ERp44 is required for endocardial cushion development by regulating VEGFA secretion in myocardium

Abstract Objectives Endocardial cushions are precursors of the valve septum complex that separates the four heart chambers. Several genes have been implicated in the development of endocardial cushions. Specifically, ERp44 has been found to play a role in the early secretory pathway, but its function in heart development has not been well studied. Materials and Methods In this study, we established conditional and tissue‐specific knockout mouse models. The morphology, survival rate, the development of heart and endocardial cushion were under evaluation. The relationship between ERp44 and VEGFA was investigated by transcriptome, qPCR, WB, immunofluorescence and immunohistochemistry. Results ERp44 knockout (KO) mice were smaller in size, and most mice died during early postnatal life. KO hearts exhibited the typical phenotypes of congenital heart diseases, such as abnormal heart shapes and severe septal and valvular defects. Similar phenotypes were found in cTNT‐Cre +/−; ERp44fl / fl mice, which indicated that myocardial ERp44 principally controls endocardial cushion formation. Further studies demonstrated that the deletion of ERp44 significantly decreased the proliferation of cushion cells and impaired the endocardial‐mesenchymal transition (EndMT), which was followed by endocardial cushion dysplasia. Finally, we found that ERp44 was directly bound to VEGFA and controlled its release, further regulating EndMT. Conclusion We demonstrated that ERp44 plays a specific role in heart development. ERp44 contributes to the development of the endocardial cushion by affecting VEGFA‐mediated EndMT.

for CHD. Morphologically, the normal extracellular matrix secreted by the myocardium promises the normal development of EC. 3,4 Many genes have been shown to play essential roles in endocardial cushion development, including molecules involved in transcription, epigenetics, adhesion and migration, such as TGF-β2, Wnt3, VEGF. 5 These cytokines induce the depolarization of the endothelial cells in the endocardial cushion area and weaken the connections between cells, resulting in the migration of mesenchymal cells (MSCs) to the myocardium. The proliferated MSCs together contribute to forming EC. The above process is named the endothelial-mesenchymal transition (EMT). 6 Although the effect of EMT, to some extent, has been clarified on the EC development, additional details remain to be elucidated given the complexity and importance to regenerative or therapeutic purposes.
ERp44 is a pH-regulated chaperone and belongs to the protein disulphide isomerase family. 7 It contains three thioredoxin domains, a, b and b', and a flexible carboxy-terminal tail. 8 The CRFS motif in the domain is thought to form a disulphide bridge with ERp44 client proteins. [9][10][11] This flexible tail masks the substrate-binding site; thus, the protein is in the off-state at the neutral pH of the endoplasmic reticulum (ER) and in the on-state at the lower pH of the cis-Golgi apparatus, allowing it to capture KDEL receptors. 12,13 In this manner, ERp44 orchestrates the balance between client protein retention in the ER and their secretion by covalent interactions. 14,15 Several important secretory factors are reportedly regulated by ERp44, including IgM, 16,17 adiponectin, 18 SUMF1 19 and serotonin. 20,21 Moreover, ERp44 has been implicated in calcium homeostasis. Specifically, ERp44 binds directly to the L3V of inositol 1,4,5-trisphosphate receptor type 1 (IP 3 R1) to inhibit its activity. 22 C160/C212, but not C29, participates in the regulation of IP 3 R1. 23 Huang et al reported that ERp44 may also regulate IP 3 R2. 24 Recently, Wang et al reported that ERp44 mutant leads to embryonic lethality using mouse model, but they did not provide a direct reason for this phenomenon. 25 Moreover, Hisatsune et al reported that ERp44 regulates blood pressure. 26 In the present study, we report that the deletion of ERp44 leads to congenital heart defects. Specifically, a loss of ERp44 causes AVC dysplasia arising from the aberrant proliferation of cushion cells and reduced EndMT by directly regulating VEGFA during heart development. Thus, our findings revealed a critical role for ERp44 in endocardial cushion development.

| Mouse breeding and genotyping
All animal studies were performed according to the relevant guidelines and regulations approved by the Committee on Animal Care of the Institute of Biophysics at the Chinese Academy of Sciences in China. Mice were mated at 5:00 pm, and E0.5 was defined as 9:00 am of the next day when the vaginal plug was detected. Genotyping procedure was performed as previously described. 27 According to the 3R' principles, all animal-relevant experiments should be designed to reduce the number of animals used to meet scientific objectives. 28

| Generation of ERp44 conventional knockout mice
ERp44 knockout mice were generated by gene targeting the ESCs of the 129 mouse strain and subsequently injecting positive cells into C57BL/6 blastocysts. A Loxp-neomycin-Loxp cassette with a homologous arm was used to replace exon 2 and exon 3 of the ERp44 gene. The Loxp-neomycin-Loxp cassette was deleted in mutant mice by crossing them with CMV-Cre mice. 129 × C57BL/6 genetic background mice were backcrossed to C57BL/6 for at least 5 generations. For genotyping, a set of two primers was designed within or without exon 2 and exon 3 for both ERp44-WT-F/R and ERp44-G--F/R. The sequences of the genotyping primers are listed in Table S1.

| Generation of ERp44 conditional knockout mice
The gRNA (GTTTTAGAGCTAGAAATAGC) sequence was designed to help Cas9 cut DNA. The sgRNA with T7 promoter was transcribed in vitro using the MEGAshortscript™ Kit (Thermo Fisher, Cat No. AM1354). Cas9 (Addgene, Cat No. 41815) including the T7 promoter was transcribed using the mMACHINE T7 ULTRA Kit (Thermo Fisher, Cat No. AM1345) and purified using the MEGAclear Kit (Thermo Fisher, Cat No. AM1908). The targeting donor was designed to replace exon 2 in the wild-type allele with two flanking loxP sequences and two homologous arms. gRNA, Cas9 and donor DNA were microinjected into C57BL/6 zygotes and transplanted into pseudopregnant mice. The founder was genotyped with primers against sequence-F/R, which detected the left loxP sequence, and screen-F/R, which spanned two loxP sequences. The genotype was then verified with Sanger sequencing. The sequences of relative primers are listed in Table S1.

| Quantitative real-time PCR (qPCR)
Total RNA was isolated with TRIzol, and cDNA was synthesized with the PrimeScriptTM RT Reagent Kit and gDNA Eraser (TaKaRa, Cat No. RR037A). Real-time PCR was performed using the SYBR Green PCR Mastermix (Solarbio, Cat No. SR1120) and a Rotor-Gene-Q instrument (Qiagen). The fold change in target expression between WT and KO mice was calculated using the 2 −ΔΔCt method, 29 and GAPDH was used as an internal reference. The primers are shown in Table S2.

| Immunohistochemistry and immunofluorescence
The embryos were excised in cold PBS (10 g/L NaCl, 0.25 g/L KCl,

| AV cushion explant assay
The AV cushion explant assay was modified from the previous reports. 30 Briefly, endocardial cushions from the AV cushion were explanted on rat tail collagen gel (Thermo Fisher, Cat No. A1048301).
After overnight incubation, the medium was added (M199, Gibco; 1% FBS, Invitrogen; 0.1% ITS, Gibco; 100 U/ml penicillin, 100 mg/ ml streptomycin, INALCO), and elongated or spindle-shaped mesenchymal cells were counted after 48 h and analysed according to Xiong et al. 31 The degree of EndMT of a given explant was assessed based on the number of mesenchymal cells relative to the mean in all control groups, which was defined as 100% EndMT.
Then together with VSV-G and pHIT, they were transfected into Plat E cells for retrovirus packaging. Quality-checked retroviruses infected HEK-293T cells supported by polybrene for follow-up puromycin screening until obtaining modified HEK-293T cell line stably expressing VEGFA165-myc. Details of the above procedure referred to previous documents, 32,33 and it also guided the establishment of the H9C2 cell line overexpressing ERp44 (ERp44 OE).

| Western blot
For the Western blot analysis, the prepared samples were homog- Actin or GAPDH was used as an internal reference to assess protein levels. Three independent repetitive trials were performed.

| Transfection and co-immunoprecipitation (Co-IP)
HEK-293T cells were cultured in high-glucose Dulbecco's modified Eagle's medium containing 10% foetal bovine serum, 100 U/ml penicillin and 0.1 mg/ml streptomycin (culture medium). Around 10 5 cells were seeded in each well of a 6-well plate for next-day transfection with TurboFect™ (Thermo Fisher, Cat No. R0533). The cells were washed twice with PBS, and 2 ml fresh culture medium was added

| RNA-seq and bioinformatic analysis
The total RNA was extracted from the heart AV cushion tissues of E9.5 according to the protocol of RNA-ease Isolation Reagent (Vazyme, Cat No. R701-01). After quality quantification, the total RNA was converted to cDNA library and performed with RNA sequencing by BGI Genomics Co., Ltd. The raw data were subjected to ggplot2 assay in R software (version 3.6.1) for describing advanced volcano plot. Identified genes were mapped to Gene Ontology (GO) terms to determine their biological and functional properties.

| Statistical analysis
Experiments were performed at least three independent biological replicates for each group. Differences were considered significant at * p < 0.05, ** p < 0.01 and *** p < 0.001 using one-way ANOVA or Student's t test. Unless otherwise indicated, all results are expressed as mean ± SEM.

| Generation and characterization of ERp44 knockout mice
To reveal the role of ERp44 in vivo, we generated knockout mice by disrupting the ERp44 gene in embryonic stem cells (ESCs) by homologous recombination ( Figure 1A). A genotyping study indicated that exon 2 and exon 3 were successfully deleted ( Figure 1B), and ERp44 expression was undetectable in knockout mice, whereas it was observed in WT mice via Western blot ( Figure 1C).

| ERp44 deficiency leads to perinatal embryonic lethality
To assess the effect of ERp44 deletion in mice, we carefully examined the phenotypes of mice. Specifically, ERp44 KO mice were smaller ( Figure 1D and G) and weighed significantly less ( Figure 1E) than their littermate controls, and this difference persisted into adulthood. Because the mice were crossed with heterozygotes, the birth rate of KO mice was 21.2% (31/146), slightly lower than the expected Mendelian rate (25%). Most KO mice died within 24 h after birth and exhibited marked cyanosis, and only a few (15% of all KO mice identified after birth) survived to adulthood ( Figure 1F).
Newborn KO mice initially exhibited normal breathing but turned pale ( Figure 1G) and showed symptoms of tachypnoea within a few hours (data not shown), suggesting abnormal cardiopulmonary function. By E9.5, the sizes of KO and WT littermate embryos were similar, but the development of KO embryos was retarded starting at E10.5 ( Figure 1H and I).

| Embryos deficient in ERp44 exhibit cardiac defects
The observed retarded development of the body and high mortality of ERp44 KO mice after birth suggested that cardiac function is developmentally impaired in these mice. To test this hypothesis, we examined the heart under a stereoscope and found that the hearts of newborn KO mice were biventricularly enlarged and exhibited dilated atria (Figure 2A

| ERp44 deficiency impairs endocardial cushion cell proliferation
Atrioventricular cushions (AVCs) give rise to atrioventricular septation and valves via a series of complex cellular processes. 5 We hypothesized that the heart defects observed in ERp44 KO mice were due to the abnormal development of AVCs. To test this hypothesis, we histologically analysed both WT and KO embryo hearts at E10.5 and E11.5. The H&E staining of hearts showed fewer cushion cells in the AVCs of KO mice compared with WT mice (Figure 3A 35 Alcian blue staining was performed at E9.5, which did not show significant differences in acid glycosaminoglycans in the cardiac jelly between WT and KO AVCs ( Figure S2).
These results indicate that the hypoplastic AVCs in ERp44 KO mice are primarily due to a proliferative defect of cushion cells.

| ERp44 is required for endocardialmesenchymal transition (EndMT)
EndMT is a major cellular process during the development of AVCs. 36 To test whether EndMT is affected in ERp44 KO AVCs, we employed an in vitro assay modified from Feng et al. 30 Briefly, AVCs from WT and KO E9.5 were explanted ( Figure 4A Figure 4D). Also, the mesenchymal migration of WT mice remained unchanged between co-culture and single cultivation ( Figure 4E). These indicated that co-culture brought out compensation to the mesenchymal cell migration of heart atrioventricular septum in KO mice. Hence, we deduced that the lack of some secreted factors, not excessive secretion of inhibitory factors, was responsible for the above phenotype.

| Myocardial loss of ERp44 is the primary cause of heart defects
To further assess the autonomous contribution of cells observed in ERp44 knockout mice, ERp44 conditional knockout mice were generated using the CRISPR/Cas9 technique. Briefly, two loxP sequences were inserted between ERp44 exon 2 in C57BL/6 mice ( Figure S3A), and the ERp44 floxed allele was inactivated in myocardial cells by crossing the mice with cTNT-Cre mice. cTNT-Cre delivers Cre within the myocardium after E7.5. 38 The hearts and other organs of cTNT-Cre +/− ; ERp44 fl/+ mice were collected, and the level of ERp44 mRNA was measured by RT-PCR to confirm the deletion.
ERp44 showed the shortened mRNA mutant band lacking exon 2 (73 bp) in the heart but not in other tissues ( Figure S3B).
Because cTNT-Cre +/− ; ERp44 fl/+ mice were viable and normal, we crossed them with ERp44 fl/fl mice ( Figure S3C). Like the conventional KO mice, cTNT-Cre +/− ; ERp44 fl/fl mice were smaller ( Figure 5A and B) and exhibited foetal lethality; that is, most mice died at birth  Figure 5G and H) and did not show obvious congenital heart defects ( Figure 5I), and the birth rate was normal ( Figure 5J).

| ERp44 KO suppresses VEGFA secretion
Multiple signal pathways affect the development of mouse endocardial cushions, such as TGFβ/BMP, Notch, ErbB, Wnt/catenin and VEGF. [41][42][43][44] Transcriptome assay and qPCR revealed multiple EndMTrelated genes, such as ErbB3, Itga4, Shh, Tgfβ1, Uty, Vegfa and Wnt3, were significantly decreased ( Figure 6A and C). The bioinformatic analysis also depicted that differential genes were mainly enriched in these biological processes and pathways related to EndMT, such as extracellular matrix binding, cell proliferation, extracellular region, hedgehog signalling pathway and ECM-receptor interaction ( Figure 6B). At the protein level, only VEGF expression was significantly reduced in KO mice compared with WT ( Figure 6D). We next investigated the expression of VEGF at different development stages of the heart. Immunohistochemical trial discovered VEGF had a significant decrease in the defective atrioventricular septum ( Figure 6E). Co-IP trial showed that ERp44 interacted with VEGFA at different stages of heart development ( Figure 6F). Western blot assay also indicated VEGF was all markedly decreased in E9.5, E12.5 and the P0 septal tissues of KO mouse hearts compared with WT ( Figure 6G and H).
To further confirm the relationship between ERp44 and VEGF, we investigated their cytoplasmic localization in H9C2 cells. Figure 7A and Figure S4) and Co-IP ( Figure 7D and dimer decreased in the medium ( Figure 7G). This suggested N101 was crucial for the maturation and secretion of VEGF. Yet, co-expression of ERp44 with VEGF165 or mutant N101Y revealed that glycosylation site N101 did not affect the interaction between ERp44 and VEGF165 ( Figure 7H). We also detected VEGF expression in ERp44 OE and ERp44 KO H9C2, respectively. Results showed that VEGF increased in OE H9C2 but notably decreased in KO H9C2 ( Figure 7C). Besides, CCK-8 analysis revealed ERp44 KO suppressed cell proliferation, and VEGF overexpression in ERp44 KO H9C2 cells, to some extent, rescued the above process ( Figure   S5). These data suggest that ERp44 is involved in the trafficking of VEGF protein.

Micrographs (
In summary, ERp44 interacts with immature VEGF in endoplasmic reticulum to facilitate its correct fold, maturation and extracellular secretion, and is further involved in the normal EndMT process during EC development. Adversely, ERp44 KO directly results in the decreased extracellular VEGF, which impairs the EndMT and is responsible for the dysplasia EC ( Figure 7I).

| DISCUSS ION
Here, we report that ERp44 contributes to the development of en-  26 We found that ERp44 KO mice were predominantly AVSD, with some PVSD and ASD and no apparent defect in the myocardium.
These results showed dysplasia in the AV cushion development. The proportion of abnormal KO heart was approximately 85%, which was in line with the lethality rate (about 80%). The fact that CHD phenotype penetrance was lower than 100% was also reported by other studies. 47,48 Possible explanations for this reason are as follows: 1, the defects were not found by serial section; and 2, the mild defects may heal at late-developing stages. As no obvious defects were found in KO mice related to OFT cushion development, we did not investigate particularly whether the OFT cushion EndMT was impaired. An explanation of these anomalies is that mesenchymal cells derived from cardiac neural crest cells can migrate in this area to a certain extent. 49,50 AV cushion morphogenesis proceeds in the time window from E9.5 to E11.5 in mice and has been widely reported to be controlled by signals from both the endocardium and myocardium. 36,51 After E11.5, heart valve remodelling proceeds. 40 53 Here, we confirmed that ERp44 interacted with VEGFA and positively controlled VEGFA secretion. VEGFA signalling plays a spatiotemporal role in controlling EndMT, and disrupting one VEGFA allele or overexpressing VEGFA causes AV cushion dysplasia. 54 In the myocardium, the resting calcium level also regulates EndMT.
Specifically, an increase in the intracellular calcium concentration can activate calcineurin, which promotes the translocation of NFATc2/3/4 into the nucleus to regulate the expression of a set of genes, including VEGFA. 40,55 However, we did not observe differences in the resting Ca 2+ level in isolated cardiomyocytes between In particular, the atrium, ventricular wall and septum of neonatal heart tissue were separated as previous description. 59 All statistical data are represented as means ± s. * p < 0.05, ** p < 0.01 and *** p < 0.001 | 13 of 15 BI et al. KO and WT mice at P0 (data not shown). The importance of ERp44 in the early secretion pathway has been well studied. 16,17,26,56,57 Several proteins have been found to interact with ERp44, including IgM, 17 adiponectin 58 and ERAP1, 26 and other clients, such as VEGFA, are speculated to bind to ERp44. 53 Here, we found that ERp44 directly binds to VEGFA, and C29 in ERp44 was important for the binding with VEGFA. Also, the glycosylation site of VEGF N101 was critical for its maturation and secretion but affects less on the interaction between ERp44 and VEGF165.
In conclusion, we discovered a novel and close link between the ER chaperone protein ERp44 and endocardial cushion defects.
Specifically, myocardial ERp44 contributes to the development of the endocardial cushion by affecting the proliferation of cushion cells and EndMT processes, and ERp44 directly regulates VEGFA by binding to the C29 site. The information gained from this study will improve our understanding of the mechanisms underlying AV cushion defects and may also provide a potential diagnostic strategy for congenital heart defects in humans.

E TH I C S S TATEM ENT
All animal experiments were performed following the guidelines of laboratory animal care (NIH publication NO. 85 -23, revised 1996) and with approval from the Institute of Biophysics Committee for Animal Care (Approval No. SYXK2019025).

ACK N OWLED G EM ENTS
The cTNT-Cre mouse train was kindly gifted by Prof. Weinian Shou

CO N FLI C T O F I NTE R E S T S
The authors declare no competing or financial interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
Main data generated or analysed during this study are included in this article, and detailed data are available from the corresponding authors on reasonable request.