ABSTRACT
Fetal development relies on adequate iron supply by the placenta. The placental syncytiotrophoblasts (SCTB) express high levels of iron transporters, including ferroportin1 (Fpn1). Whether they are essential in the placenta has not been tested directly, mainly due to the lack of gene manipulation tools in SCTB. Here, we aimed to generate a SCTB-specific Cre mouse and use it to determine the role of placental Fpn1. Using CRISPR/Cas9 technology, we created a syncytin b (Synb) Cre line (SynbCre) targeting the fetal-facing SCTB layer in mouse placental labyrinth. SynbCre deleted Fpn1 in late gestation mouse placentas reliably with high efficiency. Embryos without placental Fpn1 were pale and runted, and died before birth. Fpn1 null placentas had reduced transferrin receptor expression, increased oxidative stress and detoxification responses, and accumulated ferritin in the SCTB instead of the fetal endothelium. In summary, we demonstrate that SynbCre is an effective and specific tool to investigate placental gene function in vivo. The loss of Fpn1 in late gestation mouse placenta is embryonically lethal, providing direct evidence for an essential role of Fpn1 in placental iron transport.
INTRODUCTION
The placenta is increasingly recognized as an essential contributor to maternal and fetal health during pregnancy, as well as setting the foundation for lifelong well-being. Despite its crucial role in fetal development, the placenta is among the least well-understood organ systems. The importance and paucity of placental research is cited as the driving force for the 2014 launch of the ‘Human Placenta Project’ at the National Institutes of Health, aiming to better understand the role of the placenta in health and disease (Guttmacher et al., 2014; Sadovsky et al., 2014).
Owing to its evolutionary proximity and structural similarity to the human placenta, the mouse placenta has been used to study various aspects of placental development, physiology and pathology (Roberts et al., 2016; Hemberger et al., 2020). The mature mouse placenta consists of three anatomical layers – maternal decidua, junctional zone and labyrinth – each with distinct function and cell composition (Watson and Cross, 2005). The study of global knockout (KO) mice helped identify many essential genes, but it is unclear whether the embryonic phenotype (often lethality) is caused by loss of gene function in the embryo, the placenta, or both.
Many tools have been developed to target the placenta for gene modification, with the Cre-LoxP-based approach being the most cell lineage-specific and more commonly used than cell transplantation and viral infection-based strategies (Renaud et al., 2011). There are three commonly used placental Cre transgenic lines, driven by different trophoblast-specific genes (Li et al., 2014). TpbpaCre targets a subset of ectoplacental cone-derived trophoblasts including junctional zone spongiotrophoblasts, glycogen trophoblasts and trophoblast giant cells (Calzonetti et al., 1995; Simmons et al., 2007). Cyp19Cre is pan-trophoblast and is expressed in both labyrinth and spongiotrophoblast layers (Wenzel and Leone, 2007). Gcm1Cre is the only placental Cre that targets the labyrinth syncytiotrophoblasts (SCTB), which function as the metabolic hub at the maternal-fetal interface, in charge of nutrient transport, waste removal and hormone production. In the Gcm1Cre mouse, Cre expression is driven by Gcm1, an SCTB layer II (SCTBII) transcription factor expressed in early-to-mid-gestation placentas (Nadeau et al., 2009; Nadeau and Charron, 2014). However, few publications have used Gcm1Cre to study SCTB genes since its generation in 2009, possibly due to its inconsistent and incomplete recombination efficiency (C.C. and M.D.F., unpublished).
To address this gap, we used CRISPR/CAS9 technology to generate an SCTB-specific Cre, driven by syncytin b (Synb). Synb is a retro-viral envelope gene in SCTBII and is essential for SCTB formation (Dupressoir et al., 2011). Importantly, Synb is exclusively expressed in mature placentas after embryonic day (E) 11.5 (Dupressoir et al., 2011), making the Cre particularly useful to study gene functions in late gestation placentas.
To demonstrate its effectiveness, we used SynbCre to study the physiological role of placental iron exporter ferroportin 1 (Fpn1; also known as Slc40a1). Fpn1 is the only known mammalian non-heme iron exporter and is crucial for iron efflux in high iron trafficking cells including duodenal enterocytes, iron-recycling macrophages and hepatocytes (Nemeth and Ganz, 2021). Fpn1 is an ideal gene to interrogate with SynbCre. First, Fpn1 is among the most highly expressed genes involved in iron transport and metabolism in the mouse placenta (Cao et al., 2021) and its protein is expressed in SCTBII, which also exclusively expresses Synb (Cao and Fleming, 2021). Second, the current knowledge of placental Fpn1 function is largely deduced from studies in whole-body Fpn1 KO (Donovan et al., 2005) and hypomorphs (Mok et al., 2004) that have significant limitations. For example, it is unclear whether embryonic anemia in Fpn1 hypomorphic mutants is due to the absence of Fpn1 in the embryo, the placenta, or both. Furthermore, Fpn1 KO embryos die before maturation of the placenta, thus precluding any conclusions regarding placental Fpn1. Deletion of Fpn1 in the SCTBII is needed to unequivocally study its role in placental iron transport.
We hypothesized that loss of placental Fpn1 will cause iron deficiency in the embryo as well as compensatory upregulation of other genes involved in iron transport in the placenta. However, these changes will not meet the high fetal iron demand in late gestation and embryonic death will ensue.
RESULTS AND DISCUSSION
SynbCre is expressed in late gestation placentas and is specific to SCTBII
Of the 98 pups born from CRISPR microinjections, 33 (21 males and 12 females) had correct Cre targeting as indicated by PCR amplicons of expected size at both insertion junctions (Fig. 1A). The Cre transgene was inherited at the expected ratios in both male and female offspring.
Generation and characterization of the SynbCre mouse. (A) Design of SynbCre knock-in allele. P2A-Cre was inserted immediately before the stop codon (TAA) of Synb. Primer locations for 5′ and 3′ junction PCRs are shown (left) along with expected amplicon sizes (right). (B) SynbCre-induced YFP expression in the mouse placenta. Left: SynbCre was bred to the R26R-EYFP reporter allele and resulted in excision of the LoxP flanked STOP sequence and subsequent expression of YFP in Cre-expressing cells. Right: immunofluorescence of YFP (green) and SCTB marker CX26 (red) in SynbCre-positive and -negative placentas at E15.5. fc, fetal circulation; mbs, maternal blood space. Scale bars: 15 μm. (C) SynbCre mRNA expression in mouse placentas from E10.5 to E18.5 (n=3-7/time point). Placentas were collected from 1-2 dams/time point. SynbCre-positive placentas are represented by filled circles and SynbCre-negative controls are represented by unfilled circles. Box plots show median values (middle bars) and first to third interquartile ranges (boxes); whiskers indicate 1.5× the interquartile ranges. (D) Correlation between Synb and Cre mRNA expression in SynbCre-positive placentas (n=17). Pearson's correlation=0.85, P<0.0001. (E) Cre mRNA expression in SynbCre-carrying tissues including placentas (n=17), fetal liver (n=5), yolk sac (n=3) and adult tissues (n=2). Black circles represent fetal tissues and red circles represent adult tissues.
Generation and characterization of the SynbCre mouse. (A) Design of SynbCre knock-in allele. P2A-Cre was inserted immediately before the stop codon (TAA) of Synb. Primer locations for 5′ and 3′ junction PCRs are shown (left) along with expected amplicon sizes (right). (B) SynbCre-induced YFP expression in the mouse placenta. Left: SynbCre was bred to the R26R-EYFP reporter allele and resulted in excision of the LoxP flanked STOP sequence and subsequent expression of YFP in Cre-expressing cells. Right: immunofluorescence of YFP (green) and SCTB marker CX26 (red) in SynbCre-positive and -negative placentas at E15.5. fc, fetal circulation; mbs, maternal blood space. Scale bars: 15 μm. (C) SynbCre mRNA expression in mouse placentas from E10.5 to E18.5 (n=3-7/time point). Placentas were collected from 1-2 dams/time point. SynbCre-positive placentas are represented by filled circles and SynbCre-negative controls are represented by unfilled circles. Box plots show median values (middle bars) and first to third interquartile ranges (boxes); whiskers indicate 1.5× the interquartile ranges. (D) Correlation between Synb and Cre mRNA expression in SynbCre-positive placentas (n=17). Pearson's correlation=0.85, P<0.0001. (E) Cre mRNA expression in SynbCre-carrying tissues including placentas (n=17), fetal liver (n=5), yolk sac (n=3) and adult tissues (n=2). Black circles represent fetal tissues and red circles represent adult tissues.
To localize SynbCre expression in the placenta, SynbCre was crossed to a reporter line containing a yellow fluorescent protein (YFP) gene downstream of a floxed STOP sequence (Fig. 1B). Cre-induced removal of the STOP sequence and subsequent YFP expression was detected in the SCTBII of the Cre-positive placenta. Cre-negative placentas did not express YFP. These data confirm the specificity of SynbCre expression and recombination efficiency in the SCTBII.
Next, we determined the time course of SynbCre after mid-gestation (Fig. 1C). Transcript abundance of the Cre transgene rose sharply at E12.5 and remained elevated throughout pregnancy. In addition, Cre was highly correlated with Synb mRNA expression (Fig. 1D), consistent with our design of using Synb to drive Cre expression. Cre was not detected in any adult or fetal tissues analyzed (Fig. 1E), a pattern that mirrors the placental exclusivity of Synb (Dupressoir et al., 2011).
SynbCre is effective at deleting Fpn1 in the mouse placenta
To demonstrate the utility of the SynbCre line, we crossed the line to mice carrying a conditional allele of Fpn1. SynbCre-mediated deletion of the Fpn1 floxed allele was confirmed by the appearance of the Fpn1 KO allele in Fpn1fl/+ placentas carrying SynbCre (Fig. 2A). Placental Fpn1 KO (pKO) was generated by crossing homozygous Fpn1 floxed females with males carrying SynbCre harboring a heterozygous Fpn1 KO allele. No viable pKO pups were observed at birth (n=29 pups from four litters, Chi-square test P=0.02). A pale and small phenotype was consistently observed in pKO embryos after E15.5 (Fig. 2B). The crown-rump length of the pKO embryos at E15.5 was on average 11.3% shorter than the control embryos (12.6±0.7 mm versus 14.2±1.0 mm, P=0.0006).
Effects of placental Fpn1 deletion by SynbCre. (A) Genotyping PCRs of SynbCre-positive and -negative placentas carrying the same inherited parental Fpn1 alleles (one floxed and one WT). (B) Representative image of placental Fpn1 KO and the control embryos at E15.5. (C) Hematoxylin and Eosin staining of control and pKO placentas. (D) Relative expression of Fpn1 mRNA in control (n=7) and pKO (n=15) placentas. **P<0.001 (unpaired two-tailed Student's t-test). (E) Top: western blots of iron metabolic proteins in the control (n=5) and pKO placentas (n=6). Bottom: densitometric analysis of protein quantification between control and pKO placentas. Control placentas are shown as filled circles and pKO placentas are represented as unfilled circles. *P<0.01, **P<0.001 (unpaired two-tailed Student's t-test). (F) Immunofluorescence of FPN1 in control and pKO placentas and fetal liver. (G) Immunofluorescence of FPN1 (red) and TFRC (green) in control and pKO placentas. (H) Ferritin distribution in the labyrinth cell layers in control and pKO placentas. Ferritin light chain (FTL1) is co-stained with SCTB marker CX26 (top) and endothelium marker IB4 (bottom). fc, fetal circulation; mbs, maternal blood space. (I) Graphic summary of the effects of Fpn1 deletion in late gestation placenta on the placenta and the embryo. pKO placentas have reduced TFRC and altered ferritin accumulation in the SCTB. pKO embryos are pale and small and die before birth. (J) Volcano plot of differentially expressed gene (DEG) analysis results of control (n=4) and pKO (n=6) placentas. DEGs are shown in red and non-significant genes are in gray. The 20 genes with the lowest FDR-adjusted P-values are labeled with the gene names. (K) Dot plot of the ten most significantly enriched gene ontology (GO) terms in the DEGs. The size of the dot corresponds to the number of genes associated with the GO term. The color of the dot represents the FDR-adjusted P-value of the GO enrichment. Scale bars: 100 μm (C); 50 μm (F); 20 μm (G); 10 µm (H).
Effects of placental Fpn1 deletion by SynbCre. (A) Genotyping PCRs of SynbCre-positive and -negative placentas carrying the same inherited parental Fpn1 alleles (one floxed and one WT). (B) Representative image of placental Fpn1 KO and the control embryos at E15.5. (C) Hematoxylin and Eosin staining of control and pKO placentas. (D) Relative expression of Fpn1 mRNA in control (n=7) and pKO (n=15) placentas. **P<0.001 (unpaired two-tailed Student's t-test). (E) Top: western blots of iron metabolic proteins in the control (n=5) and pKO placentas (n=6). Bottom: densitometric analysis of protein quantification between control and pKO placentas. Control placentas are shown as filled circles and pKO placentas are represented as unfilled circles. *P<0.01, **P<0.001 (unpaired two-tailed Student's t-test). (F) Immunofluorescence of FPN1 in control and pKO placentas and fetal liver. (G) Immunofluorescence of FPN1 (red) and TFRC (green) in control and pKO placentas. (H) Ferritin distribution in the labyrinth cell layers in control and pKO placentas. Ferritin light chain (FTL1) is co-stained with SCTB marker CX26 (top) and endothelium marker IB4 (bottom). fc, fetal circulation; mbs, maternal blood space. (I) Graphic summary of the effects of Fpn1 deletion in late gestation placenta on the placenta and the embryo. pKO placentas have reduced TFRC and altered ferritin accumulation in the SCTB. pKO embryos are pale and small and die before birth. (J) Volcano plot of differentially expressed gene (DEG) analysis results of control (n=4) and pKO (n=6) placentas. DEGs are shown in red and non-significant genes are in gray. The 20 genes with the lowest FDR-adjusted P-values are labeled with the gene names. (K) Dot plot of the ten most significantly enriched gene ontology (GO) terms in the DEGs. The size of the dot corresponds to the number of genes associated with the GO term. The color of the dot represents the FDR-adjusted P-value of the GO enrichment. Scale bars: 100 μm (C); 50 μm (F); 20 μm (G); 10 µm (H).
Compared with the controls, pKO embryos had elevated liver transferrin receptor (Tfrc) mRNA (P=0.01) and non-significantly lower Hamp mRNA (P=0.13) (Fig. S1). Liver iron staining in the pKO embryo was also less intense than in the control (Fig. S1). This mild iron deficiency phenotype echoes the observation in late gestation Fpn1 hypomorph embryos (Mok et al., 2004). Placental size as measured by placental disk diameter did not differ between pKO and control placentas (7.6±0.3 mm versus 7.5±0.4 mm, P=0.6). In addition, like the Fpn1 hypomorph placentas (Mok et al., 2004), there were no obvious structural abnormalities in the pKO placental labyrinth (Fig. 2C). This may be consistent with the sole function of Fpn1 being an iron transporter with no direct involvement in cellular pathways that have multifaceted and/or systemic effects.
We verified Fpn1 deletion in the pKO placenta at the mRNA and protein levels. Despite relatively high variation in Cre expression among placentas, the degree of Fpn1 deletion was comparable in all pKO placentas compared with the Cre-negative littermate placentas. This indicates a threshold effect of SynbCre-mediated recombination of the Fpn1 floxed allele and the efficiency of Cre-mediated recombination needs to be empirically validated when used with other floxed genes. Using quantitative PCR (qPCR) primers that target one of the two deleted Fpn1 exons, we showed that the Fpn1 transcript level was 50% lower in pKO placentas than in the controls (Fig. 2D). The residual Fpn1 transcript expression in the pKO placenta is likely derived from macrophages and other types of trophoblasts not targeted by the Cre. Exon usage analysis of the RNA-sequencing data showed that pKO placentas had significantly fewer reads from the two deleted Fpn1 exons (Fig. S2). Remarkably, there was a 90% reduction of FPN1 protein in pKO placentas compared with the controls (Fig. 2E). Immunofluorescence demonstrated a near-complete absence of FPN1 in the SCTBII of pKO, which displayed normal FPN1 in the fetal liver (Fig. 2F). This confirms that the anemic phenotype of pKO embryos is a direct consequence of Fpn1 deletion in the placenta, not in the embryo. Using the epiblast-specific Meox2Cre, Donovan et al. (2005) showed that embryos with Fpn1 deletion were viable and attributed their survival to the preservation of Fpn1 expression in the placenta. Here, the lethality of embryos lacking placental Fpn1 directly demonstrates an essential role of placental Fpn1 in supporting fetal growth in late gestation.
In addition to FPN1, Tfrc, the key cellular iron importer, was reduced by 70% in pKO placentas (Fig. 2E). Immunofluorescence localized the TFRC downregulation to the SCTB layer I (SCTBI) (Fig. 2G). Thus, the mild iron deficiency in the pKO embryos was likely a consequence of the combined deficiencies of the two primary placental iron transporters, TFRC and FPN1.
pKO placenta accumulates ferritin in SCTB
Because Fpn1 is an iron exporter, we hypothesized that loss of Fpn1 would block iron export and cause iron accumulation inside the SCTB. To examine iron distribution in the placental layers, we performed diaminobenzidine tetrahydrochloride (DAB)-enhanced Perls Iron Stain in pKO and control placentas. There was no stainable iron in the SCTB of either genotype (Fig. S1). In our experience, it is common for the labyrinth to have little stainable iron, even with DAB enhancement to increase signal sensitivity. The relatively low storage iron in the labyrinth may reflect its transit role in iron transfer and the potential toxicity of high intracellular iron. Consistent with the iron staining results, non-heme iron concentrations did not differ between pKO and control placentas (18.8±13.3 versus 16.1±15.1 μg/g, P=0.6).
Thus, we explored using ferritin immunofluorescence as an alternative to determine iron distribution in the labyrinth layers (Fig. 2H). In the control placenta, most ferritin was localized in the fetal endothelium and SCTBII, with little staining in the SCTBI. In the pKO placenta, ferritin was found in both SCTB layers and was nearly absent from the fetal endothelium. The shift of ferritin storage from fetal endothelium to SCTB in the pKO supports the hypothesis that Fpn1 is needed for iron export from SCTB to the fetal circulation, as illustrated in Fig. 2I.
pKO placenta has increased oxidative stress response and detoxification genes
RNA-sequencing identified 12 downregulated and 36 upregulated differentially expressed genes (DEGs) with greater than 1.5-fold-change in the pKO placentas compared with the controls (Fig. 2J). Tfrc transcript expression was 74% lower in the pKO placentas compared with the controls, consistent with marked reductions of TFRC protein in the pKO placentas shown by western blot (Fig. 2E) and immunofluorescence (Fig. 2G). Because Tfrc mediates the primary route of placental iron uptake, its downregulation may be a very effective and compensatory response to prevent iron accumulation and its associated toxicity in the pKO placenta. We have previously shown that Tfrc expression in mouse placentas is a reflection of both a programmed/temporal control by gestational age as well as an acute/localized regulation by intracellular iron levels (Cao et al., 2021). We and others have shown that placental Tfrc regulation by iron alone is of low magnitude (Cao et al., 2021; Sangkhae et al., 2021). Thus, the high degree of Tfrc downregulation in the pKO placenta is likely an accumulated effect of early and sustained suppression by subtle changes in intracellular iron caused by Fpn1 deletion.
None of the other putative placental iron transporter genes, including Slc11a2 (Gunshin et al., 2005) and Slc39a8 (Gálvez-Peralta et al., 2012), were significantly different in the pKO placentas. Expression of erythroid differentiation genes – including ferrochelatase (Fech), 2,3-bisphosphoglycerate mutase (Bpgm) and tripartite motif-containing 10 (Trim10) – were also lower in pKO placentas. This may indicate lower erythropoietic activity of the nucleated fetal red blood cells in pKO placentas.
On the other hand, genes involved in oxidative stress response and detoxification such as NAD(P)H dehydrogenase, quinone 1 (Nqo1), peroxiredoxin 6 (Prdx6), catalase (Cat) and glutathione S-transferase, alpha 3 (Gsta3), were upregulated in the pKO placenta (Fig. 2K). Complete lists of DEGs and gene ontology (GO) enrichment pathways are provided in Tables S2 and S3. qPCR analysis of selected iron transport and oxidative stress genes in the pKO and control placentas is shown in Fig. S3.
Activation of the oxidative stress response and hydrogen peroxide catabolic pathway in the pKO placentas may be a direct result of increased free intracellular iron in the SCTB, as paralleled by its ferritin accumulation (Fig. 2H). However, the increase in free iron was likely mild because there was no significant iron accumulation in pKO placentas, as indicated by iron stain and ferritin western blot. In addition, elevations of oxidative stress processes were moderate and limited to a few genes in the pathways. Thus, gene expression in pKO placentas likely captured the initial and sensitive transcriptomic response to mild increases in intracellular free iron and its associated oxidative stress.
A major challenge in placental research has been the inability to manipulate genes specifically in the placenta without affecting other maternal and fetal organs. Currently, there are few reliable tools to target the multi-functional placental SCTB for gene manipulation. In this study, we demonstrate the efficiency and specificity of the SynbCre transgenic line at deleting Fpn1 from the SCTB in late gestation mouse placentas. Our data provide direct evidence of an essential role of placental Fpn1 for fetal survival. Like Fpn1, many placental genes, including Tfrc, are considered essential due to the embryonic lethality in the whole-body KO models. Placental specific deletions of these genes using tools like SynbCre are needed to directly address their functional significance in supporting fetal growth. The SynbCre is uniquely valuable in studying SCTB gene functions in vivo and will facilitate the generation of mouse models of pregnancy complications and fetal abnormalities.
MATERIALS AND METHODS
Animals
Wild-type (WT) 129S6/SvEvTac mice were obtained from Taconic Biosciences. Fpn1-floxed mice (129S-Slc40a1tm2Nca/J) (Donovan et al., 2005) were purchased from The Jackson Laboratory. A germline Fpn1 KO allele was generated by crossing the Fpn1 floxed mice with the GATA1 Cre line [129S.Cg-Tg(Gata1-cre)1Sho/MdfJ] (Mao et al., 1999). Both the Fpn1 floxed and KO alleles had been backcrossed to 129S6/SvEvTac for more than 10 generations. Enhanced yellow fluorescent protein (YFP) reporter mice [B6.129X1-Gt(ROSA)26Sortm1(EYFP)Cos/J] (Srinivas et al., 2001) were a gift from Dr Yuko Fujiwara (Boston Children's Hospital, MA, USA) and were originally purchased from the Jackson Laboratory. Unless otherwise stated, genotyping was performed using specific probes by the automated genotyping service at Transnetyx. Mice had ad libitum access to water and standard chow. All animal protocols were approved by the Institutional Animal Care and Use Committee at Boston Children's Hospital.
Generation of SynbCre transgenic mice
SynbCre mice were generated using a protocol modified from the Easi-CRISPR method (Quadros et al., 2017). Briefly, a single guide RNA (sgRNA) targeting the mouse Synb gene (ENSMUSG00000047977) was selected based on high on-target and off-target scores using the Benchling CRISPR design tool (https://www.benchling.com/crispr/). The sgRNA oligo (GACACCATTCCTACATAACA) was synthesized by Synthego. The presence of the sgRNA site in the Synb gene was verified in the FVB/NJ genome by PCR and Sanger sequencing (Genewiz LLC). A megamer single-stranded DNA (ssDNA) donor containing the sequence of the self-cleaving peptide porcine teschovirus-1 2A (P2A) and Cre inserted immediately before the stop codon of Synb was synthesized as the repair template by Integrated DNA Technologies. The donor ssDNA was injected into FVB/NJ zygotes along with the sgRNA and CAS9 protein (CP01, PNA Bio). Tail DNA was extracted from live offspring and screened for correct Cre insertions at the 5′ and 3′ junctions by PCR. SynbCre founders were bred to 129S6/SvEvTac WT mice to pass down the Cre transgene, and the SynbCre animals used in this study were backcrossed for at least three generations. Primers used in genomic PCR assays are provided in Table S1.
Time course and cell specificity of SynbCre in the placenta
To determine the time course of SynbCre expression, timed matings were set up between 129S6/SvEvTac males carrying SynbCre and WT females. Placentas were collected from E10.5 (the formation of the placenta) to E18.5 (end of gestation) (n=3-7 placentas from 1-2 dams/time point). RNA extraction and qPCR analysis were performed as described previously (Cao et al., 2021). Various organs were harvested from two SynbCre-positive adult mice (6-10 weeks of age) to assess Cre expression in adult tissues. Transcript abundance of Cre was calculated by the 2(−ΔΔCT) method, using beta-actin (Actb) as the housekeeping control and SynbCre-positive placentas as the reference group. RNA samples in the same experiments were run in one batch on the same plate. Primers used in qPCR are included in Table S1.
Cell specificity and recombination activity of SynbCre were determined by crossing SynbCre to the YFP reporter line. Placentas were collected at E15.5, fixed in 4% paraformaldehyde, cryosectioned and immunostained for YFP (ab13970, 1:500; Abcam) alongside the SCTB marker connexin 26 (CX26; also known as Gjb2) (71-0500, 1:200; Invitrogen) using methods previously reported (Cao and Fleming, 2021).
Generation of placental Fpn1 KO (pKO)
SynbCre was bred to Fpn1 heterozygous animals to generate Cre-positive Fpn1 heterozygous males (Fpn1−/+; SynbCre). Conditional deletion of Fpn1 in the placenta was achieved by crossing Fpn1−/+; SynbCre males with homozygous Fpn1 floxed females (Fpn1fl/fl). By introducing the Fpn1 KO allele in the male parent, this pairing strategy would reduce the risk of mosaicism due to ineffective Cre-mediated excision when there are two floxed alleles. This design also circumvented the use of parents carrying both Cre and Fpn1 floxed alleles, thus avoiding the impact of placental Fpn1 deletion during embryogenesis of the parental generation.
During dissection, maternal decidua of the placenta was removed, and the fetal portion enriched for the labyrinth was divided into quadrants, processed for paraformaldehyde fixation or stored at −80°C for RNA and protein extractions as previously described (Cao et al., 2021). In most cases, additional placental tissue was weighed and processed for non-heme iron determination. Sizes of the embryo and the placenta were measured as crown rump length (CRL) and the diameter of the placental disk. Embryos and placentas were separately genotyped for Fpn1 floxed, KO, WT, Cre and the sex chromosome Y. Phenotyping comparisons were carried out between the pKO (Fpn1fl/−; SynbCre) and the Cre-negative control (Fpn1fl/−). During the initial phase of pKO phenotyping, we performed cryosectioning and FPN1 immunofluorescence in placental tissues immediately (1-2 days) after dissection to confirm FPN1 protein deletion in the pKO. The consistency of placental FPN1 deletion and pale appearance in the pKO embryos provided us reassurance to remove this confirmatory step and save all the fixed placentas for paraffin embedding for better structure preservation and batched analysis of placentas from different litters.
A total of 15 pKO and seven control placentas were collected from five different litters at E15.5. All 22 placentas were used for RNA extraction and analyzed for Fpn1 transcript abundance by qPCR as an initial confirmation for Fpn1 deletion. Sequences of primers used in qPCR are provided in Table S1. In three of the E15.5 litters, fetal liver (n=5 pKO and n=5 control) was collected for RNA extraction. Whole embryos (n=2 pKO and n=2 control) were fixed and processed for histology.
Placental tissue iron
Iron was extracted from placental tissues (n=13 pKO and n=5 control) in a trichloroacetic and hydrochloric acid mix at 65°C for 2 days. Non-heme iron concentrations in the extracts were determined by the bathophenanthroline quantification method as described previously (Torrance and Bothwell, 1980; Cao et al., 2021).
Histology and immunofluorescence
Fixed placentas were embedded in paraffin and sectioned at 10 µm using a Leica RM2255 microtome. General cellular morphology was assessed by Hematoxylin and Eosin staining. Cellular iron was visualized by Perls Prussian Blue staining followed by signal enhancement with 3,3′ diaminobenzidine tetrahydrochloride (DAB). Bright-field images were acquired using an Olympus BX50 microscope.
Immunofluorescence was performed using antibodies against FPN1 (MTP11-A, 1:200; Alpha Diagnostic International), TFRC (13-6800, 1:200; Invitrogen), FTL1 (HPA041602, 1:100; Sigma-Aldrich), CX26 (71-0500, 1:200; Invitrogen) and isolectin B4 (I32450, 1:50; Thermo Fisher Scientific), as described previously (Cao et al., 2021). At least two placentas of each genotype were stained in the same batch and imaged with a Zeiss Observer D1 fluorescent microscope using the same settings.
Western blotting
Protein extraction and western blotting were carried out as described previously (Cao et al., 2021). A random sample of 11 placentas including six pKO and five controls were selected so they could be run on the same 12-well TruPAGE gel (Sigma-Aldrich) and analyzed in the same experiment. Primary antibodies used include: FPN1 (MTP11-A, 1:1000; Alpha Diagnostic International), TFRC (13-6800, 1:1000; Invitrogen), FTL1 (HPA041602; 1:500 Sigma-Aldrich), FTH1 (4393S, 1:2000; Cell Signaling Technology) and ACTB (4970S, 1:2000; Cell Signaling Technology). Band intensities were quantified using Image Lab software (Bio-Rad Laboratories) and normalized to the intensity of ACTB in the same sample.
Placental RNA-sequencing
RNA-sequencing was performed in six pKO and four control male placentas. These placentas were from four litters at E15.5, each contributing one control and one or two pKO placentas. Only male placentas were used to avoid differential sequencing coverage of the sex chromosomes and the resulting distortion of non-sex genes in transcriptomic analysis. We chose the males because there were more matched male pKO and control placentas than female pairs at the time of RNA-sequencing. Analysis of all 22 male and female pKO and control placentas did not show significant gender differences in the degree of Fpn1 deletion, placental genes, placental iron concentrations or embryo phenotype. Male and female mouse placentas may have different transcriptomic responses to environmental stimuli such as maternal iron deficiency as we demonstrated in a previous study (Cao et al., 2021). However, we did not find gender differences in how the placental iron metabolic genes/proteins responded to maternal iron deficiency and there were no transcriptomic differences (other than the sex specific genes) between male and female placentas at baseline. Nonetheless, it is essential to study and confirm key outcomes including placental and embryonic phenotype in both genders when using Cre deleters such as SynbCre. RNA integrity was analyzed on an Agilent TapeStation 4200. The RNA integrity numbers (RIN) of the placental samples were greater than 8 and averaged 9.4.
Library construction and sequencing were performed by Genewiz, LLC, using NEBNext Ultra RNA Library Prep Kit (New England Biolabs) and Illumina HiSeq 4000 with a 2×150 bp paired-end configuration. Bioinformatic analysis was carried out using a similar pipeline as described previously (Cao et al., 2021). In short, reads were trimmed by Trimmomatic (v.0.36), aligned to the mouse reference genome (GRCm38.p6) by STAR aligner (v.2.5.2b) and counted by featureCounts from the Subread package (v.1.5.2). Counts were used for exon usage analysis using the R package DEXSeq v.1.42.0 (Anders et al., 2012). DEGs were identified using DESeq2 package (v.1.32.0) (Love et al., 2014), with false discovery rate (FDR)-adjusted P-values<0.05. Multiple testing correction is a standard procedure in DEG analysis of RNA-sequencing data and the DESeq2 package implements the Benjamini-Hochberg algorithm by default to control the overall FDR in multiple tests of independent P-values (Benjamini and Hochberg, 1995; Love et al., 2014; Koch et al., 2018). GO enrichment analysis was performed using the clusterProfiler package (v.4.0.5) (Yu et al., 2012). Pathways with FDR-adjusted P<0.05 were considered significant.
Statistical analysis
Statistical analyses (except for RNA-sequencing) were performed in JMP 16.0 (SAS Institute). Differences between two samples were compared by two-tailed unpaired Student's t-tests for normally distributed variables and by Wilcoxon rank sum tests for non-normally distributed variables. Pearson correlation was used to examine the relationship between placental Synb and Cre mRNA expression. Unless otherwise stated, data were expressed as mean±standard deviation (s.d.). Significance was defined as P<0.05.
Acknowledgements
We are grateful to Dr Chen Wu, Dr Joel Lawitt and Jennifer Mark at the Transgenic Core of Beth Israel Deaconess Medical Center for their expertise and efficiency in generating the SynbCre mouse. We thank Dr Liang Sun at the Boston Children's Hospital for his advice on RNA-sequencing analysis. We extend our gratitude to the Bioinformatics Core at Harvard T.H. Chan School for providing training courses on RNA-sequencing data processing and analysis. We are indebted to the staff at Animal Resources of Children's Hospital at Boston Children's Hospital for keeping the animal facility running through the COVID-19 pandemic, which allowed this research to come to fruition.
Footnotes
Author contributions
Conceptualization: C.C.; Methodology: C.C.; Formal analysis: C.C.; Investigation: C.C.; Resources: M.D.F.; Data curation: C.C.; Writing - original draft: C.C.; Writing - review & editing: C.C., M.D.F.; Visualization: C.C.; Supervision: M.D.F.; Funding acquisition: C.C., M.D.F.
Funding
Funding for this study came from the Allen Foundation and the Children's Hospital Pathology Foundation, Boston Children's Hospital.
Data availability
The RNA-sequencing data have been deposited in GEO under accession number GSE206109.
Peer review history
The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.201160.reviewer-comments.pdf
References
Competing interests
The authors declare no competing or financial interests.