A Null Mutation in the Inflammation-Associated S100 S100A8 Causes Early Resorption of the Mouse Embryo This information is current as of September 25, 2021. Robert J. Passey, Elizabeth Williams, Agnieszka M. Lichanska, Christine Wells, Shengping Hu, Carolyn L. Geczy, Melissa H. Little and David A. Hume J Immunol 1999; 163:2209-2216; ; http://www.jimmunol.org/content/163/4/2209 Downloaded from

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 25, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. A Null Mutation in the Inflammation-Associated S100A8 Causes Early Resorption of the Mouse Embryo1

Robert J. Passey,* Elizabeth Williams,† Agnieszka M. Lichanska,† Christine Wells,† Shengping Hu,* Carolyn L. Geczy,* Melissa H. Little,† and David A. Hume2†

S100A8 (also known as CP10 or MRP8) was the first member of the S100 family of calcium-binding shown to be chemotactic for myeloid cells. The is expressed together with its dimerization partner during myelopoiesis in the fetal liver and in adult bone marrow as well as in mature granulocytes. In this paper we show that S100A8 mRNA is expressed without S100A9 mRNA between 6.5 and 8.5 days postcoitum within fetal cells infiltrating the deciduum in the vicinity of the ectoplacental cone. Targeted disruption of the S100A8 gene caused rapid and synchronous embryo resorption by day 9.5 of development in of homozygous null embryos. Until this point there was no evidence of developmental delay in S100A8؊/؊ embryos and 100% decidualization was normal. The results of PCR genotyping around 7.5–8.5 days postcoitum suggest that the null embryos are Downloaded from infiltrated with maternal cells before overt signs of resorption. This work is the first evidence for nonredundant function of a member of the S100 gene family and implies a role in prevention of maternal rejection of the implanting embryo. The S100A8 null provides a new model for studying fetal-maternal interactions during implantation. The Journal of Immunology, 1999, 163: 2209–2216.

he S100 proteins are low m.w. calcium-binding proteins expressed exclusively in this location from around 10.5 days of http://www.jimmunol.org/ belonging to the EF hand superfamily (1). One member of embryological development (10.5 days postcoitum (dpc)3). Ex- T this family, S100A8, is a secreted protein and acts as a pression was not observed during yolk sac hematopoiesis and was potent pure chemoattractant for mouse and human neutrophils and also absent from mature macrophages that leave the liver in very macrophages (2, 3). At least two other members of the S100 family large numbers during development (9). This pattern was consistent also possess chemotactic activity (4, 5). Intradermal injection of with in vitro studies in which both S100A8 and S100A9 mRNAs the S100A8 protein into rats or mice elicits a sustained inflamma- were induced transiently in bone marrow cultures stimulated with tory response closely resembling delayed-type hypersensitivity re- the macrophage-specific growth factor, CSF-1 (10). In this study sponses (2, 3). The S100 gene family also includes S100␣, S100␤, we sought to define the possible roles of S100A8 by targeted dis- calbindins, and calcyclins (6) and forms a cluster of linked ruption of the mouse gene. Given the very restricted pattern of by guest on September 25, 2021 in the mouse and human genomes (1). Human S100A8 (also expression of S100A8 during embryogenesis and its biological ac- known as MRP8, calgranulin A, L1 light chain, cystic fibrosis Ag, tivities, we anticipated that a targeted disruption of the gene might and ) is the most closely related member of the human cause a late embryonic lethality if macrophage/granulocyte pro- S100 family to mouse S100A8, although the level of homology is duction by the liver was perturbed, or it might compromise some Ͻ60%, and the human protein lacks chemotactic activity (3). In aspect of inflammation or LPS responsiveness in adult animals if both species, S100A8 forms complexes with S100A9 (MRP14) in fetal expression was redundant. Because of the large size of the the cytoplasm of neutrophils and macrophages in various states of S100 family, it was possible that the function of S100A8 might be differentiation and activation and has been implicated in calcium- substituted by some other family member and no phenotype would dependent regulation of myeloid cell function (7). In other studies be evident. A more extensive examination of sites of expression of in the laboratory we have found that S100A8 mRNA and protein S100A8 during embryogenesis herein revealed that the gene is also is massively induced in the lung in response to i.v. injection of expressed in extra-embryonic tissues immediately following im- bacterial LPS (our unpublished data). plantation, where it might regulate fetal-maternal interactions. In The first isolation of the mouse S100A8 and S100A9 gene re- keeping with this hypothesis, we show that 100% of null embryos ported upon the restricted expression of the two genes in the fetal are resorbed by the mother at exactly the time that S100A8 is liver (8). More recently, we have demonstrated that both genes are expressed.

*School of Pathology, University of New South Wales, Sydney, Australia; and †De- Materials and Methods partments of Biochemistry and Microbiology and Centre for Molecular and Cellular Location of S100A8 mRNA Biology, University of Queensland, Queensland, Australia Whole mount in situ hybridization using digoxygenin-labeled (DIG) ribo- Received for publication March 26, 1999. Accepted for publication June 3, 1999. probes was performed on embryos at the stages indicated as described The costs of publication of this article were defrayed in part by the payment of page previously (11). S100A8 (CP10/MRP8) and S100A9 (MRP14) cDNA plas- charges. This article must therefore be hereby marked advertisement in accordance mids are detailed in Hu et al. (12). In each case sense and antisense mRNA with 18 U.S.C. Section 1734 solely to indicate this fact. probes were produced; any signal detected with the sense probes is noted 1 Grant support was provided by the National Health and Medical Research Council in the figure legends. For the analysis of the implantation sites, individual of Australia. The Centre for Molecular and Cellular Biology is a Special Research decidua were hemisected along the longest axis using a scalpel under a Centre of the Australian Research Council. 2 Address correspondence and reprint requests to Dr. D. A. Hume, Department of Microbiology, University of Queensland, Q4072 Queensland, Australia. E-mail ad- 3 Abbreviations used in this paper: dpc, days postcoitum; CSF-1, macrophage colony- dress: [email protected] stimulating factor; ES cell, embryonic stem cell; EPC, ectoplacental cone.

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 2210 S100A8 IN FETAL IMPLANTATION Downloaded from

FIGURE 1. Targeted disruption of the S100A8 gene in the mouse germline. A, Targeting strategy used to disrupt one allele of the S100A8 gene in ES http://www.jimmunol.org/ cells. G418-resistant transfectants of ES cells, selected secondarily for the absence of the thymidine kinase gene, were screened by Southern blotting for appropriate shift in restriction fragment length as indicated. Four homologous recombinants were identified. Two of these were used to produce chimeric animals. The transmission of the introduced mutation was monitored subsequently by PCR (see Materials and Methods). A Southern blot of DNA from heterozygous animals probed with a 350-bp probe from the 3Ј end of the targeting vector gave the same restriction pattern as wild type, indicating that no rearrangement or deletion of the gene at the 3Ј end occurred in the targeted allele. B, Results of PCR screening of a litter from a mating between two heterozygous animals, showing the presence of the wild-type S100A8 allele in all of the progeny.

dissecting microscope before fixation in 4% paraformaldehyde and pro- tocysts hatched, and trophoblast migration was observed. Genotyping was by guest on September 25, 2021 cessing as described previously for embryos (11). Following hybridization performed on DNA isolated from hatched blastocyst cells. and extensive washing, the DIG was detected using alkaline-phosphatase- conjugated anti-DIG Abs followed by histochemical staining which yields Immunolabeling of disaggregated decidual cells a blue-purple color. The pattern of expression of S100A8 and c-fms ob- served in Fig. 2 was observed in at least four independent litters at each For immunostaining, embryos were removed and decidua were washed in developmental age; the approximate gestational age (dpc) was confirmed PBS, cut into small pieces, and drawn through a 23-guage needle. Decidua by examination of the stage of development of the embryos. Typically, were incubated in Petri dishes in 0.1% collagenase, 0.2% dispase (Boehr- multiple litters of each gestational age were combined, then 8–10 embryos inger Mannheim, Indianapolis, IN), 20% FCS in PBS for 60 min at 37°C. were stained for expression of each marker examined. The timing of ges- The digest was drawn through the 23-guage needle, pelleted, and washed. tational age is based upon the assumption that coitus occurred at midnight The single cell suspension was blocked with 5% rat serum, then incubated on the evening before confirmation of a vaginal plug. with PE-conjugated anti-Mac-1 Ab (Caltag, South San Francisco, CA), washed, and analyzed on a FACSCalibur (Becton Dickinson, Mountain Targeted disruption of the S100A8 gene View, CA) flow cytometer. The S100A8 targeting vector described in Fig. 1 is based upon a positive- Transfer of embryos into lacZ-expressing mothers negative selection strategy in which cells are initially selected for recom- bination using G418 resistance and then nonhomologous insertions are For the investigation of infiltration of embryos by maternal cells, 3.5-dpc ϩ Ϫ selected against because of their retention of the viral thymidine kinase embryos were collected from S100A8 ( / ) matings and transferred into gene using ganciclovir (13). Following homologous recombination at the pseudopregnant 253 strain female mice, which express nuclear lacZ in all S100A8 locus, the coding sequence is replaced by 1.8 kb of neomycin cells under the control of the HMG-CoA reductase promoter (15). The phosphotransferase (neo) gene driven by the phosphoglycerate kinase 1 mice were sacrificed at 6.5–9.5 dpc. Decidua and embryos were bisected as (pgk1) promoter. Transfection of R1 embryonic stem (ES) cells (a gift from for in situ hybridization, fixed, and stained for lacZ expression as described ␮ Dr. A. Nagy; Samuel Lunenfeld Research Centre, Toronto, Canada) by (15). They were then embedded in paraffin wax, sectioned at 8 and electroporation, selection of candidate homologous recombinants, screen- counterstained with Neutral Red. ing by Southern blotting, and generation of chimeras by the morula aggre- gation method were performed as described by Monkley et al. (14). We Results selected two independent targeted ES cell lines (clone 52 and clone 79) for S100A8 but not S100A9 is expressed in the vicinity of the further study. ectoplacental cone (EPC) at 6.5 to 7.5 dpc Genotyping of S100A8 null mice Previous studies indicated S100A8 is expressed for the first time in Identification of the targeted allele in mice was based upon detection of the the mouse embryo together with S100A9 in presumptive myeloid neo gene using the primers described (14), and identification of the wild- cells in developing liver around 11.5 dpc (8, 9). As a prelude to type, S100A8 allele using the primers 5Ј-GCTCCGTCTTCAAGA performance of targeted disruption of the gene, we examined the CATCGT-3Ј (ϩ22 to ϩ41) and 5Ј-GGCTGTCTTTGTGAGATGCC-3Ј (ϩ898 to ϩ880). For genotyping of preimplantation embryos, the uterus expression of S100A8 mRNA at earlier stages of development. was flushed at 3.5 dpc and blastocysts were placed into 200 ␮l microcul- This study revealed a second site of S100A8 expression in ex- tures in DMEM ϩ 10% FBS for up to 7 days. During this time the blas- traembryonic tissue. Following initial attachment and implantation The Journal of Immunology 2211 Downloaded from http://www.jimmunol.org/

FIGURE 2. Expression of S100A8 and c-fms mRNA in sites of implantation. Sites of expression of mRNA were detected by whole-mount in situ hybridization using digoxygenin-labeled RNA probes as described in Materials and Methods. Dark blue/purple formazan product indicates the sites of . For A–C, decidua were removed from the uterus and cut sagitally just lateral to the midline to expose the embryo. (Bar in A, B, D, and E ϭ 300 ␮). A, Expression of c-fms mRNA in 7.5-dpc embryo; note the expression of the gene in cells completely surrounding the embryo (the primary trophoblasts) as well as extensive expression in the ectoplacental cone (epc). Occasional staining of the embryo itself (middle embryo in B) is due to by guest on September 25, 2021 nonspecific trapping and was also observed with the sense probe. B, Expression of S100A8 mRNA in a subset of cells surrounding the embryo at 7.5 dpc. Unlike c-fms, S100A8 mRNA is not expressed in the cells of the EPC, but is restricted to cells apparently infiltrating the deciduum. C, The 7.5-dpc embryos stained for S100A8 mRNA were embedded in paraffin, sectioned, and counterstained with neutral red. Blue stain denoting S100A8 expression flanking the EPC is restricted to large mononuclear trophoblast-like cells. D and E, Localization of c-fms mRNA (D) in the placenta (stripped of decidual tissue) of an 11.5-dpc embryo, viewed from the maternal side, reflect the very high levels of expression in trophoblast giant cells observed previously (8). By contrast, S100A8 mRNA is undetectable except for a reticular pattern confined to the surface of the maternal face (E). Sections indicate this staining is associated with the vasculature (data not shown).

of the mouse embryo, the polar trophectoderm proliferates to form the method is not amenable to double-labeling to confirm identity the EPC and extraembryonic ectoderm, from which the various (Fig. 2B). Sections of in situ-stained embryos demonstrated that differentiated secondary trophoblast cell types arise. In many re- the S100A8-positive cells flanking the EPC were large mononu- spects, EPC-derived trophoblasts resemble granulocytes and mac- clear cells with abundant cytoplasm (Fig. 2C). The S100A8-pos- rophages. They are capable of extensive phagocytosis (16) and like itive cells were indistinguishable from cells expressing the macrophages, express the gene encoding the receptor for CSF-1, CSF-1R (c-fms) in sections (data not shown, see Ref. 8), though the c-fms protooncogene (11). Whole mount in situ hybridization less numerous, but one cannot distinguish unequivocally between on hemisected decidua at 6.5–8.5 dpc provides a three-dimen- maternally derived decidual cells and trophoblasts at this stage. A sional perspective of gene expression in the implantation site. The separate intense S100A8 signal at the very tip of the EPC was application of this method, which is used routinely on mouse em- present in all decidua (though visible to varying extents in indi- bryos, to gene regulation in the implantation site has not been vidual half-decidua depending upon the plane of hemisection). reported previously to our knowledge. The c-fms mRNA was de- Histological examination of this region confirmed the presence of tected within primary trophoblasts infiltrating surrounding tissues large numbers of neutrophils (not shown). Maternally derived neu- from the full perimeter of implanted embryos at 6.5 dpc, through- trophils are present within the uterine lumen (17), and these, to- out the EPC and extending outwards from that pole into the de- gether with activated macrophages, may contribute to the localized ciduum. By contrast, S100A8 mRNA was only detected in very inflammatory response during initiation of blastocyst implantation small subsets of cells that seemed to delineate the external perim- (18). Murine neutrophils (19) and activated macrophages (9) ex- eter of the EPC (not shown). By 7.5 dpc, the c-fms positive cells press high levels of S100A8 mRNA. were more extensively infiltrated into the deciduum surrounding The induction of S100A8 mRNA of the expected size in decidua the embryo, particularly around the EPC (Fig. 2A). S100A8 was was confirmed by Northern blot analysis (data not shown), indi- clearly restricted to a halo of cells surrounding the EPC. The dis- cating that the signal was unlikely to be attributable to cross hy- tribution apparently partly overlaps the distribution of c-fms but bridization with any other S-100 gene. At 8.5 dpc, S100A8 mRNA 2212 S100A8 IN FETAL IMPLANTATION

Table I. Genotype of progeny of matings of heterozygous S100A8 null heterozygous (ϩ/Ϫ) offspring in the ratio 1:2 (Table I), indicating micea that the null mutation was embryonic lethal. Additional timed mat- ings were conducted for each of the lines and the progeny geno- Cell Line 79 Cell Line 52 typed at various times during embryonic development. Combining No. of progeny screened 189 82 the results for all embryos examined from 10.5 to 13.5 dpc, ho- Percent heterozygous 69 70 mozygous normal, heterozygous (ϩ/Ϫ) and resorbing embryos that could not be genotyped (presumably null) occurred in the a Data from clone 52 and clone 79-derived lines have been combined. No differ- ences between the lines were observed. expected ratios of 1:2:1 (Table II). Development of S100A8 null mice is normal until they are expression in the vicinity of the EPC was no longer detectable by resorbed by the mother in situ hybridization or Northern blot analysis (data not shown). At 6.5 and 7.5 dpc, intact embryos could be extracted from all Whereas c-fms mRNA was maintained at very high levels in tro- decidua. PCR genotyping revealed an excess of apparent heterozy- phoblasts and trophoblast giant cells (Fig. 2D) at later stages of gotes over the expected 1:2:1 ratio, which is probably an artifact embryonic development (10.5–11.0 dpc), S100A8 mRNA expres- due to the sensitivity of PCR and the difficulty of completely sion was only detected in cells associated with the vasculature at avoiding maternal tissue which could result in both ϩ/ϩ and Ϫ/Ϫ the maternal face of the placenta (Fig. 2E). In contrast to the sit- decidua being scored as ϩ/Ϫ. Any detectable signal was scored as uation in the liver, where S100A9 was co-expressed with S100A8 positive. The problem could be avoided with quantitative PCR, but in myeloid cells (14), S100A9 mRNA was not detected in the the key point is that the ratio of definite homozygous wild-type to Downloaded from vicinity of the EPC at any stage except in decidual neutrophils null mutants is one, indicating that there is no selective loss of (data not shown). Although most studies in human and mouse have S100A8-deficient embryos. This conclusion is supported indepen- indicated coexpression of the two S100 proteins, at least one other dently below. The definitive S100A8 null embryos were com- precedent for expression of secreted S100A8, in the absence of pletely indistinguishable from their wild-type or heterozygous lit- S100A9, occurs in mature macrophages responding to bacterial termates in terms of size or developmental stage at 7.5 dpc. LPS (9). The cells expressing S100A8 and S100A9 in the fetal

To confirm that pre-implantation development and trophoblast http://www.jimmunol.org/ liver are also not necessarily coincident and appear morphologi- development was normal, blastocysts were isolated at 3.5 dpc, cally distinct (8), although double labeling is required to confirm hatched in vitro, and subsequently genotyped. This experiment de- this proposal. tected the expected 1:2:1 ratio of wild-type, heterozygote, and ho- mozygous mutant embryos (Table II). The hatched blastocysts, Resorption of S100A8 null embryos occurs between 9.5 and 13.5 including the (Ϫ/Ϫ) embryos, were indistinguishable from one dpc another, and large trophoblasts were adherent to the substratum To address the possible functions of S100A8 in myeloid and tro- surrounding the inner cell mass in each case (data not shown). phoblast cells, we created a disruption of the S100A8 gene in the Infiltration of the embryo by maternal cells and of the deciduum mouse genome. The mouse S100A8 (MRP8) genomic DNA se- by guest on September 25, 2021 quence reported by others (20) has been confirmed in our labora- by fetal cells tory. The targeting strategy is outlined in Fig. 1. Because the gene By 8.5 and 9.5 dpc, despite the lack of any obvious gross abnor- is comparatively small, the introduced mutation removed the ma- mality, the ratio of embryos genotyped as null compared with wild jority of the coding sequence and no functional protein product type was significantly less than one. This pattern suggests that was possible. The targeting vector was transfected into embryonic homozygous null embryos contain greater numbers of cells of ma- stem cells, and after positive-negative selection (for G418 resis- ternal genotype that can be detected by PCR, possibly presaging tance and ganciclovir-resistance respectively) the transfectant the overt resorption that becomes visibly obvious only 24 h later clones were screened for correct targeting of the S100A8 gene. (9.5–10.5 dpc; Table II). Transplacental leukocyte infiltration oc- Two homologous recombinants were used to produce chimeras by curs in normal pregnancies (21) but appears to be strictly con- morula aggregation, and male chimeras from both lines transmitted trolled (22) and is presumably below the limits of detection by the null allele with high frequency to their progeny. Heterozygous PCR in wild-type embryos. animals revealed no obvious gross phenotype, and mated nor- To seek evidence that the apparent excess of heterozygotes de- mally. After S100A8ϩ/Ϫ mice were interbred, analysis of the prog- tected by PCR was indeed due to infiltration by maternal cells, we eny derived from both of the two independent targeting events in transplanted 3.5-dpc embryos from heterozygous matings into separate ES cell lines revealed homozygous normal (ϩ/ϩ) and mothers in which a nuclear lacZ transgene is expressed in all cells

Table II. Genotype/phenotype analysis of embryos derived from mating heterozygous S100A8 null mice

Genotype Embryo Age Total Resorbing (dpc) Embryos (ϩ/ϩ)(ϩ/Ϫ)(Ϫ/Ϫ) Embryos

13.5 33 11 22 0 8 12.5 27 2 23 0 6 10.5 15 7 8 0 4 9.5 56 14 42 0 5 8.5 98 18 76 4a 0 6.5–7.5 109 16 75 18 0 Blastocyst 27 5 15 7

a The distribution of genotypes at 8.5 dpc compared to that at 6.5–7.5 dpc using a ␹2 contingency test. The results indicate that the two sets of genotypes are independent; ␹2 ϭ 7.613, 2 degrees of freedom, p ϭ 0.02. Hence the apparent loss of S100A8 null homozygotes at 8.5 dpc is statistically significant. The Journal of Immunology 2213

FIGURE 3. lacZ staining to monitor maternal infil- tration of the embryo and fetal infiltration of the de- ciduum. The 3.5-dpc embryos were transferred to 253 strain lacZ-expressing pseudopregnant mothers and nu- clear lacZ activity was detected as described in Materi- als and Methods. A, Normal implantation site from an 8.5dpc S100A8 (ϩ/?) mouse. Cells of fetal origin (red nuclei) can be seen migrating away from the EPC and infiltrating into the adjacent deciduum (arrows). B, Typ- ical early stage of resorption in an 8.5dpc embryo from the S100A8 (ϩ/Ϫ) heterozygous cross. Extensive infil- tration of the EPC and the embryo by lacZ-positive (blue) cells (arrows) is evident. Downloaded from

(15). This experiment also permits the detection of cells of fetal mates. No gross deficiency in the EPC was evident, and decidua origin infiltrating the deciduum in normal implantation sites. Fig. were the same size and shape as those of wild-type embryos (Fig.

3A shows a normal implantation site at 8.5 dpc, the maternal cells 4A). It is possible to obtain further differentiation of trophoblasts in http://www.jimmunol.org/ expressing nuclear lacZ are excluded from the EPC region, vitro by taking explant cultures of the EPC (23). A series of het- whereas fetal cells with red nuclear staining (i.e., absence of lacZ) erozygous matings was performed, and 7.5-dpc EPC explants were can be detected infiltrating the deciduum, corresponding in loca- cultivated for 4–7 days as described without exogenous growth tion to the S100A8-positive cells detected by in situ hybridization stimulus. All cultures formed adherent trophoblast-like cells, and in Fig. 2. Conversely, Fig. 3B shows an embryo in early stages of there was no clear difference in the extent of growth or spreading resorption, at 8.5 dpc, demonstrating extensive infiltration by lacZ- that would identify a subset of putative S100A8 null EPC cultures. positive maternal cells. Apart from this embryo, where there is To assess the decidual reaction further, we digested 29 individ- macroscopic evidence of early resorption, and others in which re- ual decidua from three separate heterozygous matings with colla- sorption was already complete, in eight litters examined we did not genase/dispase and identified myeloid cells by flow cytometry. A by guest on September 25, 2021 detect lacZ-positive cells infiltrating morphologically normal em- representative profile is shown in Fig. 4B. Two populations of cells bryos using this approach and thus did not provide a clear expla- of different size and light scattering properties expressed the type nation for the PCR genotyping. It may be that maternal leukocytes 3 complement receptor (CD11b, Mac-1 Ag) present on granulo- are contributed from extraembryonic sources during dissection, or cytes and macrophages. There was remarkably little variation in the number detectable by PCR is below the limits of detection by the numbers or proportions of the two cell populations between lacZ staining or the cells infiltrating initially express lacZ at low littermates (Fig. 4B). The data indicate that there is no gross effect levels. of the null mutation on the initial myeloid cell infiltration of the decidua. Identification of S100A8 null embryos by localization of S100A8 mRNA confirms that development is normal until resorption Discussion occurs The S100 family contains some 20 functionally and structurally Apart from direct genotyping, the only alternative approach to related proteins, including at least two other members (calbindin Ϫ Ϫ identify S100A8 / embryos is to examine expression of the and mts1) expressed in trophoblasts (24, 25). S100A8 has previ- S100A8 gene. Whole-mount in situ detection of S100A8 mRNA ously been detected only in myeloid cells. In this paper we have on hemisected decidua from a series of 7.5-dpc embryos from provided the first evidence that S100A8 is expressed specifically in heterozygous crosses (Fig. 4A) showed no S100A8 signal in re- a subset of fetal cells in the vicinity of the EPC and that this gions flanking the EPC in 11/46 decidua examined, whereas the expression is required for normal implantation. Soares et al. (26) signal was detected in every one of the embryos from wild-type and Rickenberger et al. (27) have reviewed natural and introduced matings examined in the course of localization of the gene in Fig. mutations affecting murine trophoblast development and implan- 2. The disappearance of the S100A8 signal in the expected 25% of tation. Introduced mutations with some similarities to the S100A8 embryos from a heterozygous (ϩ/Ϫ) cross further argues that the null include the trophoblast transcription factors, Mash-2 (28), cells expressing the S100A8 gene in the region flanking the EPC Ets-2 (29), ERR-␤ (30), Hand1 (31), Mfa1 (32), and TFEB (33). (Fig. 4) are of fetal origin and confirms the evidence based upon However, each of these mutations differs from the S100A8 null in PCR genotyping that null embryos develop normally to this point. that the target gene is also expressed elsewhere in the embryo proper and/or there is a clear defect in growth and development of The decidual reaction and trophoblast outgrowth occurs the embryo itself or the placenta tissues before the onset of re- normally in S100A8 null embryos sorption. In fact, the nature of the cells expressing S100A8 in the The decidua of S100A8Ϫ/Ϫ embryos identified based upon the vicinity of the EPC is somewhat unclear. Most of the genes noted absence of S100A8 mRNA, and the embryos within them, were above are either expressed in a similar manner to c-fms, which we indistinguishable from the (ϩ/?) embryos and decidua of litter- have used herein as a trophoblast marker, or are not expressed until 2214 S100A8 IN FETAL IMPLANTATION Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 4. Normal development and decidualization of the implantation site of the S100A8 null mouse to 7.5 dpc. A, Comparison of hemisected decidua from S100A8Ϫ/Ϫ (left) or S100A8ϩ/? (right) implantation sites at 7.5 dpc. In both cases, S100A8 mRNA has been localized by whole-mount in situ hybridization. Identification of Ϫ/Ϫ embryos was based on the absence of detectable S100A8 mRNA in the vicinity of EPC. This phenotype was observed in 11/46 embryos. The staining at the very tip of the EPC was present in all embryos examined and was of maternal origin. Embryonic and decidual development in the S100A8Ϫ/Ϫ is indistinguishable from the normal. B, Representative flow cytometry profile for expression of CD11b (Mac-1 Ag) on isolated decidual cells. The two apparent classes of CD11b-positive cells separable on the basis of size (FSC-H) probably represent granulocytes (R1) and macrophages (R2). The range of values for percentages of cells within each of the gates is given at the right of the panel. Note that the range varied between litters, presumably reflecting minor differences in precise gestational age and rapidity of the decidual reaction. No clear outlying values were observed in any of the three litters (total, 29 embryos). later stages of placental development and trophoblast differentia- guished morphologically from trophoblasts. Hence, on the basis of tion. Activated decidual cells are also concentrated in the vicinity cell morphology and limited precedent, S100A8 expression could of the EPC, shown clearly by induced expression of the TIMP-3 met- have been either fetal or maternal in origin. In this paper, we have allo-proteinase inhibitor (34), and these cells are not readily distin- shown that there are fetally derived cells infiltrating the deciduum in The Journal of Immunology 2215 the vicinity of the EPC (Fig. 3), and the data in Fig. 4 confirm that the Acknowledgments S100A8-positive cells are of fetal origin. Indeed, it would be difficult We thank Drs. Brandon Wainwright and Steven Delaney for helpful dis- to see a mechanism for early embryonic loss in the S100A8 null if the cussion, advice and assistance at the outset of this project. cells expressing the gene were maternal. One gene that does resemble S100A8 in expression pattern is urokinase plasminogen activator, which is induced around 7 dpc surrounding the EPC, in a halo of cells References presumed to be migrating trophoblasts (34). 1. Schafer, B. W., R. Wicki, D. Engelkamp, M. G. Mattei, and C. W. Heizmann. Given the lack of any overt defect in either embryo or tropho- 1995. Isolation of a YAC clone covering a cluster of nine S100 genes on human 1q21: rationale for a new nomenclature of the s100 calcium-binding blast development in vivo or in vitro before resorption of the protein family. Genomics 25:638. Ϫ/Ϫ S100A8 embryo we propose that S100A8 is secreted, as occurs 2. Lackmann, M., C. J. Cornish, R. J. Simpson, R. L. Moritz, and C. L. Geczy. 1992. in activated macrophages, and contributes to the regulation of fe- Purification and structural analysis of a murine chemotactic cytokine (CP-10) with to S-100 proteins. J. Biol. Chem. 267:7499. tal-maternal interaction. In support of this hypothesis, recent stud- 3. Lackmann, M., P. Rajasekariah, S. E. Iismaa, G. Jones, C. J. Cornish, S. P. Hu, ies in our laboratory (S. Leung and C. L. Geczy, unpublished data) R. J. Simpson, R. L. Moritz, and C. L. Geczy. 1993. Identification of a chemo- have shown that although abundant S100A8 protein can be de- tactic domain of the pro-inflammatory S100 protein CP-10. J. Immunol. 150: 2981. tected in decidual extracts at 8.5 dpc, S100A8 apparently does not 4. Jinquan, T., H. Vorum, C. G. Larsen, P. Madsen, H. H. Rasmussen, B. Gesser, accumulate within cells in the vicinity of the EPC (as evidenced by M. Etzerodt, B. Honore, J. E. Celis, and K. Thestrup-Pedersen. 1996. Psoriasin: immunocytochemistry in which decidual neutrophils provide a a novel chemotactic protein. J. Invest. Dermatol. 107:5. 5. Komada, T., R. Araki, K. Nakatani, I. Yada, M. Naka, and T. Tanaka. 1996. positive control). As noted in the Introduction, we have shown that Novel specific chemotactic receptor for S100L protein on guinea pig eosinophils.

S100A8 mRNA is massively inducible in mouse organs following Biochem. Biophys. Res. Commun. 220:871. Downloaded from i.v. LPS injection (our unpublished data) indicating that expression 6. Fano, G., S. Biocca, S. Fulle, M. A. Mariggio, S. Belia, and P. Calissano. 1995. The S-100: a protein family in search of a function. Prog. Neurobiol. 46:71. in mice, as in humans, is associated with inflammation. The de- 7. van den Bos, C., J. Roth, H. G. Koch, M. Hartmann, and C. Sorg. 1996. Phos- cidual reaction to embryo implantation is essentially a form of phorylation of MRP14, an S100 protein expressed during monocytic differenti- ϩ acute inflammation (17, 18, 35). The data in Fig. 4 indicate that the ation, modulates Ca2 -dependent translocation from cytoplasm to membranes and cytoskeleton. J. Immunol. 156:1247. primary decidual reaction occurred normally in S100A8 null mice, 8. Lagasse, E., and I. L. Weissman. 1992. Mouse MRP8 and MRP14, two intracel- not surprisingly because much of the process occurs in response to lular calcium-binding proteins associated with the development of the myeloid http://www.jimmunol.org/ blastocyst hatching, well before S100A8 is expressed maximally at lineage. Blood 79:1907. 9. Lichanska, A. M., C. M. Browne, G. W. Henkel, K. M. Murphy, 7.5 dpc. M. C. Ostrowski, S. R. McKercher, R. A. Maki, and D. A. Hume. 1999. Differ- The actual process of resorption in S100A8 null embryos must entiation of the embryonic mononuclear phagocyte system: the role of transcrip- occur very rapidly, because among the hundreds of implantation tion factor PU.1. Blood 194:127. 10. Goebeler, M., J. Roth, U. Henseleit, C. Sunderko¨tter, and C. Sorg. 1993. Expres- sites we have examined (Table II), the one shown in Fig. 3B is the sion and complex assembly of calcium-binding proteins MRP8 and MRP14 dur- only one in which the embryo was not either completely normal, ing differentiation of murine myelomonocytic cells. J. Leukocyte Biol. 53:11. or completely destroyed. A similar pattern of resorption is ob- 11. Hume, D. A., S. J. Monkley, and B. J. Wainwright. 1995. Detection of c-fms ϫ protooncogene in early mouse embryos by whole mount in situ hybridisation served in a natural model of early embryo loss, the CBA/J indicates roles for macrophages in tissue remodelling. Br. J. Haematol. 90:939. DBA/2 cross. In this model, embryonic development is also com- 12. Hu, S. P., C. Harrison, K. Xu, C. J. Cornish, and C. L. Geczy. 1996. Induction of by guest on September 25, 2021 pletely normal until 8 dpc, and embryo loss has been attributed to the chemotactic S100 protein, CP-10, in monocyte/macrophages by lipopolysac- charide. Blood 87:3919. activation of macrophages and/or NK cells, or to ischemia due to 13. Mansour, S. L., K. R. Thomas, and M. R. Capecchi. 1988. Disruption of the vascular thrombosis (36) and inflammation (36, 37). S100A8 could proto-oncogene int-2 in mouse embryo-derived stem cells: a general strategy for contribute to the prevention of such pathology in a normal implan- targeting mutations to non-selectable genes. Nature 336:348. 14. Monkley, S. J., S. J. Delaney, D. J. Pennisi, J. J. Christiansen, and tation site. It is chemotactic at low concentrations and the macro- B. J. Wainwright. 1996. Targeted disruption of the Wnt2 gene results in placen- phages elicited have a particular phenotype including efficiency of tation defects. Development 122:3343. uptake of lipid (38) and a high phagocytic index (our unpublished 15. Tan, S. S., E. Williams, and P. P. L. Tam. 1993. X-chromosome inactivation occurs at different times in different tissues of the post-implantation mouse em- data), properties similar to those described for circulating mono- bryo. Nat. Genet. 3:170. cytes in normal pregnancy (39). Hence the S100A8 secreted in the 16. Albieri, A., and E. Bevilacqua. 1996. Induction of erythrophagocytic activity in vicinity of the EPC could regulate the state of macrophage acti- cultured mouse trophoblast cells by phorbol myristate acetate and all-trans reti- nal. Placenta 17:507. vation and procoagulant activities, thus protecting the embryo 17. Brandon, J. M. 1993. Leukocyte distribution in the uterus during the preimplan- from immune attack by maternal cells. Alternatively, or in addi- tation period of pregnancy and phagocyte recruitment to sites of blastocyst at- tion, high levels of S100A8 may be protective by virtue of its tachment in mice. J. Reprod. Fertil. 98:567. 18. McMaster, M. T., S. K. Dey, and G. K. Andrews. 1993. Association of mono- ability to be readily oxidized by reactive oxygen intermediates cytes and neutrophils with early events of blastocyst implantation in mice. including hypochlorite, the major oxidant produced by activated J. Reprod. Fertil. 99:561. granulocytes (40). By analogy, the anti-oxidant defense provided 19. Iismaa, S. E., S. Hu, M. Kocher, M. Lackmann, C. A. Harrison, S. Thliveris, and C. L. Geczy. 1994. Recombinant and cellular expression of the murine chemo- by thioredoxin functions is an essential protective factor against tactic protein, CP-10. DNA Cell Biol. 13:183. oxidative stress to embryos, though the thioredoxin null mutant 20. Nacken, W., M. P. Manitz, and C. Sorg. 1996. Molecular characterization of the dies earlier than the S100A8 null (41). In the future it will be of genomic locus of the mouse MRP8 gene. Biochim. Biophys. Acta 1315:1. ϫ 21. Piotrowski, P., and B. A. Croy. 1996. Maternal cells are widely distributed in interest to compare the S100A8 null with the CBA/J DBA/2 murine fetuses in utero. Biol. Reprod. 54:1103. cross as a model for early embryo loss, particularly once the null 22. Hunt, J. S., D. Vassmer, T. A. Ferguson, and L. Miller. 1997. Fas ligand is allele has been crossed onto the appropriate inbred genetic back- positioned in mouse uterus and placenta to prevent trafficking of activated leu- ␤ kocytes between the mother and the conceptus. J. Immunol. 158:4122. grounds. PGE2, CSF-1, leukemia inhibitory factor, FasL, IL-1 , 23. Kanai-Azuma, M., Y. Kanai, M. Kurohmaru, S. Sakai, and Y. Hayashi. 1993. TGF-␤, and IL-10, have variously been attributed roles in regu- Insulin-like growth factor (IGF)-1 stimulates proliferation and migration of lating infiltration of the embryo by maternal cells and the latter two mouse ectoplacental cone cells, while IGF-II transforms them into trophoblastic ϫ giant cells in vitro. Biol. Reprod. 48:252. prevent resorption in the CBA/J DBA/2 model (22, 36, 42). 24. Ford, H. L., and S. B. Zain. 1995. Interaction of metastasis associated Mts1 Among these candidate regulators, S100A8, a known regulator of protein with non muscle myosin. Oncogene 10:1597. myeloid cell function and inflammation (2, 3, 38, 40, 43), is unique 25. Shamley, D. R., G. Veale, J. M. Pettifor, and R. Buffenstein. 1996. Trophoblastic giant cells of the mouse placenta contain calbindin-D9k but not the vitamin D in being expressed at precisely the critical time and place where receptor. J. Endocrinol. 150:25. infiltration of the embryo by maternal cells must be regulated. 26. Soares, M. J., B. M. Chapman, T. Kamei, and T. Yamamoto. 1995. Control of 2216 S100A8 IN FETAL IMPLANTATION

trophoblast differentiation: lessons from the genetics of early pregnancy loss and 35. Orlando-Mathur, C. E., and T. G. Kennedy. 1993. An investigation into the role trophoblast neoplasia. Dev. Growth Differ. 37:355. of neutrophils in decidualization and early pregnancy in mice. Biol. Reprod. 27. Rinkenberger, J. L., J. C. Cross, and Z. Werb. 1997. Molecular genetics of im- Fertil. 48:1258. plantation in the mouse. Dev. Genet. 21:6. 36. Clark, D. A., G. Chaouat, P. A. Arck, H. W. Mittruecker, and G. A. Levy. 1998. 28. Guillemot, F. A., A. Nagy, J. Auerbach, J. Rossant, and A. Joyner. 1994. Essen- Cytokine-dependent abortion in CBA ϫ DBA/2 mice is mediated by the proco- tial role of Mash-2 in extraembryonic development. Nature 371:333. agulant fgl2 prothombinase. J. Immunol. 160:545. 29. Yamamoto, H., M. L. Flannery, S. Kupriyanov, J. Pearce, S. R. McKercher, 37. Haddad, E. K., A. J. Duclos, W. S. Lapp, and M. G. Baines. 1997. Early embryo G. W. Henkel, R. A. Maki, Z. Werb, and R. G. Oshima. 1998. Defective tro- loss is associated with the prior expression of macrophage activation markers in phoblast function in mice with a targeted mutation of Ets2. Genes Dev. 12:1315. the decidua. J. Immunol. 158:4886. 30. Luo, J., R. Sladek, J. A. Bader, A. Matthyssen, J. Rossant, and V. Giguere. 1997. 38. Lau, W., J. M. Devery, and C. L. Geczy. 1995. A chemotactic S100 peptide Placental abnormalities in mouse embryos lacking the orphan nuclear receptor ␤ enhances scavenger receptor and Mac-1 expression and cholesteryl ester accu- ERR- . Nature 388:778. mulation in murine peritoneal macrophages in vivo. J. Clin. Invest. 95:1957. 31. Riley, P., L.-A. Cartwright, and J. C. Cross. 1998. The Hand1 bHLH transcription 39. Sacks, G., I. Sargent, and C. Redman. 1999. An innate view of human pregnancy. factor is essential for placentation and cardiac morphogenesis. Nat. Genet. 18: Immunol. Today 20. 271. 32. Kraut, N., L. Snider, C. Chen, S. J. Tapscott, and M. Groudine. 1998. Require- 40. Harrison, C. A. M. J. Raftery, J. Walsh, P. Alewood, S. E. Iismaa, S. Thliveris, ment of the mouse I-mfa gene for placental development and skeletal patterning. and C. L. Geczy. 1999. Oxidation regulates the inflammatory properties of the EMBO J. 17:6276. murine S100 protein S100A8. J. Biol. Chem. 274:8561. 33. Steingrimsson, E., L. Tessarollo, S. W. Reid, N. A. Jenkins, and N. G. Copeland. 41. Matsui, M., M. Oshima, H. Oshima, K. Takaku, T. Maruyama, J. Yodoi, and 1998. The bHLH-zip transcription factor Tfeb is essential for placental vascu- M. M. Taketo. 1996. Early embryonic lethality caused by targeted disruption of larization. Development 125:4607. the mouse thioredoxin gene. Dev. Biol. 178:179. 34. Harvey, M. B., K. J. Leco, M. Y. Arcellana-Panlilio, X. Zhang, D. R. Edwards, 42. Cross, J. C., Z. Werb, and S. J. Fisher. 1994. Implantation and the placenta: key and G. A. Schultz. 1995. Proteinase expression in early mouse embryos is reg- pieces of the development puzzle. Science 266:1508. ulated by leukemia inhibitory factor and epidermal growth factor. Development 43. Geczy, C. 1996. Regulation and proinflammatory properties of the chemotactic 121:1005. protein, CP-10. Biochim. Biophys. Acta 1313:246. Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021