The human placenta shapes the phenotype of decidual macrophages

During human pregnancy, placenta-derived extravillous trophoblasts (EVT) invade the decidua and communicate with maternal immune cells. The decidua can be distinguished into basalis (decB) and parietalis (decP), the latter being unaffected by placentation. By defining a novel gating strategy, we report accumulation of myeloid cells in decB. We identified a decidua basalis-associated macrophage (decBAM) population with a differential transcriptome and secretome when compared to decidua parietalis-associated macrophages (decPAMs). decBAMs are CD11chi and efficient inducers of Tregs, proliferate in situ and secrete high levels of CXCL1, CXCL5, M-CSF, and IL-10. In contrast, decPAMs exert a dendritic cell-like, motile phenotype characterized by induced expression of HLA class II molecules, enhanced phagocytosis, and the ability to activate T cells. Strikingly, EVT-conditioned media are able to convert decPAMs into a decBAM phenotype. Cumulatively, these findings assign distinct macrophage phenotypes to decidual areas depending on placentation and further highlight a critical role for EVTs in the induction of pregnancy-tolerant macrophage polarization. Graphical Abstract Highlights In this study, we identified so far unrecognized, placenta-induced immune responses at the maternal-fetal interface. Altogether, we imply that placenta-derived trophoblasts induce a pregnancy-tolerant phenotype by suppressing antigen-presenting cell-like functions in maternal tissue macrophages.

isolates (right panel). The gates represent the frequency of low-density CD66b + neutrophils of the CD45 + PBMC matched areas that do not interact with villous placental tissues (decidua parietalis, decP) (Fig. 137 1A). The absence or presence of HLA-G, a marker for EVTs, served as an additional indicator 138 to distinguish decP from decB (Fig. 1, B and C), respectively. An overview of our tissue 139 sampling strategy is outlined in Fig. S1, A and B. First, decB and decP samples were 140 morphologically identified based on well-described characteristics (Loke, 1996). Decidua 141 parietalis tissue has a visible smooth mucosal covering (Fig. S1 B), shows the presence of 142 glands and luminal epithelium (Fig. S1 C) as well as a significantly higher cellular density when 143 compared to matched decB samples (Fig. S1 D). In contrast, we identified decB by its

175
Subsequent flow cytometry analysis of decidual cell isolates revealed two CD14 + populations, 176 one lacking CD163 and CD206 expression and the other highly positive for both tissue 177 macrophage markers (Fig. 2 A). While the latter CD163 + CD206 + population showed a typical genes including L-selectin (SELL), S100A12, S100A8, and S100A9 (Fig. 2 D and Fig. S3 D).

185
Indeed, IF-analysis shows that CD14 + S100A12 + monocytes are found in the lumen of decidual   (Fig. 2 E). By considering our established markers for tissue-residency 212 we were able to show a decB-associated accumulation of tissue macrophages (Fig. 2 F).

213
Finally, we termed macrophages obtained from different decidual compartments decBAMs and 214 decPAMs, respectively. These data show that placentation is associated with increased 215 numbers of true tissue-resident macrophages.

217
We next sought to determine the potential of decB to recruit neutrophils and macrophages.

218
Therefore, we generated conditioned media (CM) from cultivated patient-matched decB and 219 decP tissue samples and performed chemoattractant assays using isolated PBMCs or 220 neutrophils from peripheral blood samples. These assays revealed no differential effect 221 between decB-and decP-CM in the recruitment of CD14 + blood-derived monocytes (Fig. 3 A) 222 or other PBMCs, including T cell subtypes, B cells or NK cells (Fig. 3 A, Fig. S4 A). In contrast, 223 separate assays showed that decB-CM exert a significantly higher potential to attract blood 224 neutrophils (Fig. 3 B). Well in line, culture media of decB tissues contained high levels of well-225 described neutrophil chemoattractants including CXCL1, CXCL8, S100A8, and S100A9 showed significantly more CD14 + Ki67 + macrophages in decB tissues when compared with 228 decP sections (Fig. 3 D). Decidual tissue macrophages also expressed a wide array of cell 229 cycle markers including CCNA1, pRB, and CDK1 (Fig. 3 E). In addition, we observed the 230 presence of mitotic figures in tissue-resident CD14 + decidual macrophages (Fig. 3 E). Flow

262
These results suggest that decB-specific accumulation of tissue macrophages is driven by in 263 situ proliferation and that, as expected, neutrophils are specifically recruited from blood.

322
We determined secretion of cytokines and growth factors, which were selected based on our 323 previous analyses and their involvement in myeloid or neutrophil immunity and function. In amounts of S100A8, S100A9, and MMP9 (Fig. S6 B).

337
In summary, decBAMs differ substantially from decPAMs by a specific transcriptome and Our IF analyses show a strong difference in decBAM and decPAM appearance, the latter 342 exhibiting a more dendritic cell-like morphology (Fig. 5 A). Higher magnification clearly shows 343 an enhanced spindle-shaped morphology and formation of pseudopodia in decPAMs, implying 344 a more motile phenotype (Fig. 5 A). Interestingly, CD11c lo decPAMs also showed a more 345 prominent staining with antibody against HLA class II molecules (Fig. 5 B). Single channel and

392
By studying this phenomenon, we found that induction of CD4 + CD25 + FOXP3 + Tregs is not a 393 uniform feature of decidual macrophages since Treg induction was in particular executed by 394 decBAMs (Fig. 5 H). To further test for dendritic cell-like function, we studied the potential of 395 decidual macrophages to activate blood-derived T cells. To this end, we cultivated T cells from 396 healthy donors with decBAMs or decPAMs in a ratio of 10:1. In line with our previous results, 397 mainly decPAMs are able to trigger proliferation in both CD4 + helper T cells and CD8 + cytotoxic 398 T cells (Fig. 5, I and J). We observed no induction of CD8 + T cells when blood-derived T cells 399 were incubated with decidual stroma cells (DSCs), confirming that our assay is not disturbed 400 by reactivity against allogeneic HLA class I molecules (Fig. 5 K).

401
Altogether, we show that decBAM are impaired in their ability to migrate, phagocytose, and  (Fig. 6 A). Immunofluorescence-based analysis indeed showed that addition of EVT-CM 412 significantly increases CD44 expression of decPAMs in situ to a comparable level as detected 413 in decB tissue explants (Fig. 6 B). Next, we exposed isolated decPAMs to 50% EVT-CM and 414 found a significantly reduced motility and phagocytic activity (Fig. 6, C  affected Treg differentiation (Fig. 6 F), we found a significantly reduced decPAM-dependent 423 activation of CD4 + and CD8 + T cells upon treatment with EVT-CM (Fig. 6 G).

458 459
Given that EVT-CM alone did not suppress CD4 + or CD8 + T cell proliferation (Fig. 6 G) we

468
In summary, these data show that EVTs have the capacity to trigger a pregnancy-tolerant 469 phenotype in macrophages via their pleiotropically acting secreted factors.

472
In this study, we describe profound phenotypical and functional differences of tissue-resident 473 macrophages in dependence on their localization within the human decidua. Moreover, the 474 substantial difference between decBAMs and decPAMs in terms of transcripts, cellular 475 proteome, and secretome also translated into different functional phenotypes. Finally, we show 476 that EVT-CM is able to alter decPAMs into a phenotype functionally reminiscent of decBAMs.
477 Surprisingly, we identified a significant increase in neutrophils and macrophages at the

516
In particular, the latter two exert critical functions during pregnancy, for example, by inducing  Tregs. This phenomenon is well-described for IL10-and TGFβ-secreting anti-inflammatory 547 macrophages (Ruffell and Coussens, 2015) and has also been assigned to decidual

570
To test this, we incubated isolated decPAMs with CM of cultivated primary EVTs to mimic 571 trophoblast-dependent effects on decidual macrophages unaffected by placentation in utero.

572
Functional studies on viable tissue explants revealed that placenta-derived signals could 573 override the tissue-specific phenotype of decPAMs. In this setting, the addition of EVT-CM led

598
The generation of an immunosuppressive environment is decisive for a successful pregnancy.

599
However, our data impose that this phenomenon does not affect the entire decidua. In fact, we 600 find that macrophages further away from the placentation site, referred to as decPAM in this 601 study, show functionality of tissue-patrolling macrophages. In contrast, decBAMs residing at 602 the maternal-fetal interface exhibit significantly less pronounced APC-like function. Our study 603 further implies that this pregnancy tolerant phenotype is dictated by the placenta as EVTs can 604 be used to modify the functionality of decPAMs towards a state reminiscent of decBAMs.

643
Primary EVTs were isolated according to a previously published protocol (Haider et al., 2016).

644
Briefly, villi were scraped with a scalpel and digested in two sequential digestion steps using    (Table S1) overnight at 4°C. Finally, blots 679 were incubated with horseradish peroxidase-conjugated secondary antibodies (Table S1).

683
For analysis of immune cell distributions within decidual tissues, cells were isolated as 684 described above and stained for flow cytometry using the antibodies listed in Table S1. Immune

700
For staining, sections were blocked for 1 hour in TBS/T containing 5% normal goat serum (Cell

701
Signaling Technology) and then incubated with primary antibodies listed in Table S1 overnight 702 at 4°C. Subsequently, the slides were washed in TBS/T 3 times and then incubated with 703 secondary antibodies (2 µg/ml) and DAPI (1 µg/ml, Roche) for 1 hour at room temperature.

704
Images were acquired on a fluorescence microscope (Olympus BX50) using a Hamamatsu

714
To assess the effect of EVT-CM on macrophages in situ, explants were cultured in 50% EVT-

715
CM overnight, then fixed in 7.5% formaldehyde and processed for immunofluorescence as 716 described above. For proliferation assays, 10 µM EdU (EdU-Click 488 kit, baseclick) was 717 added to the explant culture and the EdU detection protocol was performed according to 718 manufacturer's user manual before blocking. Sections were co-stained with CD14, and 719 macrophage EdU incorporation was assessed using the ImageJ software.

738
The count matrices were calculated with STAR/htseq-count. Differential gene expression was