Folate regulation of planar cell polarity pathway and F-actin through folate receptor alpha

Abstract

Folate deficiency contribute to neural tube defects (NTDs) which could be rescued by folate supplementation. However, the underlying mechanisms are still not fully understood. Besides, there is considerable controversy concerning the forms of folate used for supplementation. To address this controversy, we prepared culture medium with different forms of folate, Folic acid (FA) and 5-methyltetrahydrofolate(5mTHF), at concentration of 5uM, 500nM, 50nM and folate free, respectively. Mouse embryonic fibroblasts (MEFs) were treated with different folates continuously for three passages and cell proliferation and F-actin was monitored. We determined that compared to 5mTHF, FA showed stronger effects on promoting cell proliferation and F-actin formation. We also found that FOLR1 protein level was positively regulated by folate concentration and the non-canonical Wnt/planar cell polarity (PCP) pathway signaling was significantly enriched among different folate conditions in RNA sequencing analyses. We demonstrated for the first time that FOLR1 could promote the transcription of Vangl2, one of PCP core genes. The transcription of Vangl2 was down regulated under folate-deficient condition, which resulted in a decrease of PCP activity and F-actin formation. In summary, we identified a distinct advantage of FA in cell proliferation and F-actin formation over 5mTHF, as well as demonstrating that FOLR1 could promote transcription of Vangl2 and provide a new mechanism by which folate deficiency can contribute to the etiology of NTDs.

Graphical Abstract

Folic acid and its reduced derivatives 5-methyltetrahydrofolate could be transported into cells via FOLR1 mediating receptor-mediated endocytosis. FOLR1 could then translocate into nucleus and promote the transcription of Vangl2, the core gene of PCP signaling pathway. Increased expression of Vangl2 then up-regulate the expression of its downstream genes RhoA and Fmn1, the latter of which play a crucial role in the polymerization of actin microfilaments.

1. INTRODUCTION

Folate, also known as vitamin B9, plays crucial roles in multiple biological processes, including DNA synthesis, one carbon metabolism, DNA and protein methylation and neurodevelopment. Folates exist in either a reduced or oxidized form(1). The majority of natural occurring folate exists in reduced form and has been found in a range of foods such as green leafy vegetables, liver and fruits(2). Folic acid (FA) refers to the oxidized synthetic form of folate and is the major synthetic form which is used in dietary supplements and food fortification programs(1). FA is inactive in the human body and is converted to tetrahydrofolate(THF) and then 5,10-methylene-THF, and then later is reduced to the active form, 5-methyltetrahydrofolate(5mTHF) by a series of enzymatic reactions. 5mTHF could provide methyl groups to the S-adenosylmethionine (SAM) cycle, and SAM is a ubiquitous cofactor in all cellular methyltransferase reactions(3). Folate is also crucial for DNA synthesis and repair. It helps convert uracil to thymine(4). When folate is deficient, the possibility of uracil misincorporation into DNA increases(5).

Folate deficiency has been widely demonstrated to be associated with neural tube defects (NTDs), which are severe congenital malformations resulting from the failure of proper neural tube closure(6). It has been reported that the dietary folate supplementation decreased the occurrence of NTDs by 72%(7). However, the underlying mechanism is still not well understood. Neural tube closure morphogenesis involves a complex series of steps involving convergent extension, apical constriction and interkinetic nuclear migration, as well as precise molecular control via signaling pathways such as the non-canonical Wnt/planar cell polarity pathway and Shh/BMP signaling(8). Rapid cell proliferation during neural tube closure requires the synthesis of large amounts of nucleotides for DNA replication(9), while an inadequate supply of nucleotides to neuroepithelial cells might result in neural fold retardation and NTDs(10). Extensive studies in human and vertebrate models have strongly implicated PCP pathway genes in the etiology of NTDs(11). It is known that both folate and PCP pathway signaling is involved in cytoskeleton regulation(12,13), but whether and how folate directly regulate PCP gene expression is unknown.

The importance of folate supplementation during pregnancy for preventing NTDs has been widely accepted. Currently, the most commonly used folate form for supplementation occurs as the oxidized inactive form of folate, FA(14). However, recent studies have reported adverse effects in individuals receiving high FA exposure(15), as well as in mouse studies (16). In order to circumvent these side effects, it has been suggested that there might be health benefits in replacing FA with 5mTHF, as 5mTHF is naturally occurring, it is well absorbed, and its bioavailability is not affected by metabolic defects(17). Studies also suggest that using 5mTHF instead of FA could prevent unconverted folic acid serum accumulation, and reduce the potential negative effects of FA such as concealing megaloblastic anemia due to vitamin B12 deficiency(18).

Previous studies suggested that 5mTHF ensures a stability in serum folate levels that cannot be obtained with FA, and that FA supplementation has comparable effect as 5-MTHF supplementation with respect to its physiological activity, bioavailability, and absorption at equimolar doses(19). It has been shown that FA treatment increased cell proliferation compared to treatment with 5mTHF at all concentrations after 3 days, whereas there was no difference following treatment for 5 days(20). Meanwhile, a systematic review revealed that 3 out of 23 studies showed a statistically significant difference in folate bioavailability of 5mTHF, which is more effective at increasing folate levels in study participants(21). However, there remain some methodological limitations and conflicting results in this study(21), and more research is required for greater clarification regarding the bioavailability and other cellular activities of different forms of folate.

The goal of our study was to investigate the effect of different folate forms and concentrations on cellular activities such as proliferation and cytoskeleton formation. We also explored the mechanism by which folate regulates PCP pathway signaling, which might reveal the underlying mechanism by which folates serve to prevent NTDs.

2. MATERIALS AND METHODS

2.1. Cell culture and medium preparation

Basal medium (500ml) was prepared using 50ml 10× folate-free RPMI 1640 medium (Sigma, R1145-500ml) supplemented with 50ml Fetal bovine serum (Gibco, 26140079), 5ml 100×glutamax (Gibco, 35050-061), 5ml 100× Anti-Anti (Gibco, 15240-062) and 1.85g Sodium Bicarbonate. Basal medium were supplemented with FA(Sigma, F8758) and 5mTHF(Sigma, 1379081-300MG) at the concentration of 5uM, 500nM and 50nM, respectively. Prepared mediums were stored at 4°C with shielding from light exposure.

MEFs were thawed and plated in a 10cm dish coated with gelatin and were cultured with 5uM FA RPMI medium for adaptation. At 90–100% confluency, cells were passaged to 6-well plates at a density of 0.2×10^6 per well and cell culture medium contained with 5uM, 500nM, 50nM and 0nM of FA or 5-mTHF, respectively. After 4 days, cells were trypsinized and resuspended in 1ml culture medium. 20ul cell suspension was aliquoted and mixed with 20ul 0.4% trypan blue, and 10ul mixture was used for cell counting with a CellDrop FL Fluorescence Cell Counter(DeNovix).. Cells were passaged to 6-well plates at a same density of 0.2×10^6 per well and continuously cultured for 3 passages. At the endpoint, portions of the cells were harvested for protein and RNA extraction, while the rest of the cells were passaged again for Immunofluorescence (IF) staining and the scratch assay. Western Blotting and RT-QPCR were used to examine target gene expression.

2.2. EdU and TUNEL labeling

EdU labeling was performed using the BeyoClick^™ EdU Cell Proliferation Kit (C0071, Beyotime). Briefly, cells were incubated in medium containing 10 μM EdU (5-ethynyl2-deoxyuridine) for 2 hour followed by fixation and permeabilization. The coverslip was then mounted using the Pro-LongTM Glass Antifade Mountant with NucBlue reagent (ThermoFisher Scientific, P36981). Photomicrographs were collected using epifluorescence microscopy (Leica SP8). Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining was performed using the Colorimetric TUNEL Apoptosis Assay Kit (C1098, Beyotime) according to the manufacturer’s protocol. After being fixed and permeabilized, endogenous peroxidase activity in cells was blocked with hydrogen peroxide. The sections were then incubated with the TdT enzyme followed by streptavidin-HRP buffer. The labeling was visualized with diaminobenzidine. The sections were stained with hematoxylin and imaged using Olympus fluorescence microscope.

2.3. Plasmids and transfection

The plasmids pCMV3-C-GFPSpark-Ctrl (CV026) and pCMV3-C-GFPSpark-FOLR1 (HG11241-ACG) were purchased from Sino Biological. Prior to transfection, medium was replaced with the culture medium described above without antibiotics. Transfection was performed at 50–70% confluency using Lipofectamine 2000 (ThermoFisher Scientific, 11668019) according to the manufacturer’s protocol. In brief, 2ug plasmids and 6ul Lipofectamine^® 2000 reagent were mixed with 150ul Opti-MEM medium, respectively. The two tubes were then gently mixed and incubated for 20min under room temperature, and were then added to cell culture medium. Further experiments were performed 36 hours after transfection.

2.4. Immunofluorescence staining

MEFs were continuously cultured for 3 passages in medium with different forms and/or concentrations of folate. After 3 passages, 10^5 cells were plated on 35mm glass bottom dishes pretreated with 0.1% Gelatin (EMD Millipore, ES-006-B). Immunofluorescence staining (IF) was performed 48 hours later. Cells were fixed with pre-warmed 4% paraformaldehyde for 20 minutes, then permeabilized with 0.3% Triton X-100 in TBST for 20 minutes, followed by 10% NGS in TBST blocking for 1 hour at room temperature. Phalloidin (1:1000, Santa Cruz Technologies, sc-363791) and Hoechst (1:10000, Sigma, H3570) were used to label F-actin and nucleus, respectively. Images were taken using a deconvolution microscope (Nikon T2).

2.5. Scratch assay

After cells were continuously cultured for 3 passages in different folate media, cells were plated onto 6-well plates at the same density of 1×10^6/well. On the second day, a narrow wound-like gap in the middle of the plate bottom was created using a 1000ul pipette tip when cell confluency reached 80–90%, the location of which was labeled on the back of the plate. Images were taken under 4×microscope at 0h and 6h after scratching. The edges of cells migrating into the gap was depicted by a yellow line and the migrating distances at 10 different positions were picked and analyzed through image J.

2.6. Western Blotting

Cells were collected with cell scrapers and lysed with RIPA buffer (Thermo Scientific, 89900) with protease inhibitor (Sigma, 11836170001) and were then centrifuged. The supernatant mixed with SDS loading buffer before being boiled for 10 minutes at 95°C. 25μg of total protein were loaded into the wells of the 4–15% SDS-PAGE gel (Bio-Rad, 4561086) and run for 40–50min at 130V until indicator reached the bottom of the gel. The proteins were transferred onto a NC membrane (Bio-Rad, 1620115) by wet-transfer method at 4°C, 325 mA for 55min. The membrane was blocked for 1 h at room temperature using blocking buffer and the proteins were immunoblotted with Vangl2 antibody (1:1000, Proteintech, 21492-1-AP), RhoA antibody (1:500, Santa Cruz Biotechnology, sc-418), FOLR1 antibody (1:1000, Invitrogen, PA5-86666), Fmn1 antibody (ThermoFisher Scientific, PA5-89020), COBL antibody (Atlas Antibodies, HPA019167) and GAPDH antibody (Cell Signaling Technology, 5174S) overnight at 4°C. On the second day, the membrane was incubated with HRP-conjugated anti-rabbit antibody or anti-mouse antibody (1:5000, Cell Signaling Technology, 7074S and 7076S) for 1 hour at room temperature. Membranes were rinsed with ECL substrates (Thermo Scientific, 34580) and protein bands were detected using a BioRad Gel Doc XR+ Imaging System.

2.7. Real-time quantitative PCR

Total RNA of cells was extracted using TRIzol reagent. In brief, 0.2 ml of chloroform was added per 1 ml of TRIzol reagent. Tubes were vortexed vigorously for 15 seconds, incubated at room temperature for 5 minutes and then centrifuged at 12,000xg for 15 minutes at 4°C. RNA remains in the aqueous phase were recovered by precipitation with isopropanol. RNA pellets were washed using 75% ethanol and were redissolved in DEPC-treated water. RNA concentration were adjusted to 100ng/ul. Reverse transcription PCR was performed using High-Capacity cDNA Reverse Transcription Kit (ThermoFisher Scientific, 4368813) and real-time quantitative PCR was performed using SYBR GREEN PCR Master Mix(ThermoFisher Scientific, 4309155) according to the manufacturer’s protocols, respectively. Primers were synthesized by Integrated DNA Technologies (IDT). Vangl2 forward primer sequence: CGGGCTATTCCTACAAGTCG, reverse primer Sequence: TATCTCCACGACTCCCATCC; HPRT1 forward primer sequence: CCTGGCGTCGTGATTAGTGATG, reverse primer sequence: GCAAGACGTTCAGTCCTGTCCAT.

2.8. RNA sequencing and analysis

MEFs were continuously cultured for 3 passages and total RNA was extracted using TRIzol reagent as previously described. NEBNext^® Ultra^™ RNA Library Prep Kit (New England Biolabs, E7530L) was used for library preparation according to the manufacturer’s protocol. Library was sequenced by Admera Health company using HiSeq2000. Data analysis was performed by using Kallisto and DEseq2(22). Gene set enrichment analysis (GSEA) is carried out using javaGSEA2-3.0(23).

2.9. Crispr/Cas9 knock-out

CHOPCHOP online software was used to design sgRNA for Crispr/Cas9. The vector pSpCas9 BB-2A-GFP (PX458)(48138) was purchased from Addgene, and the target sequence “CCCCCATCCAGTGTCGACCC” was cloned into the vector by Genscript. The constructed plasmids were transfected into Hela cells using Lipofectamine 2000 as previously described, and cells expressing GFP signal were picked up by 10ml pipette tip and were transferred to another well for expanding. Cell clones continuously expressing GFP signal were kept expanding, then Sanger sequencing and Western Blotting were performed for verification. Primers for sequencing were synthesized by IDT, forward:GCCACTGACCACAGCTCTTTC, reverse: AGGGCAGGGATTTCCAGGTA.

2.10. Statistical analysis

Using GraphPad Prism 9 software, data were presented as the mean values±the SEM. Two group differences were determined by an Student’s t test and multiple comparisons by a two-way ANOVA test. P-values < 0.05 were considered statistically significant.

3. RESULTS

3.1. Effect of the form and concentration of folate on the proliferation of MEFs.

To comprehensively determine the effects of different forms and concentrations of folate on cell proliferation during cell passages, we prepared folate-free cell culture medium and added 5uM, 500nM and 50nM of folic acid and 5mTHF, respectively. As commercial cell lines had already been cultured in commercial medium with several micromolar of folate which is much higher than the physiological level in both human and mouse cells, so the MEFs were continuously cultured in those medium for 3 passages. As shown in Fig. 1A, cell counts in the first passage after 96 hours of culturing were reduced as folate concentration decreased for both folate forms studied. Besides, cell counts were lowered in the 5mTHF group compared to that in FA under the 5uM condition, whereas there was no difference of cell counts between two forms of folate at both 500nM and 50nM concentrations. Cells were continuously cultured for another 2 passages and cell counts were also reduced in response to decreasing folate concentration for both folate forms in the second and third passages (Fig. 1B–C), and there were statistical differences between FA and 5mTHF under all three concentrations. In addition, differences in cell counts among groups were increased as cell passages increased, which might be due to excess folate in the cells that was gradually consumed as cells were passaged, so that the response to FA and 5mTHF could be better observed. We further performed EdU and TUNEL assays to explore the mechanism of cell counts alterations, the results(as seen in Fig. 1D–E), were consistent with the published results that folate promotes cell proliferation in a dose-dependent manner(24) and that folate deficiency could cause programmed cell death(25). It is worth noting that compared to FA, 5mTHF showed weaker effects on promoting cell proliferation and stronger effects on causing cell death especially under 500nM and 50nM conditions, respectively.

Figure 1:

Effect of different forms and concentrations of folate on cell proliferation. (A-C) Cell number of MEFs continuously cultured in medium with different forms and concentrations of folate for one, two, and three passages, respectively. (D-E) EdU and TUNEL were performed on MEFs cultured for three passages to explore the effects of different forms and concentrations of folate on cell proliferation and cell death(* p < 0.05, ** p < 0.01, *** p < 0.001, orange asterisk represents significance between groups of FA, blue asterisk represents significance between groups of 5mTHF, black asterisk represents significance between FA and 5mTHF at same concentration). Data were presented as the mean values±the SEM.

3.2. Effect of the form and concentration of folate on F-actin formation

Folate supplementation is effective in decreasing the prevalence of NTDs. Neural tube closure is a series of continuous processes including neural plate bending and fusion, initiating in the mid-cervical region of the embryonic neural plate and advancing along the anterior-posterior axis in a zipper-like fashion. The molecular mechanism of neural tube closure involves apical constriction driven by the recruited F-actin cable. It has been reported that neural crest cells and neural tube explants grew under folate-deficient conditions exhibited significantly reduction in directed migration which partially attributed to cytoskeleton changes caused by perturbations in F-actin formation and/or assembly(12). To confirm the effect of folate concentration on F-actin and to further explore the effect of different forms of folate on F-actin formation, we performed immunofluorescence staining of F-actin and the scratch assay on MEFs cultured under the above conditions after 3 passages. As seen in Fig. 2A–B, fluorescence signal of F-actin labeled by phalloidin were obviously decreased in both FA and 5mTHF as the folate concentration decreased. The F-actin signal in cells treated with 500nM and 50nM 5mTHF was also obviously lowered compared to that treated with FA at the same concentration, and that in the results of the 50nM 5mTHF group appeared comparable to the folate-free group. We also performed scratch assay to measure cell migration which could be driven by cytoskeleton. As seen in Fig. 2C–D, migration distance of cells cultured with 500nM and 50nM 5mTHF was significantly lower than that treated with 500nM and 50nM FA, respectively. In addition, migration distances in 500nM groups were similar as in 5uM groups and were higher than in the low folate groups. Based on previously reported studies, the physiological serum folate level for mice is lower than 300nM(26). To further confirm the above results, we repeated the IF experiments using human retinal pigment epithelial-1(RPE1), the plasma folate level of which is less than 50nM. As seen in the Supplementary Fig.1, F-actin signal in RPE1 cells were also affected by decreased folate concentration and seemed weaker in cells treated with 5nM 5mTHF compared to 5nM FA.

Figure 2:

Effect of different forms and concentrations of folate on F-actin formation. (A) MEFs were continuously cultured in medium with different form and concentration of folate for three passages, F-actin was labeled by phalloidin, and the nucleus was labeled by Hoechst 33342. Represented images in white box were zoomed in. Scale bar 10um. (B) Quantification analysis of mean fluorescence intensity in panel A. (C) Scratching assay was performed on cells treated for three passages and were imaged under the microscope(4×) at 0 h and 6 h after scratching, respectively. Migration distances of MEFs cells at ten different positions were picked and analyzed through image J. Scale bar 1000um. (D) Quantification analysis of distance traveled in panel C. (* p < 0.05, ** p < 0.01, *** p < 0.001, orange asterisk represents significance between groups of FA, blue asterisk represents significance between groups of 5mTHF, black asterisk represents significance between FA and 5mTHF at same concentration). Data were presented as the mean values±the SEM.

3.3. RNA-seq analysis on MEFs treated with different form and concentration of folate

To further identify the differential gene expression underlying the treatment with different forms and concentrations of folate, we performed RNA-seq on MEFs cultured for 3 passages under different folate conditions. Principal component analysis (PCA) was conducted to investigate the differences in the transcriptional profiles amongst the different groups. As seen in Fig. 3A, all three biological replicates from each group showed high reproducibility. It also revealed that seven groups with different treatments were clearly discriminated as five distinct clusters among which FA-5uM, FA-500nM and 5mTHF-5uM were clustered together, and that the degree of gene expression pattern alterations were consistent with the folate concentration alterations. These results indicated that gene expression patterns among groups with different folate culture conditions were altered based upon the different concentration and form of folate used. The differentially expressed genes among the groups were further analyzed using hierarchical clustering methods. As shown in Fig. 3B, we chose 40 of the most significantly differentially expressed genes to form the heatmap. Hierarchical clustering reveals that the transcriptomic profiles under low FA or 5mTHF treatments are similar to the FA-free group. Additionally, FA-5uM, FA-500nM, and 5mTHF-5uM treatments demonstrate analogous overall profiles, further supporting the PCA results.

Figure 3:

Transcriptional profile and differentially expressed genes analysis. (A) PCA showed the difference in transcriptional profile among groups. (B) Heatmap showed clustered the top 40 differentially expressed genes among groups with different treatments.

We further performed Gene Ontology (GO) term enrichment analysis on differentially expressed genes (DEGs) between groups. GO analysis on enriched up-regulated DEGs in 50nM 5mTHF compared to 50nM FA group is shown in Fig. 4A, which included the regulation of cell motility and migration, tube morphogenesis and negative regulation of cell proliferation in terms of biological processes. GO analysis on enriched down-regulated DEGs in 50nM 5mTHF compared to 50nM FA group was shown in Fig. 4B, which included the microtubule-based process, microtubule cytoskeleton organization in terms of biological processes, as well as polymeric cytoskeletal fiber in terms of cellular components. GO analysis on enriched up-regulated DEGs in 500nM 5mTHF compared to 500nM FA group was shown in Fig. 4C, which included the negative regulation of cell proliferation in terms of biological processes. These enriched pathways further substantiate the influence of both the form and concentration of folate on regulating cytoskeletal-related pathways, which in turn affects cell motility and migration.

Figure 4:

GO term enrichment analysis on enriched DEGs between MEFs treated with different forms of folate at the same concentration. (A) GO analysis on enriched up-regulated DEGs between 50nM 5mTHF and 50nM FA. (B) GO analysis on enriched down-regulated DEGs between 50nM 5mTHF and 50nM FA. (C) GO analysis on enriched up-regulated DEGs between 500nM 5mTHF and 500nM FA. (D) GO analysis on enriched down-regulated DEGs between 500nM 5mTHF and 500nM FA.

Considering the significant role of the PCP signaling pathway in F-actin formation and neural tube closure, we also performed differentially expressed PCP signaling pathway genes analysis among groups to form heatmap and Venn diagrams. As shown in Fig. 5A, PCP genes were discriminated and clustered distinctly, which suggested that the PCP signaling pathway could be affected by different forms and concentrations of folate. Meanwhile, transcriptome Venn diagram of DEGs sets in the PCP pathway revealed that compared to 50nM FA, 37 and 12 PCP genes were up- and down-regulated respectively, in cells cultured with 500nM FA(Fig. 5B). Compared to 50nM 5mTHF, 46 and 18 PCP genes were up- and down-regulated in 500nM 5mTHF culture conditions(Fig. 5C). In the 500nM FA group, there were 5 and 11 PCP genes that were up- and down-regulated when compared with the 500nM 5mTHF treatment group (Fig. 5D). Overall, these results suggested that RNA expression profiles could be altered by treatment with different forms and concentrations of folate, and notably, the PCP signaling pathway was upregulated with increasing folate concentrations, with a great impact from FA rather than 5mTHF.

Figure 5:

Analysis of differentially expressed PCP signaling pathway genes. (A) Heatmap of enriched PCP signaling pathway genes among groups treated with different forms and concentrations of folate. (B) Transcriptome Venn diagram of DEGs sets in PCP pathway between 500nM FA and 50nM FA. (C)Transcriptome Venn diagram of DEGs sets in PCP pathway between 500nM 5mTHF and 50nM 5mTHF. (D)Transcriptome Venn diagram of DEGs sets in PCP pathway between 500nM 5mTHF and 500nM FA.

3.4. Folate regulation of F-actin through PCP signaling pathway

The PCP signaling pathway plays a crucial role in the process of neural tube closure through the regulation of convergent extension. Mutations of Vangl2, is a key component in the PCP pathway, have been associated with NTDs in both humans and mice. It has also been reported that folate deficiency inhibits expression of Vangl2(27). To determine whether folate regulates the protein level of PCP signaling pathway genes, we evaluated protein expression of Vangl2, the PCP downstream gene Ras homolog family member A (RhoA), and its targeted gene Fmn1, which plays an important role in F-actin polymerization. We also examined the expression level of COBL, a non-PCP gene which is responsible for F-actin polymerization. As shown in Fig. 6A–D, the protein levels of Vangl2, RhoA and Fmn1 were all significantly reduced in response to decreasing FA and 5mTHF concentrations. At equivalent folates concentration, these PCP target proteins had significantly higher expression in FA group compared to cells cultured with 5mTHF. The COBL protein level showed no difference irrespective of folate form (Fig. 6A,E).

Figure 6:

Folate regulation on PCP pathway proteins and F-actin. (A) Western Blotting analysis of PCP pathway proteins in MEFs treated with different forms and concentrations of folate. GAPDH was used as control. (B-E) Quantification analysis of PCP target proteins in panel A. (orange asterisk represents significance between groups of FA, blue asterisk represents significance between groups of 5mTHF, black asterisk represents significance between FA and 5mTHF at same concentration). (F) Immunofluorescence analysis of F-action with and without Rho inhibitor treatment (magnification 20x, scale bar 200um). (G) Quantification analysis of mean fluorescence intensity in panel F. (* p < 0.05, ** p < 0.01, *** p < 0.001). Data were presented as the mean values±the SEM.

To further explore whether the regulation of folate on F-actin was through PCP pathway, we treated MEFs with 30uM Rhosin, a RhoA inhibitor, for 36 hours. F-actin signal was weakened after Rhosin treatment (Fig. 6F).

3.5. Folate deficiency altered RNA transcription but not protein stability of Vangl2

We further investigated the effect of folate on Vangl2 transcription and translation. Cycloheximide (CHX), a protein synthesis inhibitor, was used to explore the protein stability. MEFs were treated with 5uM, 500nM and 50nM FA for 3 passages, and then were treated with 10ug/ml CHX and DMSO, respectively, for 16 hours. The level of Vangl2 protein from cells treated with CHX/DMSO was significantly decreased in the 5uM FA group. There was no significant difference between the 500nM and 50nM FA groups (Fig. 7A–B). Vangl2 mRNA level was positively associated with folate concentrations in both forms, and that was significantly lower in cells treated with 5mTHF compared to FA in 5uM and 500nM groups (Fig. 7C). These results showed that folate can regulate Vangl2 transcription.

Figure 7:

Folate regulation of Vangl2 expression. (A) Western blotting analysis of Vangl2 in MEFs treated with CHX or DMSO under different concentrations of FA. GAPDH was used as the loading control. (B) Quantification analysis of panel A, the stability of Vangl2 was shown by the ratio of relative protein level in cells treated with CHX to that in cells treated with DMSO. The assay was repeated three times. (C) RT-qPCR was performed to explore the relative mRNA expression of Vangl2 in cells treated with different form and concentration of folate (orange asterisk represents significance between groups of FA, blue asterisk represents significance between groups of 5mTHF, black asterisk represents significance between FA and 5mTHF at same concentration). * p < 0.05, ** p < 0.01, *** p < 0.001, data were presented as the mean values±the SEM.

3.6. Folate condition affected PCP pathway through altering FOLR1 expression

We investigated whether folate regulation of Vangl2 transcription is mediated by folate receptor alpha, FOLR1. We determined that the FOLR1 protein was also decreased in cells given a reduced folate concentration and were lower in the 5mTHF groups compared to FA groups at all equivalent concentrations (Fig. 8A–B), a trend which was similar to that observed in Vangl2 protein level. We generated FOLR1 knockout Hela cells using CRISPR/Cas9 technology. Two cell clones were selected and confirmed by Sanger DNA sequencing (Fig. 8C). FOLR1 protein levels were shown to be significantly reduced compared with control Hela cells. The levels of Vangl2 mRNA and protein expression as well as RhoA and Fmn1 protein expression were significantly decreased, consistent with FOLR1 protein expression (Fig. 8D–F). Notably, relative mRNA expression level were also significantly decreased in FOLR1 knockout Hela cell clones compared to wildtype. Geng et al(27) reported that folate deficiency alters genomic methylation profiles which may be responsible for the alteration in Vangl2 expression. Our results demonstrated that folate deficiency leads to decreased FOLR1 protein levels, which could regulate the mRNA expression of Vangl2 and protein expression of the PCP core gene Vangl2 and PCP downstream genes RhoA and Fmn1.

Figure 8:

Folate regulation of PCP pathway signaling through FOLR1. (A) Western blotting analysis of FOLR1 protein was performed in MEFs cultured in medium with different folate for three passages. GAPDH was used as the loading control. (B) Quantification analysis of Panel A. Western blot assay was repeated three times(orange asterisk represents significance between groups of FA, blue asterisk represents significance between groups of 5mTHF, black asterisk represents significance between FA and 5mTHF at same concentration). (C) Sanger sequencing FOLR1 knockout Hela cell clones. (D) Western blotting on FOLR1 and PCP pathway targets was performed in control and FOLR1 knockout Hela cells. GAPDH was used as the loading control. (E) Quantification analysis of Panel D. Western Blotting was repeated for three times. (F) RT-qPCR was performed to explore the relative mRNA expression of Vangl2 in FOLR1 knockout Hela cell clones. * p < 0.05, ** p < 0.01, *** p < 0.001, data were presented as the mean values±the SEM.

A previously published study showed that FOLR1 could translocate into the nucleus and act as a transcription factor to regulate the transcription of target genes such as Fgfr4 and Hes1(28). To further understand the underlying mechanism by which FOLR1 regulates Vangl2 expression, we transfected Hela cells with GFP-tagged FOLR1 plasmids to observe the location of GFP signal in the cells. As seen in Fig. 9A, the GFP signal of cells transfected with control plasmid were evenly distributed throughout the whole cell, while the GFP signal in cells transfected with the FOLR1 plasmid were distributed in both the nucleus and cytoplasm. Moreover, in our Z-stack analysis of single GFP-FOLR1 transfected cell as depicted in Supplementary Video 1, the cross-section views (seen in Fig. 9B) showed obvious GFP signal inside the nucleus. We further extracted nuclear and cytoplasmic protein in Hela cells using Nuclear and Cytoplasmic Protein Extraction Kit (P0027, Beyotime) according to the manufacturer’s protocol and performed Western Blotting which demonstrated that FOLR1 protein could translocate into nucleus (Fig. 9C). The function of FOLR1 nucleus translocation and that whether FOLR1 could affect Vangl2 transcription remain to be further delineated.

Figure 9:

FOLR1 could be translocated into the nucleus and stimulate Vangl2 promoter transcription. (A) Subcellular localization analysis of GFP-FOLR1. Scale bar 5um. (B) Hela nucleus cross-section view demonstrated GFP-FOLR1 localized inside the nucleus. (C) Western blotting analysis of FOLR1 protein was performed in nuclear and cytoplasmic protein extractions in Hela cells. GAPDH was used as the cytoplasm marker, Lamin B1 was used as the nuclear marker.

4. DISCUSSION

It is well known that folates promote proliferation in various type of cells, and that folate concentration could affect the cytoskeleton and the migration of neural crest cells (12). However, there is still a lack of evidence showing the effect of different forms of folate on cell proliferation and formation and function of the cytoskeleton. Our study for the first time demonstrated that compared to 5mTHF, FA treatment enhanced a higher degree of cell proliferation than did other folate forms. This finding was unexpected, as we believed that the cells would preferentially benefit more from 5mTHF rather than FA supplementation, given that it was not necessary for it to be converted to a physiologically active form. We speculated that this might be due to their different affinity to the various folate transport systems. There are three type of folate transport proteins in the human body, including folate receptors (FOLRs/FRs), the proton-coupled folate transporter (PCFT), and the reduced folate carrier (RFC)(29). RFC is a low affinity, ubiquitously expressed transporter. It has a very poor affinity for FA and is specific for 5mTHF(30). PCFT could transport oxidized and reduced folates with similar efficiency into some tissues such as the liver under a low pH environment (31). FOLRs are glycosyl phosphatidylinositol (GPI)-anchored proteins that facilitate unidirectional folate uptake through endocytosis, among which FOLR1 is the major isoform expressed in a number of normal tissues such as kidney, lung, choroid plexus, placenta and ovaries(32). It has been reported that FOLR1 has a higher affinity for FA than for naturally occurring reduced folates. For example, the affinity of FOLR1 for FA is 14-fold higher than for 5mTHF(33). It is also been shown that FOLR1 represents the highest affinity folate transport system compared to RFC and PCFT(34), and also has a higher affinity to FA at neutral pH compared with other folate transporters(35). Based on this evidence, we speculated that FA could be more readily transported into cells compared with 5mTHF at the same supplemented concentrations due to its higher affinity to FOLR1.

In this study, we reported for the first time that FOLR1 could regulate PCP signaling pathway through promoting expression of Vangl2. It has been also reported that FOLR1 is enriched in the apical surface of the neural plate in Xenopus laevis, and that FOLR1 is necessary for neural plate cell apical constriction during Xenopus neural tube formation(36). However, there is a significant data gap with respect to the role of FOLR1 in PCP pathway signaling. We performed Crispr/Cas9 knock-out experiments and demonstrated that FOLR1 could regulate Vangl2 expression. Vangl2 is core protein in the PCP pathway which has a very important role in neural tube closure. PCP signaling is required for convergent extension (CE), which is primarily driven by planar polarized cell intercalation(37), and is an important process during neural tube closure facilitating how the neural plate undergoes narrowing along its mediolateral axis and extends along anteroposterior axis(38). This is biomechanically driven by the recruited F-actin cable at the zippering point(39). Vangl2 variants were reported to be associated with NTDs in both human(40) and mice(41). Acting downstream of Vangl2, RhoA functions in regulating cell shape and migration(42) through promoting Fmn1 expression which is essential for actin nucleation and elongation(43).

In the present study, we found that under folate-deficient conditions, FOLR1 and the PCP pathway core proteins Vangl2, RhoA and Fmn1 had decreased expression of their protein level. In FOLR1 knock-out cells, Vangl2 mRNA expression level and PCP pathway core proteins were significantly decreased as well. MEFs treated with Rhosin, the RhoA inhibitor, also showed reduced cell numbers and weaker F-actin staining. Based on all of the collective results, we believe that folate deficiency results in decreased FOLR1 protein levels, which down-regulated PCP pathway signaling. This might partially explain at least one of the mechanisms by which folate deficiency increases the risk for NTDs.

One laboratory previously reported that FOLR1 could act as a transcription factor as FOLR1 could translocate from the cytoplasm to the nucleus, bind to cis-regulatory elements at promoter regions and therefore directly regulate the expression of target genes such as Fgfr4, Hes1(28), Oct4, Sox2, and Klf4(44). In the present study, we performed immunofluorescent staining and observed the presence of a GFP-FOLR1 signal in the nucleus. However, the function of FOLR1 following its translocation to the nucleus and that whether FOLR1 could affect Vangl2 transcription remains to be further explored.

Another interesting result is that FOLR1 protein level was reduced as a result of culturing cells under folate deficient conditions instead of being upregulated. Leamon and co-workers(26) performed analysis on the effects of feeding mice both high and low folate-containing diets as related to associated changes in tissue-derived folate receptor levels. They found that serum and RBC folate concentrations sharply dropped immediately after the mice were switched to low folate diets, and tissue-related folate binding capacities were also decreased(26). We speculated that the cells were exposed to the low level of folate for a prolonged duration, such that the FOLR1 level were adaptively reduced.

In this study, we determined that FA showed a stronger effect in promoting cell proliferation and F-actin formation compared to 5mTHF. This is not consistent with our previous knowledge where we believed that there was a cellular preference for 5mTHF primarily due to it being naturally bioactive, and we suspected that it was due to high affinity for FA to FOLR1. We also found that a folate deficiency leads to decreased levels of FOLR1 protein, which could down-regulate the PCP signaling pathway. Reduced activity of the PCP pathway further affected formation of F-actin. These represent reasonable mechanisms by which a folate deficiency causes NTDs. Except for the studies published by Mayanil’s laboratory(28,44) that demonstrated FOLR1 can act as a transcription factor regulating the expression of genes including Fgfr4, Hes1, Oct4, Sox2, and Klf4, we demonstrated that FOLR1 can translocate into nucleus, and that FOLR1 could regulate the mRNA and protein expression of Vangl2, the core gene of the PCP pathway. Further studies are required to delineate in greater detail of the potential underlying mechanisms by which FOLR1 regulates VANGL2 expression.

DATA AVAILABILITY STATEMENT

RNA sequencing data has been uploaded to SRA database(accession number: PRJNA898490). Other data generated and analyzed in this study can be made available upon request from the corresponding author.