Ccl2/Ccr2 signalling recruits a distinct fetal microchimeric population that rescues delayed maternal wound healing

Foetal microchimeric cells (FMCs) traffic into maternal circulation during pregnancy and persist for decades after delivery. Upon maternal injury, FMCs migrate to affected sites where they participate in tissue healing. However, the specific signals regulating the trafficking of FMCs to injury sites had to be identified. Here we report that, in mice, a subset of FMCs implicated in tissue repair displays CD11b+ CD34+ CD31+ phenotype and highly express C-C chemokine receptor 2 (Ccr2). The Ccr2 ligand chemokine ligand 2 (Ccl2) enhances the recruitment of FMCs to maternal wounds where these cells transdifferentiate into endothelial cells and stimulate angiogenesis through Cxcl1 secretion. Ccl2 administration improves delayed maternal wound healing in pregnant and postpartum mice but never in virgin ones. This role of Ccl2/Ccr2 signalling opens new strategies for tissue repair through natural stem cell therapy, a concept that can be later applied to other types of maternal diseases.

F oetal microchimeric cells (FMCs) enter maternal circulation, engraft into various organs and persist for decades after delivery, probably for the mother's remaining lifespan 1 . Foetal cells nest in maternal bone marrow (BM) and remain well tolerated by the maternal immune system 1 . FMCs have been reported to contain progenitor cells, including lymphoid 2 , hematopoietic [3][4][5] , mesenchymal 6,7 and cardiomyocyte 8 progenitors. Upon various types of maternal injury, these cells can be triggered to migrate to various maternal tissues in which they will adopt the phenotype of the damaged organ 4,9 . Indeed, FMCs are able to differentiate into neurons 10 , hepatocytes 11 endothelial cells 12,13 , thyroid, cervix and gut 9,14 cells in concerned maternal tissues.
Wound healing is an intricate, interactive biological process 15 . There is a well-coordinated interplay between tissue structures and residential cells as well as distant cells from BM that are essential for the healing process 16 . Studies using chimeric mice with tagged BM identified various types of BM-derived endothelial progenitor cells (EPCs) participating in wound angiogenesis 17,18 . The first reported one is a CD34 þ VEGFR2 þ /CD31 þ CD45 À population that can form blood vessels [19][20][21] . Interestingly, others identified a CD34 þ VEGFR2 þ / CD31 þ CD45 þ or CD11b þ population from the BM that secretes VEGF-A and indirectly supports wound angiogenesis 22 . However, more recent studies have finally failed to detect BMderived EPCs in healing adult tissues 22,23 .
We and others have reported the constant presence of fetal cells in maternal wounds during and after delivery 13,24 , including chronic wounds. In maternal wound beds, fetal cells have been shown to differentiate into endothelial cells, able to form full blood vessels, connected to maternal circulation, in such wounds 13 . The fetal cells mobilized in response to maternal wounds did not express the classically reported EPC markers CD34 þ , CD11b À and VEGFR2 þ (ref. 13). We have also detected blood and lymphatic endothelial cells of fetal origin in inflamed skin 12 and in melanomas 25 . Thus fetal progenitors constitute a reservoir of progenitors, well tolerated by the maternal immune system, that are involved in maternal neovascularization during wound healing 1,13 .
Here we investigate the molecular mechanisms that selectively recruit fetal cells into maternal skin wounds in mice. We find that C-C chemokine receptor 2 (Ccr2) was overexpressed on FMCs upon maternal wounding. Ccr2 is a chemokine receptor expressed on BM progenitor cells as well as monocytes 26,27 . Chemokine ligand 2 (Ccl2), its main ligand, is secreted by endothelial cells and macrophages 28,29 . It triggers the recruitment of Ccr2 þ various cell types from the marrow to the secreting site. Here Ccl2, given at physiological doses, was able to induce maternal neoangiogenesis and to improve skin healing by recruiting FMCs. Importantly, such effect of Ccl2 was restricted to pregnant and postpartum mice and was dependent on Ccr2 expression on FMCs. Our results pave the way for natural stem cell therapy for tissue repair in diseased females.

Results
Maternal skin wounding recruits FMCs through Ccr2 signalling. We mated virgin C57BL/6 females with males heterozygous with the enhanced green fluorescence protein (eGFP) gene. Excisional wounds were then performed on dorsal skin at day E15.5 (experimental scheme is described in detail in Supplementary  Fig. 1). One day after skin wounding, fetal cell counts rapidly increased in the BM (Fig. 1a), blood (Fig. 1b) and wound bed (Fig. 1c). The abundance of these cells in BM and blood decreased on day 3 but were still detectable at low level on days 5 and 9, indicating that FMC mobilization in these compartments was an early process. Details of sensitivity and specificity about our cell sorting experiments are given in the Methods section and Supplementary Fig. 2. To identify the signalling pathways that mediate the mobilization of fetal cells upon maternal wounding, we performed PCR array analysis on sorted fetal cells from the BM isolated from pregnant mice with and without skin wounds. Ccr2 was found to be the chemokine receptor displaying the highest upregulation upon wounding ( Fig. 1d and Supplementary Data 1). We performed also RNA sequencing on sorted fetal eGFP þ cells from the peripheral blood obtained from wounded and unwounded pregnant mice. Since fetal cells show a very early peak in the blood (at day 1) after wounding, we first selected genes that displayed high expression (4100 count per million) in wounded mothers. Then we selected the genes that were overexpressed in circulating fetal cells from wounded mice as compared to unwounded mice (41.25-fold change). Using these thresholds, we retrieved a list of 558 genes that were overexpressed in circulating fetal cells from wounded mice. Among these genes, 16 corresponded to surface receptors and the only chemokine receptor was Ccr2 with 376 versus 261 counts per million reads in wounded versus unwounded conditions leading to a fold change of 1.44 (Fig. 1e). Accordingly, we checked the Ccr2 expression of fetal cells in maternal wounds. By immunofluorescence analysis, FMCs expressed Ccr2 in wound tissue (Fig. 1f). Moreover, 90% of the circulating fetal cells expressed Ccr2 1 day after skin injury versus only 1.50% of those from pregnant unwounded mice (Fig. 1g). These results suggest that maternal wounding induces an early expansion of the circulating Ccr2 þ fetal cell population.
Monocytes secrete Ccl2 early in skin wound healing. We next investigated the Ccr2-mediated chemotactic cues in skin wounds. Ccr2 mRNA and protein levels increased strongly in skin wounds 1 day after injury, subsequently decreasing on day 3 (Fig. 1h,k). Ccl2 mRNA levels followed a parallel pattern, increasing on day 1 then decreasing on day 3 (Fig. 1i). Meanwhile, the levels of Ccl8, another Ccr2 ligand, peaked later in the wound, on day 3 (Fig. 1j). Thus Ccl2/Ccr2 signalling corresponds with the observed early peak of postwounding fetal cells.
We investigated the specific populations secreting Ccl2 in skin wounds by measuring Ccl2 levels in various types of cells sorted from the wounds. Marked Ccl2 overexpression was detected in CD45 þ leukocytes from day 1 wounds (Fig. 1m). We carried out double immunofluorescence staining on day 1 wound sections on day 1: F4/80 þ monocytes and CD31 þ endothelial cells expressed Ccl2, whereas GR-1 þ neutrophils did not (Fig. 1l). These data suggest that monocytes and endothelial cells secrete Ccl2 during the initial stage of skin wound healing.
Ccl2 recruits fetal cells to maternal wounds. Given the overexpression of Ccr2 in FMCs upon wounding, the high percentage of FMCs expressing Ccr2 on their surface and the overexpression of Ccl2 in wound tissue during early stages of healing, we hypothesized that Ccl2 mediates the early recruitment of FMCs to wounds. We therefore injected physiological dose of Ccl2 (50 ng) or PBS into the wound bed on days 0 and 2 after wounding in E15.5 pregnant or virgin mice (Fig. 2a). The numbers of eGFP þ cells in blood did not differ between the mice receiving Ccl2 and PBS on day 7 ( Supplementary Fig. 3a). In contrast, there were 2.03-folds more eGFP þ cells in skin wound tissues into which Ccl2 had been injected as in control wounds (Fig. 2b,c). Consistent with our FACS results, by performing immunofluorescence staining, we detected 3.38-folds more as many eGFP þ cells in the sections of mice receiving Ccl2 as in those    receiving PBS (Fig. 2d,e). The enhanced mobilization of eGFP þ cells in wounds upon Ccl2 administration was also found at days 3 and 5 ( Supplementary Fig. 4). Therefore, Ccl2 is able to recruit FMCs to wound sites throughout the cutaneous healing process.
Ccl2 improves wound healing by enhancing neovascularization. We then analysed the effect of local Ccl2 treatment on maternal wound healing. On days 3, 5 and 7 after wounding, the neo-epidermal tongue, which reflects the healing without contraction interference, showed a 43.44, 65.76 and 51.36% increased re-epithelialization, respectively, after Ccl2 injection as compared to controls at correspondent dates (Fig. 2h,i and Supplementary Figs 7a,b and 8a,b). Consistent with these findings, the non-healed area was decreased on days 2, 5 and 7 in wounds treated with Ccl2 ( Fig. 2f,g). Cell proliferation in the epidermis and dermis areas increased on days 3, 5 and 7 in mice treated with Ccl2 (Supplementary Figs 3b,c; 5c,d and 6c,d). Of note, on day 9, there was no more difference in Ccl2-or PBStreated mice in neo-epidermal tongue ( Supplementary Fig. 7a,b), non-healed area (Fig. 2f,g) and epidermal cell proliferation ( Supplementary Fig. 7c,d). Neovascularization, as evaluated by determining total (adult and fetal) CD31 þ vessel density and the expression of the VEGF-A, VEGFR1 and VEGFR2 genes, also increased in the Ccl2-injected wounds at all time points (Fig. 2j-l and Supplementary Figs 3h; 5e,f; 6e,f and 7e,f). In accordance, FACS analysis of total CD31 þ endothelial cells showed a twofold increase upon Ccl2 injections ( Supplementary Fig. 3j). On the contrary, Ccl2 administration had no effect on lymphatic angiogenesis ( Fig. 2j-l and Supplementary Figs 3i; 5e,f; 6e,f and 7e,f). Nassar et al. 13 reported the formation of blood vessels by FMCs in maternal wounds. We also identified Von Willebrand Factor (vWF)-positive blood vessels formed entirely or partly by eGFP þ FMCs in Ccl2-treated mice (Fig. 2m). In these mice, there was also a higher number of fetal-derived vessels as well as CD31 þ eGFP þ fetal endothelial cells measured by FACS than in those treated with PBS ( Fig. 2n,o). Of note, wound inflammation, as assessed by determining the number of GR-1 þ neutrophils and F4/80 þ macrophages in granulation tissue, displayed no difference between mice treated with Ccl2 and those receiving PBS, including at early time points (Supplementary Figs 3d-g; 5g-j and 6g-j). We therefore conclude that Ccl2 improves healing by promoting the neovascularization of maternal wounds and the formation of fetal cell-derived vessels.
To further determine whether Ccl2 improves skin healing through the effects of FMCs or directly affect wound closure, we analysed the wound healing upon Ccl2 in virgin mice. In these, Ccl2 and PBS injections had similar effects on all parameters: wound surface area ( Supplementary Fig. 8a Therefore, the presence of FMCs is mandatory to promote maternal angiogenesis and wound healing through Ccl2. The mobilization of FMCs through Ccl2 is mediated by Ccr2. At this stage, it was important to demonstrate whether the fetal cell signalling was dependent on Ccr2. To answer this question, we analysed virgin female Ccr2 KO/KO mice, female Ccr2 KO/KO mice mated with eGFP þ males and female Ccr2 KO/KO mice mated with eGFP KI Ccr2 KO males (Fig. 3a,c-e). When Ccr2 KO/KO female mice bear Ccr2 KO/KO foetuses, Ccl2 injections do not enhance wound healing at any day (Fig. 3c,d). In contrast, when Ccr2 KO/KO female mice bear Ccr2 WT/KO foetuses, Ccl2 decreases wounded area by 49.08, 28.96 and 57.58% at days 2, 5 and 7, respectively ( Fig. 3e,f). Interestingly, this ratio is similar to the ratio we found with Ccl2 in WT mice. In addition, only Ccr2 KO/KO mice bearing Ccr2 WT/KO foetuses displayed an increase in fetal cell infiltrate in granulation tissue upon Ccl2 local injections (Fig. 3g,h). Finally, the virgin Ccr2 KO/KO mice did not show any change when treated with Ccl2 ( Fig. 3a,b). Therefore, all these data demonstrate that Ccl2 enhances wound healing through Ccr2-dependent fetal cell recruitment to wound bed.
A specific subpopulation of FMCs responds to Ccl2. We further investigated the subpopulations of fetal cells in the peripheral blood during wound healing. In wounded virgin mice, the number of total myeloid progenitor cells (MPCs), defined as CD11b þ CD34 þ CD31 þ cells, persisted at the same level as in unwounded conditions during the first 2 days, then decreased at day 3 ( Fig. 4a,b). In pregnant mice, the number of maternal MPCs, defined as eGFP À CD11b þ CD34 þ CD31 þ cells, displayed the same pattern as the virgin mice ( Fig. 4c,d). By contrast, the number of fetal MPCs, defined as eGFP þ CD11b þ CD34 þ CD31 þ cells, peaked on the day after wounding (Fig. 3e,f). We also studied maternal (eGFP À ) and fetal (eGFP þ ) EPCs, defined as CD11b À CD34 þ CD31 þ cells. Both populations displayed a late increase on day 3 after wounding ( Supplementary Fig. 10a-c). These observations demonstrate that only the fetal MPCs population expands early, 1 day after maternal wounding.
We further investigated the response of this fetal population of MPCs upon Ccl2 administration. Following the injection of Ccl2, the number of fetal MPCs decreased in the peripheral blood (Fig. 4g,h) but increased in the wound tissue ( Fig. 4i,j). By contrast, the number of maternal MPCs in both blood and wound tissue was unaffected by Ccl2 ( Fig. 4g-j). Meanwhile, Ccl2 did not modify the level of maternal and fetal EPCs in the blood and wound groups ( Supplementary Fig. 10d,e). These results suggest that Ccl2 specifically recruits fetal MPCs from the blood to the wound during the healing process.
Foetal MPCs play dual role in wound angiogenesis. We investigated the fate of fetal MPCs in vivo, by purifying this population from the peripheral blood of pregnant mice with eGFP þ foetuses on day 1 after wounding (Fig. 5a). As a control, we purified eGFP þ adult MPCs from the peripheral blood of   (g,h) PBMCs and (i,j) wound tissues were analysed to determine the levels of maternal CD11b þ CD34 þ CD31 þ MPCs (eGFP À gate) and fetal CD11b þ CD34 þ CD31 þ MPCs (eGFP þ gate) present after the administration of PBS or Ccl2. (g,i) Representative FACS results for CD34/CD31 staining in the virgin wounded transgenic eGFP þ female mice (Fig. 5a). The cells were injected, on day 1, into the wounds of WT C57BL/6 virgin mice. At day 7, wounds injected with eGFP þ fetal MPCs contained vWF þ endothelial cells and were able to form a full vessel ( Fig. 5b and Supplementary Fig. 11). In addition, some eGFP þ fetal MPCs-derived cells expressed smooth muscle cell actin (a-SMA) (Fig. 5c) but never macrophage markers (F4/80) (Fig. 5d) MPCs into wounds, no endothelial or myofibroblastic eGFP þ cells were detected (Fig. 5b,c). Thus only fetal MPCs have the specific properties required for differentiation into endothelial and mural cells in maternal wounds. We also investigated how fetal MPCs enhance angiogenesis in maternal wounds. We sorted fetal and maternal MPCs from wounded tissues of mice pregnant with eGFP þ foetuses on day 3 after wounding (Fig. 5h). We studied the expression profiles of genes associated with angiogenesis pathways by PCR array analysis (Fig. 5i,k). Five mRNA transcripts were upregulated in the fetal MPCs, including Cxcl1, Sphk and TGF-b2, Cxcl1 being the highest increased one. Immunofluorescence analysis on wound sections into which fetal or adult MPCs were injected showed that only injected fetal MPCs in mice expressed Cxcl1 (Fig. 5j). Foetal MPCs expressed lower levels of Thbs2 and Bai1, two angiogenesis inhibitors (Supplementary Data 2).
Foetal MPCs form proliferative clusters and express Ccr2. The question of the clonal origin of the fetal population has repeatedly been raised and is beyond the scope of this article. We nevertheless examined wounds from virgin adult mice into which either adult or fetal MPCs had been injected. At day 7, wound sections into which we had injected adult MPCs contained isolated cells adult MPCs (Fig. 5e,f), whereas those injected with fetal cells contained clusters of eGFP þ Ki67 cells (Fig. 5e-g). Interestingl,y nearly all fetal but o1% of the adult MPCs expressed Ccr2 (Supplementary Fig. 12a-c). Furthermore, using FACS analysis performed on the blood of pregnant mice carrying eGFP þ foetuses, Ccr2 was expressed by 85.77±1.048% of fetal MPCs and by only 0.313 ± 0.128% of their maternal counterparts ( Supplementary Fig. 12e,f).
Ccl2 rescues delayed wound healing in postpartum mice. When present, wound-healing disorders affect adult females years after delivery. Since fetal cells persist in maternal BM throughout life 1 , we investigated the effects of Ccl2 in a delayed wound-healing model, in the early and late postpartum periods. Two weeks or 6 months after delivery, females that had given birth to eGFP þ pups were treated daily with a topical application of 0.05% clobetasol on the lower back for 12 days. At that stage, the skin was atrophic. Excisional skin wounds were made in the clobetasol-treated areas. Ccl2 or PBS was injected into the wounds on days 0 and 2 after wounding (Fig. 6a). Ccl2 injections improved wound closure kinetic, neoepidermal tongues, epidermal and dermal cell proliferation and blood vessel angiogenesis in these mice (Fig. 6b-k). By contrast, lymphangiogenesis, as measured by LYVE1 þ lymphatic vessel density and VEGF-C and VEGFR3 gene expression, and inflammation, as measured by the number of GR1 þ and F4/ 80 þ cells in granulation tissue, did not change after Ccl2 injections (Fig. 6h-k and Supplementary Fig. 13a-e). Importantly, Ccl2 recruited more fetal cells to the wound bed than PBS, as demonstrated by FACS analysis (Fig. 6l,m). Therefore, Ccl2 recruits FMCs and improves delayed wound healing in the early and late postpartum in mice ( Supplementary Fig. 14a-h). Finally, Ccl2 treatment did not induce any systemic effect, with no impact on blood cells count, glycaemia and gamma-glutamyltransferase dosage (Supplementary Table 1).

Discussion
We and others have shown that fetal progenitor cells are involved in maternal repair, including wound-healing processes, as well as in the liver, thyroid, myocardial infarction and brain excitotoxic lesions in humans and/or mouse models 8,10-13 . Here we purified fetal cells specifically involved in wound healing and determined their transcriptional profiles. These cells appeared to overexpress chemokine receptor Ccr2. We found that local injections of physiological doses of Ccl2, a Ccr2 ligand, improved skin healing only in pregnant and postpartum mice (Fig. 7). These data provide the first demonstration that a specific pathway involving physiological doses of Ccl2 acts exclusively through the mobilization of FMCs. Moreover; we have been able to show that Ccl2's ability to recruit FMCs in maternal wound healing was dependent upon Ccr2 expression on fetal cells. Previous studies have shown that, even at higher doses, Ccl2 has no effect on normal wound healing in virgin mice 30 . Ccl2 injections enhance wound healing in db/db mice, in which Ccl2 secretion is impaired 31 . Ccl2 plays a role in tissue repair, because complete Ccl2 deficiency results in an impairment of myeloid cell recruitment and skin healing 32 . Thus our data show for the first time the selective recruitment of fetal progenitor cells through low doses of Ccl2 into maternal injured tissue. This translates into enhanced maternal healing both during pregnancy and much later after delivery, in conditions of normal or delayed skin healing.
One key question concerns the advantages associated with the recruitment of fetal rather than adult cells. We identified the fetal cells mobilized by Ccl2 and recruited to wounds as CD11b þ CD34 þ CD31 þ cells. These proliferated in clusters and differentiated into vWF þ endothelial cells and a À SMA þ mural cells within the wound. By contrast, adult cells of the same phenotype never formed clusters; they had a low proliferative index and did not differentiate into vascular cells of any type. Our data indicate that such fetal progenitors, even when persisting in an adult, are more pluripotent and proliferative than adult progenitors in accordance with others 33 . Our results are consistent with the findings of several studies reporting an absence of adult EPCs 22,23 , whereas such progenitors are found in the fetal compartment. Furthermore, fetal CD11b þ CD34 þ CD31 þ cells have an mRNA profile different from that of their adult counterparts, as they overproduce various proangiogenic molecules, including Cxcl1. Cxcl1/Cxcr2 signalling plays a  cells are sufficient to improve both normal and delayed maternal wound healing. By contrast, stem cell therapy with adult cells requires the injection of millions of cells into wounds or myocardial infarcts [35][36][37] . Our results are in line with previous findings showing that, even in very low numbers, FMCs can rescue maternal defects 2,3 . This efficiency of small number of active fetal progenitors is consistent with the findings of Tamai et al. 38 , who showed that Hmgb1 administration led to the recruitment of small numbers of BM Pdgfra þ Lin À cells to skin injury sites, thereby improving delayed healing 38,39 . These findings can be extended to humans, as allogeneic BM transplantation rescues dystrophic epidermolysis bullosa in children through the migration of this population of cells to sites of skin injury and its differentiation into appropriate cell types 40 .
In conclusion, this study lays the foundations for a new concept natural fetal stem cell therapy. We found that, due to the higher levels of Ccr2 expression on particular fetal progenitors, these cells could rescue delayed wound repair in parous mice and that this advantage persisted for a prolonged period after delivery. This strategy may provide an interesting new alternative to induced pluripotent stem cell, embryonic stem cell or allogeneic cell therapy. The mobilization of natural fetal stem cells results in the recruitment of cells well tolerated by the maternal immune system 41 . Being semi-allogeneic, these cells should have beneficial effects in mothers with genetic or acquired diseases, such as diabetes or sickle cell disease 42 . In addition, fetal cells have several advantages over adult cells for this purpose, in terms of their pluripotency and proliferation capacity. Senescence levels are lower in fetal stem cells than in their adult counterparts 33 , limiting the effects of aging. Finally, the mobilization of such cells avoids the need for various steps associated with risks of cell selection and amplification in vitro 35,43,44 . This approach could, theoretically, be extended to other organs, such as heart, liver or even the nervous system.
Flow cytometry. The skin on the back of the mice was shaved and wounds were harvested and incubated overnight at 4°C in 0.05% trypsin-EDTA (Invitrogen) for Once these cells have reached the wound bed, they will differentiate into fetal-derived endothelial and mural cells. In addition, fetal cells secrete specific angiogenic signals such as Cxcl1 that increase maternal-derived angiogenesis. Overall, these mechanisms enhance maternal skin healing in pregnancy or postpartum. ARTICLE mechanical separation of the epidermis. Tissues were digested by incubation in collagenase IV for 60 min at 37°C, with vortexing every 10 min, and the resulting suspension was filtered through a cell strainer with 100 mm pores and then a cell strainer with 40 mm pores (BD Pharmingen), to obtain a single-cell suspension. Blood was collected from the heart of the mouse and peripheral mononuclear blood cells were separated from erythrocytes and platelets by the Ficoll 1.088 method (Health Care). The cells were harvested and the resulting suspension was washed with PBS (Life Technologies) and filtered through a cell strainer with 40 mm pores (BD Pharmingen). Bones from the legs were collected and BMs were flushed, after sectioning the two heads of the bone, with 1 ml of PBS with a syringe 1 ml and a needle of 25 G. The resulting suspension was washed with PBS and filtered through a cell strainer with 40 mm pores (BD Pharmingen). The antibodies used for cytometry were CD34-eFluor660 (1:100; eBioscience), CD11b-PERCP-Cy5.5 (1:100; eBioscience), CD31-PE-Cy7 (1:100; eBioscience) and CCR2 (1:100; Santa Cruz) crossed with an anti-goat-Alexa 555 (1:1,000 Invitrogen). Flow cytometric data were acquired with a BD LSRII (BD Pharmingen) machine and sorting was performed on a MoFlo cell sorter (Beckman Coulter). The cells were then analysed with the FlowJo software (Treestar, San Carlos, CA). In order to assess the specificity and sensitivity of FACS technique, we performed a dilution curve of eGFP þ splenocytes in WT splenocytes ranging from 100% to 0.0001%. FACS analysis always showed the expected amount of eGFP cells, even in very diluted specimens. (Supplementary Fig. 2a,b). The correlation curve between the expected percentage of eGFP cells and the detected percentage on the standard curve was perfectly linear (correlation factor R 2 ¼ 1). Furthermore, we checked the purity of the sorted eGFP þ cells from the dilution curve using quantitative PCR that showed 92-100% purity in all samples except for the most diluted one that had 81% purity. ApoB was used as a reference gene ( Supplementary Fig. 2c).
Microscopic scoring and measurements. We used a Nikon Eclipse 90i fluorescence microscope equipped with a Nikon DS-Fi1C digital camera (Nikon, Tokyo, Japan). For cell scoring, we took photographs of three different fields and counted the labelled cells by fluorescence densitometry, reporting the number of cells as a percentage of the total (assessed as the total number of nuclei). The mean percentage of labelled cells was calculated for each specimen. Measurements were made with the ImageJ software (NIH, Bethesda, MD).
RNA extraction quantitative PCR and RNA sequencing. Total RNA was extracted from cells or tissues with Trizol reagent, in accordance with the manufacturer's (Invitrogen) instructions. For low amount of cells, RNA was extracted with the NucleoSpin RNA XS Kit, in accordance with the manufacturer's (Macherey-Nagel) instructions. It was then reverse-transcribed with the iScript cDNA Synthesis Kit (Bio-Rad). The resulting cDNA was used for PCR with the SYBR-Green Master PCR Mix (Roche). PCR and data collection were performed on a LightCycler 480 (Roche). The levels of gene expression in the samples were normalized against that of the housekeeping gene (b-actin). All the primers were from Qiagen (QuantiTect).
RNA sequencing was performed on sorted fetal cells from the peripheral blood obtained from wounded and unwounded pregnant mice. Parameters used were 2 reads Â 25 million fragments, paired-end 2 Â 100 nt. Each sample was a pool of six mice, leading to 4500 fetal cells per sample.
RNA sequencing analyses were performed at ICM, Hopital La Pitié Salpétrière, with an Illumina Hi-Seq device.
Surgical wounds. Mice were anaesthetized by the inhalation of 4.9% isoflurane delivered at a flow rate of 300 ml min À 1 in ambient air. The backs of the animals were shaved, and punch biopsy devices were used to create four 6 mm surgical wounds or a single 8 mm surgical wound. All tissues above the panniculus carnosus were excised. Wounds were left uncovered until harvesting. Standardized images of the wounds were obtained at various time points, with a Sony Cybershot 10.1megapixel DSC-W180 digital camera (Sony, Tokyo, Japan). Wound tissues were harvested and either snap-frozen in liquid nitrogen or stored at À 80°C.
Corticoid treatment. Four days after delivery, mice were treated by the topical application of 200 ml clobetasol (Dermoval) per day onto shaved dorsal skin over a period of 10 or 12 days depending on age.
Chemokine/cell injections. Following the generation of an 8 mm wound, 100 ml of Ccl2 (Clinisciences, Nanterre, France) was injected into the four cardinal points of the wound bed on days 0 and 2, at a concentration of 0.5 ng ml À 1 . We therefore injected 50 ng of Ccl2 after the surgery and 48 h later 26,27 .
Alternatively, 10,000 cells were injected into the wound after FACS sorting and resuspension in PBS in accordance with the procedure used before, on day 1.
PCR array for chemokines and receptors and angiogenesis. Changes in chemokines and receptors or angiogenesis levels were measured with an RT 2 Profiler PCR Array for mouse chemokines and receptors PAMM-022ZA or mouse angiogenesis PAMM-024ZR (Qiagen, Hilden, Germany). Total RNA was extracted from sorted eGFP þ cells from BM. The expression of 86 cytokine and chemokine genes was analysed with the Lightcycler 1536 system (Roche Diagnostics, Mannheim, Germany), according to the manufacturer's instructions. The data obtained were analysed with the RT 2 Profiler PCR Array Data Analysis Template (Qiagen). Data were normalized against five housekeeping genes (Actb, B2m, Gapdh, Gusb, Hsp90ab1), and relative expression levels were calculated by the 2 À DDCt method.
PCR array analyses were performed at the Hopital La Pitié Salpétrière ICM, with a LightCycler 1536 (Roche) and RT 2 Profiler PCR Arrays.
Statistical analysis. Statistical analysis was performed with the Graphpad Prism software. The results are expressed as means±s.e.m. Pairs of groups were performed with unpaired, two-tailed Student's t-tests. Differences were considered statistically significant if Po0.05.
Data availability. The authors declare that all data supporting the findings of this study are available within the article and its Supplementary Information files or from the corresponding author upon reasonable request.