those cytokines more greatly altered.
Interestingly, we showed that for the 36 cytokines studied the profile was very similar for eSTBM and pSTBM, but mSTBM were not as stimulatory. This is consistent with a recent report that pSTBM and eSTBM but not mSTBM induce monocyte production of IL-6, IL-8 and IL-1β 
. We found that pSTBM significantly induce PBMC production of TNFα, MIP-1α, IL-1α, IL-1β, IL-6, and IL-8, whether the donors were pregnant or not. However, mSTBM were not stimulatory. To generate mSTBM trophoblast tissue is mechanically removed from villi, tissue is suspended in salt buffer at 4°C and released microvesicles are collected from the supernatant. We believe this is less likely to produce biologically normal vesicles than placental perfusion or explant cultures, which are produced under more physiological conditions (O2
, temperature, buffered solutions). We are currently performing proteomic analysis to compare mSTBM and pSTBM and this will be reported in a subsequent paper.
Detecting circulating levels of cytokines is notoriously difficult and often conflicting reports have been published citing the increased or decreased quantities, or absence, in pregnancy. The six cytokines investigated here are shown in some studies to be elevated in the peripheral blood of pregnant women, compared to non pregnant women, or are found at increased concentrations locally at the placenta and are therefore considered important for successful pregnancy. Total IL-1 concentrations have been reported to increase in the serum of pregnant women during the second and third trimester and are significantly higher than in non pregnant women, but whether this was IL-1α or IL-1β was not defined 
. Together with TNFα, the IL-1 proteins IL-1α and IL-β are pro-inflammatory, and TNFα can also be detected at increased levels in pregnant women 
. MIP-1α is a chemokine, also known as CCL3, and is produced by first trimester placenta cytotrophoblast to aid recruitment of monocytes and natural killer cells to the decidua to enable efficient placentation 
. MIP-1α instigates acute and chronic inflammatory host responses by recruiting pro-inflammatory cells expressing its receptor CCR5 to the sites of injury or infection 
. Finally, IL-8, which is also a chemokine, functions as a chemoattractant and angiogenic factor, and hence helps drive immune responses. Taken together the induction of all these cytokines (TNFα, IL-1α, IL-β and MIP-1α) in a pregnant woman would lead to a generalised inflammatory status. To counteract this we have noted that anti-inflammatory cytokines are also modulated by STBM, which may prevent excessive inflammation. IL-6 levels are raised in the peripheral blood of pregnant women 
, IL-6 has both pro-inflammatory and anti-inflammatory functions and therefore its presence in pregnancy may be to prevent excessive immune activation as it inhibits both IL-1 and TNFα. Similarly, inhibition of IP-10 would be immunomodulatory, preventing type 1 cytokine release.
Placentas exert various immunomodulatory effects 
, and the shedding of STBM which could carry regulatory placental proteins, lipids or nucleic materials extends the immunomodulatory capacity into the maternal system. Immunomodulation of T cell responses have been previously described 
; STBM inhibition of allogeneic immune responses (MLR) is attributed to syncytiotrophoblast membrane glycoproteins 
and inhibition of PHA/ionomycin activation of T cells is dependent upon microvesicle expression of FasL (where T cell Fas expression is high, for example on the Jurkat cell line) and PD-L1, the ‘programmed cell death 1 ligand 1’ immunosuppressive molecule 
. In our study we observe that pSTBM are able to inhibit IP-10 production. IP-10 is a member of the CXC chemokine family, also known as CXCL10, and is involved with monocyte, T cell, natural killer (NK) cell and dendritic cell chemoattraction, as well as promotion of T cell adhesion to endothelium 
. It is induced by pro-inflammatory stimuli and is associated with the pathogenesis of various diseases, such as diabetes mellitus type 1 
as well as allograft rejection 
. It is also an anti-angiogenic factor 
and is associated with the development and continuation of Th1 responses; as its name suggests, the interferon-inducible protein 10 KDa (IP-10) is induced by the classical type 1 cytokine IFNγ, and its receptor, CXCR3, has higher expression on Th1 than Th2 cells 
. Concentrations of serum IP-10 have been found to be significantly higher in normal pregnant women than non pregnant women 
, and even higher levels circulate in pre-eclamptic women. In line with this, we detect more IP-10 production by PBMC from pregnant women than non-pregnant women. Gotsch et al
(2007) propose that elevated maternal serum IP-10 contributes to the anti-angiogenic state of pre-eclampsia (along with sFlt-1 and endoglin). Here we suggest that in normal pregnancy STBM are able to reduce the levels of IP-10 produced by PBMC, thereby enabling angiogenesis and skewing of immunity to type 2 responses that are important for healthy pregnancy. Finally, the cytokine arrays indicate STBM induction of G-CSF. This factor aids pregnancy success in patients with recurrent miscarriage or repetitive failed implantation in IVF cycles 
and has been shown to be induced by trophoblast derived microvesicles in vitro 
, further indicating that STBMs may be important for healthy pregnancy.
We have shown that pSTBM bind monocytes, and to a lesser extent B cells. In addition, we have shown that monocytes and B cells are able to rapidly phagocytose pSTBM. This suggests that the receptor for pSTBM on monocytes and B cells may be involved with phagocytosis as well as binding. Possible receptors include Toll-like Receptors, Receptor for Advanced Glycation End Products (RAGE) and Integrins. Further work is in progress to assess the nature of this interaction.
While pSTBM bind to both monocytes and B cells, it appears that the monocytes are responsible for the production of the majority of the most abundantly produced cytokines (TNFα, IL-6, IL-8 and IL-1β). Even though monocytes are stimulated to produce cytokines they do not up-regulate HLA-DR expression, and HLA-DR expression is reduced by monocytes stimulated by pSTBM. Increased HLA-DR expression upon antigen stimulation aids antigen presentation and hence T cell stimulation. T cells are not normally activated in pregnancy, and no placental derived antigenic peptide has been reported. However down-modulation of class II molecules would reduce the potential for antigen presentation and prevent T helper cell stimulation.
The minimal binding capacity of pSTBM to T cell and NK cells, may suggest that only a small proportion of these cells may be subjected to direct immunomodulation by placental microvesicles. As mentioned above, allogenic or mitogen induced T cell responses are inhibited by STBM, and NK cells are also reported to be subjected to STBM derived immunosuppression through MIC/ULBP protein binding to their NK receptor NKG2D 
. It is unclear why our preparations are not binding to NK cells at greater levels.
PBMC from normal pregnant women showed a significantly increased production of TNFα and IL-6 when compared to the responses from non-pregnant women. IL-8, IL-1β and MIP-1α also showed increased production but not to levels of significance. This suggests immune cells are primed to respond to STBM by pregnancy which supports our previous observations of enhanced monocyte activity in pregnant compared to non-pregnant women 
. The mechanism of this is unknown, but it could be due to the increased levels of pro-inflammatory cytokines and hormones released by the placenta, or prior exposure to STBM in vivo.
In conclusion, we have shown immune cells produce of a variety of proinflammatory cytokines in response to STBM stimulation, which may contribute to the increased inflammation seen in normal pregnancy. Interestingly we have also found that constitutive IP-10 production is inhibited by STBMs, which would encourage skewing of the immune system away from excessive type 1 cytokine responses in normal pregnancy.