Evidence of macrophage modulation in the mouse pubic symphysis remodeling during the end of first pregnancy and postpartum

In mouse pregnancy, pubic symphysis (PS) remodels into an elastic interpubic ligament (IpL) in a temporally regulated process to provide safe delivery. It restores at postpartum to assure reproductive tract homeostasis. Recently, macrophage localization in the IpL and dynamic changes in the expression of inflammatory mediators observed from the end of pregnancy (D18, D19) to early days postpartum (1dpp, 3dpp) highlighted the necessity of the identification of the key molecules involved in innate immune processes in PS remodeling. Therefore, this study uses morphological and high-sensitivity molecular techniques to identify both macrophage association with extracellular matrix (ECM) remodeling and the immunological processes involved in PS changes from D18 to 3dpp. Results showed macrophage association with active gelatinases and ECM components and 25 differentially expressed genes (DEGs) related to macrophage activities in interpubic tissues from D18 to 3dpp. Additionally, microarray and proteomic analysis showed a significant association of interpubic tissue DEGs with complement system activation and differentially expressed proteins (DEPs) with phagocytosis, highlighting the involvement of macrophage-related activities in mouse PS remodeling. Therefore, the findings suggest that PS ECM remodeling is associated with evidence of macrophage modulation that ensures both IpL relaxation and fast PS recovery postpartum for first labor.

in mouse pregnancy, pubic symphysis (pS) remodels into an elastic interpubic ligament (ipL) in a temporally regulated process to provide safe delivery. it restores at postpartum to assure reproductive tract homeostasis. Recently, macrophage localization in the ipL and dynamic changes in the expression of inflammatory mediators observed from the end of pregnancy (D18, D19) to early days postpartum (1dpp, 3dpp) highlighted the necessity of the identification of the key molecules involved in innate immune processes in pS remodeling. therefore, this study uses morphological and highsensitivity molecular techniques to identify both macrophage association with extracellular matrix (ECM) remodeling and the immunological processes involved in PS changes from D18 to 3dpp. Results showed macrophage association with active gelatinases and ECM components and 25 differentially expressed genes (DEGs) related to macrophage activities in interpubic tissues from D18 to 3dpp. Additionally, microarray and proteomic analysis showed a significant association of interpubic tissue DEGs with complement system activation and differentially expressed proteins (DEPs) with phagocytosis, highlighting the involvement of macrophage-related activities in mouse pS remodeling. Therefore, the findings suggest that PS ECM remodeling is associated with evidence of macrophage modulation that ensures both IpL relaxation and fast PS recovery postpartum for first labor.
In mice, drastic modifications occur in the lower reproductive tract during pregnancy [1][2][3][4][5] . The pubic symphysis (PS) widens in a temporally regulated process to enlarge the space available for lower reproductive tract softening, ripening and expanding, providing both safe delivery of the offspring and maintenance of reproductive tract homeostasis 6,7 . During mouse pregnancy, PS extracellular matrix (ECM) remodeling gives rise to an interpubic ligament (IpL) that separates pubic bones and contributes to an optimal accommodation of the lower female reproductive tract [6][7][8][9][10][11] . After the first labor, PS histoarchitecture restoration provides mechanical stability to the reproductive tract and the ability to initiate and maintain subsequent pregnancies [12][13][14] .
Aberrant modifications of PS ECM molecules directly alter tissue biomechanical behavior, leading to the development of pelvic organ prolapse 15 . Therefore, the remodeling of mouse PS through pregnancy and postpartum without loss of pelvic floor homeostasis requires a delicate balance between deposition and degradation of the ECM 7 . The increase in compliance and extendibility of the IpL before expected parturition on day 19 of

Results
Dynamic remodeling of mouse pS at the end of pregnancy and early postpartum. At the end of pregnancy and early postpartum, PS remodeling results in a relaxed IpL that has pseudo-cavities ( Fig. 1A-C) filled with non-fibrillar versican/V0 variant-rich ECM (from now on referred as versican) and F4/80 + cells ( Fig. 1F-I). After 3dpp, the IpL rapidly remodels to tie the pelvic bones together in a near non-pregnant fibrocartilaginous joint (Fig. 1C), as it was described in previous studies 6,14,42 .
F4/80 + cells associated with active gelatinases and immunomodulatory ecM components during interpubic tissue remodeling. Considering that enzymatic activity at interpubic tissue remodeling leads to ECM structural changes 8,16 and that some ECM components can modulate macrophage differentiation and the function of Mmps 32 , we investigated the co-localization between F4/80 + cells and HA, versican, decorin or active gelatinases at the end of pregnancy and early postpartum. Additionally, to identify modifications in gelatinase production in interpubic tissue at this period, we analyzed Mmp-2 and -9 gene expression and protein production by microarray and proteomics (Fig. 2). (B,C) D18 relaxed, and 3dpp remodeled IpL containing pseudo-cavities (white arrowhead) closed the PB and osteoligamentous junctions (OJ), delimited by a grey dot line. (D,E) Light micrographs of pseudo-cavities (Pc) in both D19 relaxed and 3dpp remodeled IpL haematoxylin-phloxine B stained sections, which differ strikingly from well-preserved blood vessels (Bv). Note that in well-preserved resin-embedded specimens, the predominance of recruited mononuclear phagocyte-like cells (arrowhead) that may engulf material in the cytoplasm (detail) inside of pseudo-cavities in both D19 relaxed and 3dpp remodeled IpL. (D) D19 relaxed IpL pseudo-cavity border lined with a discontinuous layer of fibroblastlike cells (arrow) directly attached to IpL fibrous ECM. (E) 3dpp remodeled IpL pseudo-cavity border with more than one layer or discontinuous cell clusters associated with a dense collagen matrix (arrowhead). (F-I) Confocal immunolocalization of versican, VEGFR2 + cells, and F4/80 + cells present in pseudo-cavities and IpL tissue, counterstained with DAPI (blue). (F,G) versican V0 variant core protein immunodetection on D19 relaxed IpL ECM and inside of 3dpp remodeled IpL pseudo-cavity (green; marked with arrows), where immunolocalization was most intense surrounding recruited mononuclear phagocyte-like cells. (H,I) Double staining of F4/80 with VEGFR2 showed that F4/80 + -recruited phagocytes (red; marked with an arrow) do not co-localize with VEGFR2 + cells (green; marked with arrowhead) on D19 relaxed IpL, but F4/80 + / VEGFR2 + double-positive cells lining the 3dpp remodeled the IpL pseudo-cavity border (yellow; marked with arrows).   Fig. 2A-D) and with a distinct set of ECM components during interpubic tissue remodeling (Fig. 2G-R) from D18 to 3dpp. Together with in situ zymography identification of F4/80 + cells and gelatinase association at the end of pregnancy and postpartum, Mmp2 mRNA and protein levels were remarkably upregulated throughout interpubic tissue remodeling immune system biological processes involved in interpubic tissue remodeling. To evaluate how interpubic tissue DEGs and DEPs are involved in the immune system canonical pathways and biological processes during both IpL relaxation and postpartum remodeling from D18 to 3dpp, we utilized a relative analysis of transcriptomic and proteomic data. DEGs and DEPs (fold changes between 2 and − 2, p < 0.05) of whole interpubic tissues from D18 to 3dpp were obtained after normalization to D12 levels for use in enrichment data analysis (Fig. 3A).
Results from the global analysis indicated that DEGs and DEPs varied over distinct time points and 1dpp presented the highest amount of DEGs and DEPs (Fig. 3B,C). Enrichment analysis performed using the PAN-THER classification system (https ://www.panth erdb.org) correlated both DEGs and DEPs to the "immune system process" during interpubic tissue remodeling (Fig. 3D). Also, enrichment analysis by the MetaCore database showed that DEGs and DEPs were associated with distinct immune system response processes. However, while DEGs were significantly included in both the complement system inflammation process and development-related processes, DEPs were included only in the immune response of phagocytosis (Fig. 3E).
When analyzed per time point, the number of DEGs and DEPs associated with the immune system-related processes was higher at postpartum than at the end of pregnancy (Fig. 4A,B). The increased number of DEGs and DEPs associated with interpubic tissue remodeling at early postpartum was followed by expressive false discovery rate (FDR) and p values for each process listed by the MetaCore database (see Supplementary Table S1 online). Although the relationship between DEGs and DEPs at the listed immunological processes was indicated from D18 to D19 of pregnancy by the MetaCore database, these associations were not statistically significant (p > 0.05 and FDR > 10 -2 ).
Concerning immune response-related processes, the "transcriptional regulation of granulocyte development" (see Supplementary Table S2 online) and the "role of proteases in hematopoietic stem cell mobilization" (see Supplementary Table S3 online) indicated most of DEGs were downregulated. However the "complement system" indicated most of DEGs were upregulated (Table 1). Expression levels from the microarray analysis for some of the complement system-related genes were confirmed by qPCR, suggesting that the increased number of complement system-related DEGs were involved in IpL remodeling from D18 to postpartum (Fig. 5).
Once complement system activation and phagocytosis were found to be directly related to macrophage roles in homeostatic tissue remodeling processes 43 , an analysis of function designations for mice was performed using the DAVID (https ://david .abcc.ncifc rf.gov) database to identify interpubic tissue DEGs and DEPs directly related to macrophage activation, differentiation, function and regulation. The analysis identified 25 DEGs related to macrophage activities (Table 2) and none DEPs. Analyzing immune system-related processes, 13 DEPs were found, showing that although no directly related macrophage DEPs were observed in interpubic tissue remodeling, the modification in DEGs may reflect a signal for immune system-related protein production during this remodeling process (Table 3). In agreement with the hypothesis, our data showed that from D18 to 3dpp, DEGs directly related to macrophages were mainly grouped into processes, such as regulation of macrophage activation, macrophage activation, positive regulation of macrophage chemotaxis, positive regulation of macrophage differentiation and macrophage differentiation (Table 2). Similarly, DEPs were grouped into processes related to macrophage activity, such as immune system response to stimulus, antigen presentation and innate immune response differentiation (Table 3).
Although the number of macrophage-related DEGs ( Table 2) gradually increased from the end of pregnancy (D18 = 7 DEGs; D19 = 11 DEGs) to the first day after parturition (1dpp = 22 DEGs) in a similar profile observed at global analysis, DEPs (Table 3) were more abundant and markedly downregulated at D19, decreasing in number at postpartum (1dpp = 5 DEPs; 3dpp = 3DEPs). We also noticed that processes with inverse functions, such Figure 2. F4/80 + cells co-localize with gelatinase and ECM elements during interpubic tissue remodeling from D18 to 3dpp. (A-D) Positive fluorescent labeling of gelatinase activity (green) and F4/80 + cell immunolocalization (red) at osteoligamentous junction (OJ) and the interpubic ligament (IpL) at D18 (A,B) and 3dpp (C,D). F4/80 + cells co-localized with active gelatinase all over interpubic tissues at D18 and presented marked co-localization inside IpL pseudo-cavities at 3dpp. (E,F) Quantitative data of gene expression and protein production levels of metalloprotease (Mmp) with gelatinase activity (Mmp2-9) in interpubic tissues from D18 to 3dpp. While Mmp2 mRNA and protein levels were remarkably upregulated throughout interpubic tissue remodeling (E), the Mmp9 mRNA was downregulated, and its protein production was not detected from D18 to 3dpp (F). Data obtained by microarray and shotgun proteomic analysis, with a p ≤ 0.05 indicating statistical significance. as "positive regulation of macrophage cytokine production" and "negative regulation of macrophage cytokine production, " had exclusively upregulated DEGs. Additionally, "positive regulation of macrophage activation" processes presented both positive and negatively regulated macrophage-related DEGs from the end of pregnancy The total number of DEGs and DEPs progressively increased from D18 to 1dpp and decreased at 3dpp. Upregulated DEGs were abundant on D18, whereas positive DEPs were observed on 1dpp. Downregulated DEGs were abundant on 1dpp, while DEPs were more frequent at 3dpp. (D) PANTHER classification of biological processes by The Gene Ontology (GO) (a-m) associated with DEGs or DEPs identified by microarray and shotgun proteomic analysis during interpubic tissue remodeling. (E) MetaCore enrichment analysis for biological processes was completed for all time points and distributed the statistically significant data (p < 0.01/minimum FDR < 10 -2 ) for at least one day of study into processes related to immune system. The statistically significant biological processes were grouped by the authors before being displayed in the diagram.
Scientific RepoRtS | (2020) 10:12403 | https://doi.org/10.1038/s41598-020-68676-x www.nature.com/scientificreports/ to postpartum recovery (Table 2), highlighting a constant balance of macrophage-related gene activities during interpubic tissue remodeling. Complement system elements were listed as directly related to both macrophage DEGs (Table 3-C3 and C3a) and DEPs related to immune system processes such as phagocytosis (Table 2-CFB). This result suggests the possibility of macrophage and complement system interactions for IpL relaxation at the end of pregnancy and PS recovery postpartum.

Discussion
This paper highlights that PS ECM remodeling is associated with the temporal activation of innate immune response processes that may lead to the modulation of macrophage activity. This association has been demonstrated in a homeostatic remodeling of interpubic tissues that leads to safe delivery and fast PS histoarchitecture recovery after first labor, granting reproductive tract and pelvic floor integrity. Our light microscopy and SEM results illustrated the classical remodeling of interpubic tissues with pubic bone separation in labor by IpL relaxation 7 , as well as at the beginning of the postpartum restoration, which, in approximately 40 days, will give rise to a structure similar non-pregnant histoarchitecture 12 . Our morphological analysis also showed non-vascular pseudo-cavities in the IpL, filled with non-collagen ECM that presented www.nature.com/scientificreports/ similarities with previously described IpL sinusoidal like-cavities 14 or symphyseal space 6 with indistinct edges, composed mainly of newly differentiated (neutrophil 7/4 + cells) and mature macrophages (F4/80 + ) 36 . These IpL pseudo-cavities, which are easily distinguished from IpL ECM by their rich versican composition, are quite similar to the previous description of IpL sinusoidal like-cavities at the end of pregnancy 14 , which suggests that the plexus of large anastomosing cavities may accumulate fluids during the end of pregnancy, possibly helping IpL relaxation for labor 14 . Also, our results for versican V0 variant immunolocalization in IpL were in agreement  44 . Interestingly, only at postpartum, when IpL pseudo-cavities seem to be a well-organized structure and not just a random network of cavities, our results of IpL F4/80 + cells co-localized with VEGFR2, showing that a heterogeneous macrophage population participates in interpubic tissue remodeling. These macrophages presented phenotypic similarity with cells that can form vascular mimicry networks in matrigel-like ECM 45 , indicating that IpL macrophages may be engaged in the pseudo-cavity organization during IpL ECM remodeling. Although gene expression and proteomic assays agreed with previous findings of Mmp2 expression in IpL from the end of pregnancy to postpartum, the decreased expression of Mmp9 related to D12 shown by our microarray analysis pointed a distinct pattern from previous qPCR evaluations 16 . As in our proteomic analyses, previous blotting assessments of MMP9 production indicate low levels of MMP9 during PS remodeling 16 . These results highlighted that MMP2 is the main active form of gelatinase in the rich enzymatic milieu that conducts IpL remodeling. Consequently, it also determines that MMP2 was the active gelatinase detected in the in situ zymography experiments. Additionally, in situ zymography data showed that F4/80 + macrophages are one of the cells responsible for this active gelatinase production, first restrictedly associated with fibroblasts and chondrocytes 16 during PS remodeling.
Our double immunofluorescent results showed that IpL F4/80 + cell co-localize mainly with versican at the end of pregnancy and with HA, versican and decorin at postpartum. Although the presence and content of these three ECM components are well characterized in mouse interpubic tissues during pregnancy and postpartum Table 2. List of DEGs (− 2 > fold changes > 2) directly related to macrophage biological processes or functions at DAVID database and their fold changes from D18 to 3dpp. (n = 3/group).

Ensembl number
Gene ID Gene name Biological process or function  8,11,17 , this is, to our knowledge, the first description of their association with F4/80 + cells. Although a small amount of pro-inflammatory low molecular weight (MW) HA had been previously registered in IpL at the end of pregnancy 8 , apparently no co-localization between F4/80 + cells and HA was observed in our confocal results from D18 to D19. In the same period, we registered F4/80 + cells and versican co-localization in IpL. Even though the increase in versican gene expression from D18 to D19 is not associated with increased levels of versican-degrading enzyme ADAMTS1, previous data point a constancy in chondroitin and dermatan sulfated content in the IpL at labor 8 . This event is parallel to high levels of both the Mmp2 gene and protein in the IpL from D18 to D19, indicating that MMP2 may similarly cleave IpL versican molecules as previous observations in rabbit lung 46 . Therefore, F4/80 + cells may co-localize with cleaved versican molecules at IpL at the end of pregnancy and labor. At postpartum, the co-localization between decorin, versican or HA with F4/80 + cells was simultaneous with the highest MMP2 protein levels detailed here and the formerly reported constant gene expression of Hyal1 and Hyal2 in IpL 8 , suggesting that these ECM elements may also be cleaved at postpartum. MMP2 activity may release decorin and versican from the ECM to act as endogenous ligands of TLRs at macrophage, triggering sterile inflammation by enhancing the synthesis of the pro-inflammatory cytokines [31][32][33] . Interestingly, from D18 to D19, when F4/80 + cells co-localized with versican, both the presence of pro-inflammatory macrophages (F4/80 + /CD40 + ) 36 and high levels of NO, a marker of M1 activity 47 , were found in the IpL 48 . Additionally, from 1 to 3dpp, the co-localization of F4/80 + cells with decorin and versican is parallel to the previous register of increased levels of pro-inflammatory mediators in interpubic tissues 36 and the upregulation DEGs related to macrophage activation in our microarray analysis. Keeping in mind that in its soluble form, digested ECM becomes potent DAMPs acting as immunomodulators 32,33 , our results strongly suggest that decorin and versican interact with IpL macrophage and influence their pro-inflammatory activation during fast IpL ECM degradation and PS histoarchitecture. Furthermore, macrophage co-localization with versican can also indicate that these cells produce versican in IpL, a process associated with macrophage maturation and activation in vitro 44,49 . Although both possibilities may occur in the IpL, the functional enrichment of microarray data reinforces the first hypothesis, as several DEGs related to the macrophage activation process were identified in interpubic tissues from D18 to D19, and negligible levels of DEGs directly associated with the macrophage differentiation process were also observed.
Unlike decorin and versican pro-inflammatory signals, HA interaction with macrophage can lead to effects in pro-inflammatory or anti-inflammatory stimuli according to MW 32 . During labor and early postpartum, IpL content is mainly composed of high MW products 8 , capable of inhibiting leukocyte production of proinflammatory mediators 31,50 . Coincidently, immunolocalization of both F4/80 + cells with HA at postpartum is concomitant with both the presence of M2 macrophages (F4/80 + /TfR + ) and high levels of interleukin 10 (Il10) Table 3. List of DEPs (− 2 > fold changes > 2) related to immune system biological processes at DAVID database and their fold changes from D18 to 3dpp. (n = 3/group).  36 . These observations suggest that, on the one hand, degraded decorin and versican association with macrophages can result in M1 activation; on the other hand, IpL rich in high MW HA ECM may help in macrophage activity regulation by inducing an anti-inflammatory response. The presence of pro-inflammatory and anti-inflammatory signals provided by cleaved ECM components might explain the identification of upregulated DEPs related to the immune response to stimuli as well as both up-and down-regulated DEGs directly related to macrophage functional activities in IpL during postpartum. These expression patterns agree with the statement that the balance between pro-and anti-inflammatory signals along IpL remodeling regulates macrophage activity and ensures proper PS recovery at postpartum and the homeostasis of pelvic floor 36 .

Uniprot ID Protein symbol Protein name Biological process or function
As suggested by integrated data from gene and protein expression studies of cervical modifications at pregnancy and postpartum 51,52 , these similar essays may also help us to understand the immunological processes involved in mouse PS complex ECM remodeling at first pregnancy. Previous data 36,42 , along with our morphological analysis, support that immune system processes are statistically significant along with IpL remodeling by the enrichment analysis of microarray and proteomic experiments. Coincident with the highest number of F4/80 + cells and F4/80 expression levels in IpL 36 , the highest number of DEGs and DEPs related to immune system-related processes found here during postpartum remodeling, particularly at 1dpp, also corroborate our F4/80 + cell and ECM component co-localization data. Macrophages may be essential for proper remodeling of interpubic tissues because Mrc1 was found among our upregulated macrophage-related DEGs data in the IpL at 1dpp, while previous studies show the ablation of both Mrc1 and Asgr1/2 resulted in IpL relaxation failure, the arrest of labor and 61% of death in pregnant mice at parturition 40 .
The transcriptional regulation of granulocyte developmental processes from D18 to 3dpp in IpL shows constant inhibition of transcription factors (Myb, Mxd1, and Gata1), cell differentiation markers (Csf3r and Ncf1) and neutrophil activation markers (Ltf, Elane, Prtn3, Mpo, and Cybb), similarly with induced myeloid leukemia 53 or genetically modified animal neutropenia 54 . Moreover, downregulation of DEGs (Itga4, Cstg, Cxcl12, and Cxcr4) in the role of proteases in hematopoietic stem cell mobilization processes is strongly associated with impaired transmigration, activation and survival of granulocytes 55,56 , and migration of other leukocytes in association with glycosaminoglycans 57 . These inhibitions are congruent with previous morphological descriptions that neither neutrophils nor eosinophils are recruited to remodel interpubic mouse tissues from pregnancy to postpartum 12,35 . Then, among immune system-related processes associated with PS remodeling, complement system activation and phagocytosis presented the highest FDR by MetaCore, and some DEGs, such as C3 and C3ar, were directly associated with macrophage functions by DAVID analysis. The identification of phagocytosis as a significant process involved in IpL remodeling was also noted in light microscopy analysis and by recent electron microscopy data that showed phagocytic macrophage-like cells with phagosomes filled with engulfed materials in their cytoplasm, immersed in IpL ECM mostly at postpartum 36 .
Even though pregnancy complications, such as recurrent miscarriages, preterm birth and preeclampsia have been associated with excessive complement activation in animal models and humans at systemic levels 58 , local production of complement system components has indicated that complement might mediate distinct finely tuned roles in the context of tissue homeostasis and immunosurveillance in a context-driven fashion, independent of its systemic activities 43,59 . Except for C7 and serping1, upregulated DEGs involved in complement system processes detected in IpL were previously shown to be expressed by naïve and different activated subsets of bone marrow macrophages, mainly by M0, M1 and M2b subsets, in vitro studies 60 . From D18 to 3dpp, cartilage and bone tissues undergo extended reabsorption and phenotypic modulation processes, leading to less-differentiated cell composition of osteoligamentous junctions from late pregnancy to early postpartum recovery 13 . This result decreases the possibility of previously reported production of C1s, C2, C3ar and C3 by mature cartilage and bone cells 61 . The absence of C5 and C5ar from D18 to 3dpp and the high levels of C1q in postpartum the IpL particularly strengthen the possibility of the local production of complement components by IpL macrophages, once naïve, or activated macrophages do not produce C5 in vitro, and bone marrow macrophages express higher levels of C1q 60 . According to this hypothesis, the increased local expression of complement system elements found in the IpL and the parallel increase in macrophage number and expression of macrophage-related cytokines 36 suggest the existence of bidirectional feedback loops between complement elements and macrophage activation for proper PS remodeling.
Besides, although small osteoclasts were initially associated with pubic bone resorption at D14 14 , our results demonstrated that the Ncf1 osteoclast maturation-related gene is downregulated in the IpL from D18 to 3dpp. Parallel to Ncf1 downregulation, a progressive increase in F4/80 expression in the IpL was previously described during this period 36 . Once NCF1 is an essential subunit of the ROS-producing NOX 2 complex that contributes to osteoclast formation by possibly regulating pre-osteoclast maturation 62 and F4/80 acts as a suppressor of osteoclast maturation 63 , our results suggest that osteoclasts may not contribute to complement system component expression in interpubic tissues from D18 to 3dpp.
Keeping in mind that the complement system is critical to tolerogenic perception of apoptosis in many physiological processes 59,64 , the increase in both complement system activation DEGs and phagocytosis-involved DEPs shown by our enrichment analysis during the IpL remodeling process may be related to the regulation of apoptotic cell clearance. This is associated with the increase in cell death in interpubic tissues from labor to 5dpp 65 , phases of drastic remodeling of IpL and osteoligamentous junction components due to the breakdown of ECM to proper IpL relaxation for labor and fast restoration of PS histoarchitecture after delivery 7,13 . Similarly to complement system processes that can contribute to homeostasis by promoting tissue repair 43 , we found upregulation of initial elements of classical (C1q, Cr1, Cs1, C2, C3, C3ar) and lectin (Masp1) complement system pathways without the involvement of the C5-C5aR axis at IpL postpartum remodeling. These results are associated with early descriptions of increased Il10 expression from 1 to 3dpp 36 and are consistent with the immunological silent phagocytosis process mediated by C1q. In this process, the clearance of C1q and C3b-C3bi opsonized dying Scientific RepoRtS | (2020) 10:12403 | https://doi.org/10.1038/s41598-020-68676-x www.nature.com/scientificreports/ cells induces an anti-inflammatory program, increases the expression of immune checkpoint molecules and prevents the upregulation of maturation markers 64 . These results strengthen the hypothesis that complement system activation leads to phagocytic clearance of dying cells in PS remodeling. As discussed above, PS remodeling involves drastic ECM modifications, including ECM component solubilization by enzymatic activities 7 . Interestingly, it was found that ECM components, such as soluble decorin and high MW HA 66,67 , and Mmps, such as MMP2 and MMP12 68,69 , can directly interact with complement components, impairing their activity and limiting tissue damage after injury. Coincidently, previous studies showed that the IpL is rich in decorin and high MW HA 8 to the same period that increased Mmp12 expression levels (1dpp), elevated amounts of MMP2 (D18 to 3dpp) and increased expression of complement system components were observed in this study, indicating that DEPs are only involved in phagocytosis processes. Considering that decorin, HA and MMP2 can decrease C1q activity [66][67][68] and that MMP12 can degrade C3a 69 , it is possible that IpL ECM remodeling gives rise to a microenvironment that regulates both complement system activation and macrophage activity, assuring interpubic tissue remodeling without loss of homeostasis.
In summary, morphological and molecular findings suggest that mouse interpubic tissue modification for labor and postpartum recovery resembles the closure of a wound through an extensive tissue remodeling 59 , in which rigorous tissue remodeling involving protease-mediated degradation of ECM and rearrangements with intense cell turnover. This process induces the local release of danger signals that can activate immune cells, such as macrophages, into distinct phenotypes and modulate their function temporally from D18 to 3dpp. Active IpL macrophages release other immunomodulatory factors, such as cytokines and complement system elements, and activate innate immune pathways that can regulate the clearance of tissue turnover products (e.g., cell debris, apoptotic cells) without immune response activation, granting IpL relaxation for labor and fast mouse PS recovery to maintain pelvic floor integrity.

Materials and methods
Animals. Virgin female C57BL/6/JUnib mice (3 months old) were obtained from the Multidisciplinary Center for Biological Investigation on Laboratory Animal Science at UNICAMP. Mating and vaginal plug observation at day one of pregnancy (D1) were performed as described by Castelucci and coworkers 13 . Interpubic tissues obtained on days 12, 18, and 19 of pregnancy (D12, D18, D19), one and three days postpartum (1dpp, 3dpp) were arbitrarily distributed for morphological and molecular analyses. Three animals per group were used to obtain the interpubic tissues for light, electron and confocal microscopy analyses, and three (D18, D19, and 1dpp) or six (D12 and 3dpp) animals were used in the microarray, proteomic shotgun analysis, and realtime PCR analysis. In total, 102 animals were used for the experimental analyses. The animal experiments were conducted following the Guide for the Care and Use of Laboratory Animals issued by the National Institutes of Health (Bethesda, MD). All protocols using mice were approved by the Institutional Committee for Ethics in Animal Research (IB/UNICAMP, protocol 3789-1). and double immunostaining of F4/80 and VEGFR2, decorin or HA, interpubic tissues from three animals per day of study were frozen in n-hexane with liquid nitrogen and sectioned in the anteroposterior direction. The sections were fixed with acetone at − 20 °C for 3 min. Specifically for F4/80 and versican (V0 variant core protein) double immunostaining, samples were fixed in Carnoy's mixture (i.e., ethanol:chloroform:acetic acid 60:30:10 by volume) for 24 h at 4 °C and cryoprotected in a crescent sucrose gradient (10-30%) at 4 °C for 2 days. After the tissues were embedded in Tissue-Tek O.C.T. Compound (Sakura Finetek, USA, Inc.), they were finally frozen in n-hexane with liquid nitrogen and sectioned. Single versican immunostaining was performed according to Consonni and coworkers 9 , applying rabbit anti-mouse versican primary antibody (1:300/sc25831, Santa Cruz Biotechnology Inc., California, USA) and goat anti-rabbit secondary antibody (1:800/ab150077, AlexaFluor 488, Abcam plc., Cambridge, MA, USA). Double immunostaining was performed as described by Castelucci and coworkers 36  Microarray analysis. PS and IpL samples were removed and rapidly frozen in liquid nitrogen. Total RNA was extracted from D12 to 3dpp interpubic tissues following the RNeasy Mini Kit (QIAGEN, Hilden, Germany) manufacturer's instructions, and the RNA concentration was evaluated using a NanoDrop spectrophotometer, with the RNA concentration adjusted to approximately 1 μg/μl. Gene expression analysis was performed using the GeneChip Mouse Transcriptome Assay 1.0 Kit (Affymetrix, Thermo Fisher Scientific Inc., Massachusetts, USA). Microarray data analysis was performed using the R programming language. The linear model for microarray data analysis (LIMMA) was used to identify DEGs between D12 and other days of study (2 < fold changes < -2; a p < 0.05). Enrichment analysis for biological processes and immune system-related pathways and processes was performed using the PANTHER classification system and one quick analysis toll using the Meta-Core software (Clarivate Analytics, Boston, MA, USA). For the immune system biological processes found in the MetaCore database (Clarivate Analytics), only the processes in which DEGs were grouped with a false discovery rate (FDR) < 0.01 were selected for analysis. DEGs associated with macrophage-related processes and functions were selected with the Database for Annotation Visualization and Integrated Discovery (DAVID) analysis.

Histology. PS and
proteomic shotgun analysis. PS and IpL samples were removed and rapidly frozen in liquid nitrogen.
Total protein was extracted from D12 to 3dpp interpubic tissues using the PlusOne Sample Grinding Kit (GE Healthcare, Little Chalfont, Bucks, UK) in extraction buffer (4 M urea, 1 M thiourea, and 70 mM DTT). Buffer exchange into 50 mM ammonium bicarbonate was carried out using spin columns (Millipore, USA) and then quantified using a NanoDrop spectrophotometer. Samples were redigested with trypsin (Promega, USA) at a ratio of 1:50 (w/w trypsin/protein). The samples were suspended in water, and peptides were quantified using a Qubit Fluorometric Quantitation System (Thermo Fisher Scientific). Samples were diluted to 1 µg µL −1 and injected on a 2D ACQUITY UPLC M-Class system (Waters Corporation, Manchester, UK) coupled to a SYN-APT G2-Si Mass Spectrometer (Waters Corp.). Three micrograms of the sample were injected onto an XBridge BEH130 C18: 5 µm × 300 µm × 50 mm column (Waters Corp.) for peptide fractioning and elution to the analytical column. Peptides were eluted from the first column using three discontinued fractions containing 13.7%, 18.4%, and 50.0% acetonitrile on ammonium formate, pH = 10. After fractionation and elution on the first column, peptides were separated on an HSS T3: 1.8 µm × 75 µm × 150 mm column (Waters Corp.) using a gradient of 7% to 40% acetonitrile/water containing 0.1% formic acid for 54 min, followed by 10 min of 85% acetonitrile/water at a flow rate of 0.4 µL min −1 . Mass spectrometry was performed using nanoESI ionization in positive mode and data acquisition using HDMSE (Waters Corp.). The mass spectra were acquired from 50 to 2000 m/z. Protein identification and quantitation were performed using Progenesis QI for Proteomics v3.0 software (Waters Corp.).
The protein search was performed against the mouse revised databank from UniProt, using the following search parameters detailed by Silva and coworkers 73 . After identification, DEPs were associated with biological processes and immune system-related pathways and processes using the PANTHER classification system and one quick analysis toll in the MetaCore database (Clarivate Analytics). For immune system biological processes produced by MetaCore (Clarivate Analytics), only processes in which DEPs grouped with false discovery rate (FDR) < 0.01 were selected for analysis. DEPs associated with macrophage-related processes and functions or with immune system processes were selected for DAVID analysis.

Real-time pcR (qpcR).
For gene expression analysis of complement system components C2, C3, C5, C7, and Masp1, total RNA was extracted from frozen D12 to 3dpp interpubic tissues using the RNeasy Kit (QIAGEN). After cDNA synthesis using a RevertAid H Minus First Strand cDNA Synthesis Kit (Fermentas, Maryland, USA) according to the manufacturer's recommendations, real-time PCR was performed using SYBR Green qPCR Master Mix (Applied Biosystems, Foster City, CA, USA) on an Applied Biosystems 7,300 Real-Time PCR System (Applied Biosystems). Each gene was normalized to the expression of the housekeeping gene 36b4. To quantify and acquire the fold increase in gene expression, the mathematical model 2-ΔΔCt was utilized as performed by Consonni and coworkers 9 , and the results were normalized to that of the D12. All primers (see Supplementary Table S6 online) were purchased from Applied Biosystems. Three (D18, D19, and 1dpp) or six animals (D12 and 3dpp) were used to obtain 200 ng of total RNA from each day of study, and all reactions were performed in triplicate on the same plate.
Statistical analysis. The relative gene expression of qPCR analysis was performed based on Conover and Iman 74 and Montgomery 75 using rank data classification and semiparametric analysis (one-way ANOVA followed by Tukey's test or two-way ANOVA followed by Bonferroni's test) with p < 0.05 of significance performed Scientific RepoRtS | (2020) 10:12403 | https://doi.org/10.1038/s41598-020-68676-x www.nature.com/scientificreports/