https://orcid.org/0000-0002-5131-9543
Abstract
Folic acid (synthetic folate, FA) is consumed in excess in North America and may interact with common pathogenic variants in methylenetetrahydrofolate reductase (MTHFR); the most prevalent inborn error of folate metabolism with wide-ranging obesity-related comorbidities. While preclinical murine models have been valuable to inform on diet–gene interactions, a recent Folate Expert panel has encouraged validation of new animal models. In this study, we characterized a novel zebrafish model of mthfr deficiency and evaluated the effects of genetic loss of mthfr function and FA supplementation during embryonic development on energy homeostasis and metabolism. mthfr-deficient zebrafish were generated using CRISPR mutagenesis and supplemented with no FA (control, 0FA) or 100 μm FA (100FA) throughout embryonic development (0–5 days postfertilization). We show that the genetic loss of mthfr function in zebrafish recapitulates key biochemical hallmarks reported in MTHFR deficiency in humans and leads to greater lipid accumulation and aberrant cholesterol metabolism as reported in the Mthfr murine model. In mthfr-deficient zebrafish, energy homeostasis was also impaired as indicated by altered food intake, reduced metabolic rate and lower expression of central energy-regulatory genes. Microglia abundance, involved in healthy neuronal development, was also reduced. FA supplementation to control zebrafish mimicked many of the adverse effects of mthfr deficiency, some of which were also exacerbated in mthfr-deficient zebrafish. Together, these findings support the translatability of the mthfr-deficient zebrafish as a preclinical model in folate research.
Introduction
In utero nutritional and genetic disturbances in folate metabolism predispose to the development of disease risk in later life (1,2). The gene methylenetetrahydrofolate reductase (MTHFR) encodes the rate-limiting enzyme in the 1-carbon cycle, producing the only endogenous source of the bioactive folate form 5-methyltetrahydrofolate (5-MTHF) (1). 5-MTHF is integral for homocysteine (Hcy) catabolism, transmethylation reactions and neurotransmission, and is the only folate form that can cross the blood–brain barrier (1,3,4). Rare pathogenic mutations in MTHFR underlie severe MTHFR deficiency (OMIM 607093; c1p36.22), the most common inborn error of folate metabolism (5,6). To date, a total of 34 mutations have been described in patients with severe MTHFR deficiency, the majority of which are missense mutations in the catalytic N-terminal domain (7). By comparison, up to 50% of the general population contain MTHFR variants encoding a thermolabile enzyme with reduced activity (e.g. g.677C > T), underlying a milder form of the disorder (8). Notably, disease consequences associated with both mild and severe MTHFR deficiency are wide-ranging, varying in severity and onset, and include neural tube defects, early and later-life neurobehavioral impairments and/or various cardio-metabolic and immune dysfunctions, including obesity (9–12). This clinical heterogeneity poses a challenge for timely diagnosis and treatment via nutritional therapeutic strategies (e.g. methyl donor supplement) (13–16) and emphasizes a need to identify modifiable risk factors.
To ascertain key elements that limit folate status and contribute to health and disease risk, understanding the interaction between folate intake and MTHFR variants is critical. In North America, widespread food fortification and use of prenatal supplements with folic acid (FA, synthetic folate) have been public health strategies for neural tube defect prevention (17), but have also resulted in gestational intakes that may exceed the tolerable upper level (18–20). Consequently, unmetabolized FA (UMFA) has been widely detected in plasma, breast milk and umbilical cord blood (20–24) and may interfere with 1-carbon metabolic processes.
Addressing current gaps in folate research using the rodent model poses challenges and has resulted in the need to validate new animal models for 1-carbon metabolic studies (25). The metabolism of folate is much faster in rodents compared with humans (26) and their use for high-throughput analyses during embryogenesis is limited and costly. Moreover, determining the effects of folate versus UMFA during gestation is not possible because of the confounding effects of the maternal-fetal nutrient supply. By comparison, the zebrafish (Danio rerio) offers a unique compromise between in vivo complexity and in vitro simplicity as it develops rapidly and externally, is optically clear and highly permeable and is readily genetically manipulable (27–29). Most importantly, the folate pathway in zebrafish has been reported to be structurally and kinetically similar to humans, and therefore can be an informative preclinical model for 1-carbon metabolic investigations (30–32).
In this study, our primary objective was to characterize and evaluate the effects of genetic loss of function of mthfr and external FA supplementation during embryonic development on energy homeostasis and metabolism in the developing zebrafish. We evaluated energy balance (food intake and metabolic rate), central regulatory gene expression, myeloid cell abundance, lipid accumulation (neutral lipids and cholesterol) and cellular metabolism (1-carbon metabolites). We identify changes consistent with human MTHFR deficiency, thus supporting the suitability of the zebrafish model and system for future folate metabolism and gene–environment studies.
Results
mthfr knockout (KO) in zebrafish does not result in overt morphological differences during embryonic development
In this study, we characterized 4 guide RNA (4gRNA) mthfr crispants and zebrafish containing a germline homozygous mutation (mthfr−/− mutant zebrafish) via disruption of the 5′ untranslated region (UTR) and transcriptional start site (TSS) of the mthfr gene (Fig. 1). The germline mutagenesis approach resulted in a 2058 bp deletion, which included the TSS (Fig. 2A). qRT-PCR confirmed mthfr mRNA to be downregulated in whole-4gRNA mthfr crispants and mthfr−/− zebrafish at 5 days postfertilization (dpf) by ~ 60% and ~80%, respectively, compared with their corresponding controls (Fig. 2B, P < 0.001). Both 4gRNA mthfr crispants and mthfr −/− mutants were viable (Fig. 2C) and did not exhibit gross morphological defects up to 5 dpf (Fig. 2D and E), with no differences noted in body length, body width, jaw length or eye diameter (Supplementary Material, Fig. S1), as well as no observable craniofacial skeleton defects as assayed by Alcian blue stain (Supplementary Material, Fig. S2).
mthfr zebrafish model creation and study design. Schematic of the steps to generate (A) 4gRNA mthfr crispant zebrafish and (B) germline mthfr −/− mutant zebrafish (HSC194). (C) Simplified study design. Metabolic analyses included measures of food intake, metabolic rate and lipid accumulation. Analysis indicated by the gray arrowhead was performed following a 10-day HCD challenge. Abbreviations: F0, mutagenized generation; F1 +/−, first-generation zebrafish heterozygous for mthfr; F2 −/−, second-generation zebrafish nullizygous for mthfr; hpf, hours postfertilization; MCFA, medium chain fatty acid; WT, wild-type zebrafish from AB background.
Generation of mthfr-deficient zebrafish by CRISPR mutagenesis. (A) Schematic representation of the zebrafish genome with guide RNA sequences identified in the mthfr gene. The 4gRNA target sequences (gRNA1/2/3/4-4) reside in exons 3, 5, 6 and 9. The two-gRNA approach (gRNA2 and gRNA6) targets sequences in the 5′ UTR and in exon 2, respectively, and results in a 2058 bp deletion (mthfr-atgdel) that includes the TSS. Black arrowheads show where gRNA cutting occurred. Boxes indicate exons, with white boxes representing UTR and gray boxes representing protein coding regions. Primer locations used for sequencing are shown. (B) Relative mthfr mRNA expression in 4gRNA mthfr crispants (4gRNA) and mthfr −/− mutant zebrafish (mthfr −/−, HSC194) compared with their respective controls at 5 dpf. Values are mean ± SEM, n = 2–3 biological replicates of n = 20 pooled zebrafish/group. Each pool was composed of n = 20 larvae. Analyzed by Student’s t-test. Significant at *P < 0.05 and ***P < 0.001. (C) Kaplan–Meier curve showing percent survival of WT and mthfr −/− mutant zebrafish larvae throughout embryonic development from 0 to 5 dpf. n = 60 larvae/group/petri dish at the start of analysis. (D, E) Representative image of (D) WT and (E) mthfr −/− mutant zebrafish at 5 dpf showing no obvious morphological differences. 10× magnification, scale bar = 200 μm.
mthfr disruption alters energy regulatory systems during embryonic development
We next measured food intake in first-time exogenously fed zebrafish larvae at both 5 and 8 dpf (via fluorescently labeled paramecia), as well as metabolic rate (via a NADH2 fluorometric assay) in larvae that received no exogenous feeds. At 5 dpf, mthfr −/− mutants had lower food intake (~30%, P < 0.001, Fig. 3A–C) and metabolic rate (~38%, P < 0.001, Fig. 3D) compared with wild-type (WT) controls. By comparison, food intake of mthfr −/− zebrafish was normalized to that of control by 8 dpf (Fig. 3E–G), whereas metabolic rate remained lower (~48%, P < 0.001, Fig. 3H) highlighting a persistent impact of loss of mthfr function on metabolic regulation.
Food intake and metabolic rate are affected by mthfr deficiency in larvae. (A, B) Representative images of zebrafish intestinal content following feeding with fluorescent paramecia in (A) WT controls and (B) mthfr −/− mutant zebrafish (mthfr −/−, HSC194) at 5 dpf. (C) Food intake, measured by relative CTCF, at 5 dpf in WT compared with mthfr −/− mutants. (D) Relative metabolic rate in WT versus mthfr −/− at 5 dpf. (E, F) Representative images following feeding with fluorescent paramecia in (D) WT and (E) mthfr −/− at 8 dpf. (F) Relative CTCF at 8 dpf in WT compared with mthfr −/− mutants. (G) Relative CTCF at 8 dpf in WT compared with mthfr −/− mutants. (H) Relative metabolic rate in WT versus mthfr −/− at 8 dpf. For all food intake measures: 80× magnification, scale bar = 200 μm, n = 20/group. All values are mean ± SEM. Analyzed by Student’s t-test. Significant at *P < 0.05, **P < 0.01, ***P < 0.0001.
mthfr deficiency reduced central regulatory gene expression
To offer insight into the central regulatory pathways that may be affected by mthfr gene disruption, select gene targets previously reported to be sensitive to fluctuations in folate status (33–35) were assayed in heads of mthfr −/− versus WT larvae at 5 dpf (Fig. 4). Overall, mthfr disruption induced the downregulation of target gene expression. Specifically, lower expression of several genes involved in neurotransmission and neuronal energy regulation was observed in mthfr −/− zebrafish including lower tyrosine hydroxylase (th) (1.3-fold, P = 0.05), dopamine signaling receptor 2a (drd2a) (2-fold, P < 0.05), catechol-o-methyltransferase (comta) (1.5-fold, P < 0.01), gamma-aminobutyric acid type A receptor-4alpha (gabra4) (1.5-fold, P < 0.05), glutamate receptor, ionotropic, N-methyl D-aspartate 1a (grin1a) (1.6-fold, P < 0.05) and brain-derived neurotrophic factor (bdnf) (1.6-fold, P < 0.01). mRNA expression of neuropeptide Y (npy), cocaine- and amphetamine-regulated transcript 4 (cart4) and gamma-aminobutyric acid type A receptor-3alpha (gabra3) was not affected.
Central regulatory gene expression was altered in mthfr −/− at 5 dpf. All values are relative to control. Mean ± SEM, n = 5 larval heads/pool, 4–5 pools/group. Analyzed by Student’s t-test. Significance at *P < 0.05, **P < 0.01. Abbreviations: mthfr −/−, 5-methylenetrahydrofolate reductase homozygous KO germline zebrafish mutants (HSC194); WT, wild-type control zebrafish.
mthfr deficiency in zebrafish affects tissue-specific myeloid cell abundance during embryonic development
Pathogenic variants in MTHFR may induce immune cell dysregulation underlying several of the disease consequences (36). We therefore investigated myeloid regulatory cells during embryonic development by generating 4gRNA mthfr crispants in the double transgenic line Tg(mpx:GFP); Tg(mpeg1:mCherry). Neutrophil number (as determined by quantifying mpx:GFP positive cells) in the dorsal head of both 4gRNA mthfr crispants and control larvae was low and did not localize to the central brain region (Supplementary Material, Fig. S3). This is consistent with the limited neutrophil CNS-infiltration seen with an intact blood–brain barrier (37). However, we observed a reduction in the number of macrophages (mpeg1:mCherry positive cells) in the dorsal head of 4gRNA mthfr crispants compared with cas9 controls (~25%, P < 0.01, Supplementary Material, Fig. S3), suggesting fewer microglia in mthfr crispant zebrafish. Notably, myeloid cells in the periphery were not affected at this time as a similar number and distribution pattern of labeled macrophages and neutrophils were observed in crispants and controls both under homeostatic conditions and following a tail-wounding assay (Supplementary Material, Figs S4 and S5). To confirm whether reduced macrophages were reflective of fewer microglia, we performed vital staining with neutral red (NR), a dye that marks microglia, at 4 dpf. Indeed, 4gRNA mthfr crispants had ~20% less NR+ cells in the optic tectum compared with cas9 controls (P < 0.01, Fig 5A, B and E). We next corroborated these results in mthfr −/− mutant zebrafish and confirmed a similar ~20% reduction at 4 dpf (P < 0.001, Fig. 5C–E) that was maintained to 6 dpf (data not shown). To ascertain whether reduced microglia in mthfr −/− zebrafish were because of defects in embryonic macrophage migration, we performed NR staining at 3 dpf (a time point when microglia rapidly populate the midbrain) but found no differences in number or distribution pattern of NR+ cells (data not shown). Microglia in the midbrain are responsible for clearance of dead/apoptotic cells, a process called efferocytosis. Thus, we next investigated whether reduced microglia in mthfr −/− mutant zebrafish was a consequence of increased efferocytic demand because of an increase in apoptotic cells. By performing Acridine Orange (AO) staining at 3 dpf, we found that mthfr−/− zebrafish had ~ 52% more AO+ cells compared with WT (P < 0.0001, Fig. 5F–H).
mthfr deficiency in zebrafish reduces microglia and increases apoptosis in the optic tectum. (A–D) Representative bright-field images of NR+ staining at 4 dpf for microglia in (A) cas9, (B) 4gRNA, (C) WT and (D) mthfr −/− mutant zebrafish (mthfr −/−, HSC194). 80× magnification, scale bar = 100 μm. (E) Number of NR+ cells at 4 dpf. n = 8–14 larvae/group. Analyzed by one-way ANCOVA with ‘batch’ as a covariate. Significance at **P < 0.01 and ***P < 0.001. (F, G) Representative fluorescent images of 3 dpf (F) WT and (G) mthfr −/− zebrafish stained with AO. (H) Number of AO+ cells at 3 dpf. White dotted line outlines area of quantification. 80× magnification, scale bar = 200 μm. Analyzed by Student’s t-test. Significant at ***P < 0.001. All values are mean ± SEM. Circles represent one individual zebrafish from batch 1 and squares represent zebrafish from batch 2.
Loss of mthfr in zebrafish recapitulates a biochemical profile of severe MTHFR deficiency
The folate (Fig. 6A–D) and methyl metabolite (Fig. 6E–L) contents of both 4gRNA mthfr crispants and mthfr −/− mutants were evaluated to determine whether they mirror the profile reported in severe MTHFR deficiency in humans (38), in support of the observed phenotypic analysis. Accordingly, 4gRNA mthfr crispants had ~80% and mthfr −/− mutants had ~90% lower 5-MTHF concentrations relative to their respective controls (P < 0.001, Fig. 6A). The concentration of 5,10-methylenetetrahydrofolate, which interconverts to 5-MTHF, was also lower in both 4gRNA mthfr crispants (P < 0.05) and mthfr −/− zebrafish (P < 0.01) compared with controls. The concentrations of synthetic FA, dihydrofolate and 5,10-methenyltetrahydrofolate were below the lower detection limit of quantification for all larvae pools at this time point. Both 4gRNA mthfr crispants and mthfr −/− zebrafish also had lower concentrations of methionine (MET, P < 0.05), and the alternative methyl donor betaine (P < 0.05), along with >2-fold higher concentrations of the methyltransferase inhibitor s-adenosylhomocysteine (SAH, P < 0.01) and higher cystathionine (P < 0.01). The SAM:SAH ratio was also lower in both 4gRNA mthfr crispants and mthfr −/− zebrafish, likely indicative of reduced methylation potential. Notably, mthfr −/− zebrafish also had significantly lower concentrations of tetrahydrofolate (THF) and 5–10-THF, and significantly higher choline and total Hcy concentrations (P < 0.05) compared with controls, suggesting a more robust response to mthfr disruption on methyl metabolites in the germline model.
mthfr deficiency in zebrafish reduces folates and alters 1-carbon cycle methyl metabolites. Relative concentrations of whole-body (A–D) folates and (E–L) related 1-carbon metabolites in zebrafish larvae at 5 dpf. Comparisons were conducted between cas9 versus 4gRNA mthfr crispants (4gRNA) or age-matched WT controls versus mthfr −/− mutants (mthfr −/−, HSC194). Values are mean ± SEM, n = 3–4 biological replicates of pooled zebrafish/group. Each pool was composed of n = 40 larvae for folates and n = 20 larvae for methyl metabolites for cas9 versus 4gRNA mthfr crispants and n = 50 larvae for folates and n = 30 larvae for methyl metabolites for WT versus mthfr −/− mutants. No differences in pooled weight were observed. Analyzed by Student’s t-test stratified by mutagenesis approach. Significant at *P < 0.05, **P < 0.01, ***P < 0.0001. Abbreviations: tHcy, total homocysteine; 5,10-CH-THF, 5,10-methenyltetrahydrofolate; 5,10-CH2-THF, 5,10-methylenetetrahydrofolate.
Loss of mthfr in zebrafish induces lipid dysregulation up to 15 dpf
Next, we explored whether mthfr disruption in zebrafish results in the accumulation of lipids, and specifically cholesterol, as has been reported in the Mthfr murine model (39–41). We first utilized our 4gRNA mthfr crispants to perform a long-term in vivo CE-BoDipy-C12® fluorescent feeding assay up to 15 dpf. 4gRNA mthfr crispants compared with cas9 control zebrafish had greater hepatic accumulation of the fluorescent CE analog (P < 0.001, Fig. 7A–G), which has been suggested to reflect aberrant cholesterol metabolism and/or transport (42). A greater percentage (~50%) of 4gRNA mthfr crispants also had fluorescent lipid punctae throughout their vasculature (P < 0.05, Fig. 7H, Supplementary Material, Video S1) confirming an increased predisposition to vascular lipid accumulation induced by mthfr disruptions. Notably, 4gRNA mthfr crispants also had greater neutral lipid accumulation, as assessed by Oil-ed-O (ORO) staining, in the viscera region that included the heart and liver area (P < 0.05) and in their whole body (P < 0.05), accounting for the more intensely stained dorsal artery, intersegmental veins and caudal vein compared with cas9 control zebrafish (Fig. 7I–M). Body length at 15 dpf (data not shown) was not different between zebrafish groups, limiting confounding effects on lipid parameters reported. We next investigated whether mthfr disruption resulted in the accumulation of lipids during early development by performing ORO staining in mthfr −/− mutants. At 5 dpf, mthfr −/− zebrafish had ~75% greater lipid deposition in their vasculature, head and viscera region that included the yolk-sac and heart (P < 0.0001, Fig. 7N–P), findings which were persistent up to 8 dpf (Fig. 7Q).
Lipid accumulation is greater in mthfr zebrafish compared with controls. (A–F) Representative bright field (BF) and fluorescent confocal images of (A–C) cas9 control (cas9) and (D–F) 4gRNA mthfr crispants (4gRNA) fed a 4% HCD with ce-BoDipy-C12® from 5 to 15 dpf. Fluorescent liver area is outlined by a white dotted line and green fluorescence represents CEs. 10× magnification, scale bar = 200 μm. (G) Relative fluorescent liver area and (H) percentage of larvae with excess circulating lipids were compared between cas9 and 4gRNA zebrafish. n = 13–18 larvae/group. Analyzed by one-way ANCOVA with ‘batch’ as a covariate. Significance at *P < 0.05. (I–L) Representative bright-field images of (I, J) cas9 and (K, L) 4gRNA zebrafish stained with ORO at 15 dpf following 10 days of HCD feeding. White dotted line outlines liver area. 20× magnification, scale bar = 500 μm for images I and K and 40× magnification, scale bar = 200 μm for panels J and L. (M) Calculated mean gray area of viscera region at 15 dpf. n = 4–5 larvae/group. Analyzed by Student’s t-test. Significance at *P < 0.05, **P < 0.01. (N, O) Representative BF images of WT controls and 5-methylenetrahydrofolate reductase homozygous KO germline zebrafish mutants (mthfr −/−, HSC194) at 5 dpf stained with ORO. 20× magnification, scale bar = 500 μm. Calculated mean gray area of viscera region at (P) 5 dpf and (Q) 8 dpf. n = 12–13 larvae/group. Analyzed by one-way ANCOVA with ‘batch’ as a covariate. Significance at *P < 0.05 and ***P < 0.0001. Letters and arrows show areas of interest: h, heart; da, dorsal artery; isv, intersegmental area; pcv, posterior cardinal vein; ys, yolk syncytial layer; L, liver; g, gallbladder; i, intestine. All values are mean ± SEM. Circles represent one individual zebrafish from batch 1 and squares represent zebrafish from batch 2.
FA supplementation does not rescue and may exacerbate effects of mthfr disruption in zebrafish independent of changes in 1-carbon metabolites at 5 dpf
To determine the effect of excess supplemental FA exposure on the developing zebrafish larvae, and to examine the interplay between mthfr deficiency and FA supplementation, zebrafish larvae were exposed to either 0 or 100 μm FA up to 5 dpf. FA exposure increased whole-body lipid accumulation, as assessed by ORO staining, in WT and mthfr −/− zebrafish as indicated by a significant FA Dose (P < 0.001) and Genotype (P < 0.01) effect but not an interaction effect (Fig. 8A). Metabolic rate (Fig. 8B) and microglia number (Fig. 8C) were reduced in WT treated with FA to an extent similar to that seen with mthfr −/− mutants. By comparisons, food intake (Fig. 8D) of first-time fed zebrafish at 5 dpf was affected by FA Dose (P < 0.0001) that was dependent on the Genotype as shown by a significant Genotype × Dose interaction effect (P < 0.0001). Specifically, while FA exposure did not affect the amount of food consumed by WT zebrafish at 5 dpf, mthfr −/− mutants exposed to FA consumed >50% more food than all other.
FA dose and mthfr genotype interact to affect metabolic outcomes. (A) Calculated mean gray area of viscera region at 5 dpf in WT and mthfr −/− mutants (mthfr −/−, HSC194) exposed to either 0 or 100 mm FA from 0 to 5 dpf and stained with ORO. (B) Relative metabolic rate at 5 dpf of WT and mthfr −/− exposed to either 0 or 100 mm FA from 0 to 5 dpf. (C) Number of NR+ cells at 4 dpf WT and mthfr −/− exposed to either 0 or 100 mm FA from 0 to 4 dpf. (D) Relative food intake, measured by CTCF, at 5 dpf in WT and mthfr −/− exposed to either 0 or 100 mm FA from 0 to 5 dpf. All values are mean ± SEM. Analyzed by two-way ANOVA with FA dose (0 versus 100 mm) and mthfr genotype (WT versus mthfr −/−) as main factors and a Dose × Genotype interaction term. Significant at P < 0.05. Values in bold indicate significant effects. Circles represent one individual zebrafish from batch 1 and squares represent zebrafish from batch 2.
As expected, FA was only detected in WT and mthfr −/− zebrafish exposed to 100 μm of FA at 5 dpf (Table 1). However, concentrations of methionine, SAM, SAH, SAM:SAH ratio, betaine and choline were unaffected by FA dose as indicated by a significant Genotype effect (P < 0.01) but not by either Dose (P > 0.05) or interaction effect. A significant Genotype × Dose interaction was observed on concentrations of cystathionine, with high FA exacerbating mthfr deficiency-induced increases.
Whole-body methyl metabolite concentrations in larvae exposed to 0 or 100 μm FA up to 5 dpf
| WT | mthfr −/− | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 μm FA | 100 μm FA | 0 μm FA | 100 μm FA | Two-way ANOVA P-value | |||||||||||
| Measure | Mean | ± | SEM | Mean | ± | SEM | Mean | ± | SEM | Mean | ± | SEM | D | G | D × G |
| 5 dpf | |||||||||||||||
| FA (nmol/g) | <LOD | 0.95 | ± | 0.33 | <LOD | 1.19 | ± | 0.40 | |||||||
| Methionine (nmol/g) | 1.00 | ± | 0.15 | 1.01 | ± | 0.09 | 0.70 | ± | 0.07 | 0.77 | ± | 0.07 | 0.68 | 0.03 | 0.79 |
| SAM (nmol/g) | 1.00 | ± | 0.26 | 1.54 | ± | 0.12 | 0.86 | ± | 0.26 | 1.08 | ± | 0.16 | 0.09 | 0.45 | 0.17 |
| SAH (nmol/g) | 1.00 | ± | 0.04 | 1.32 | ± | 0.15 | 4.56 | ± | 0.33 | 4.42 | ± | 0.43 | 0.78 | <0.0001 | 0.47 |
| SAM:SAH (nmol/g) | 1.00 | ± | 0.16 | 1.35 | ± | 0.12 | 0.26 | ± | 0.07 | 0.34 | ± | 0.07 | 0.06 | <0.0001 | 0.21 |
| Cystathionine (nmol/g) | 1.00a | ± | 0.19 | 1.09b | ± | 0.05 | 1.89a | ± | 0.07 | 2.73c | ± | 0.16 | 0.006 | <0.0001 | 0.02 |
| Choline (nmol/g) | 0.87 | ± | 0.18 | 0.84 | ± | 0.10 | 1.22 | ± | 0.09 | 1.08 | ± | 0.11 | 0.50 | 0.029 | 0.68 |
| Betaine (nmol/g) | 1.00 | ± | 0.13 | 1.01 | ± | 0.07 | 0.70 | ± | 0.04 | 0.73 | ± | 0.02 | 0.80 | 0.003 | 0.85 |
| WT | mthfr −/− | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 μm FA | 100 μm FA | 0 μm FA | 100 μm FA | Two-way ANOVA P-value | |||||||||||
| Measure | Mean | ± | SEM | Mean | ± | SEM | Mean | ± | SEM | Mean | ± | SEM | D | G | D × G |
| 5 dpf | |||||||||||||||
| FA (nmol/g) | <LOD | 0.95 | ± | 0.33 | <LOD | 1.19 | ± | 0.40 | |||||||
| Methionine (nmol/g) | 1.00 | ± | 0.15 | 1.01 | ± | 0.09 | 0.70 | ± | 0.07 | 0.77 | ± | 0.07 | 0.68 | 0.03 | 0.79 |
| SAM (nmol/g) | 1.00 | ± | 0.26 | 1.54 | ± | 0.12 | 0.86 | ± | 0.26 | 1.08 | ± | 0.16 | 0.09 | 0.45 | 0.17 |
| SAH (nmol/g) | 1.00 | ± | 0.04 | 1.32 | ± | 0.15 | 4.56 | ± | 0.33 | 4.42 | ± | 0.43 | 0.78 | <0.0001 | 0.47 |
| SAM:SAH (nmol/g) | 1.00 | ± | 0.16 | 1.35 | ± | 0.12 | 0.26 | ± | 0.07 | 0.34 | ± | 0.07 | 0.06 | <0.0001 | 0.21 |
| Cystathionine (nmol/g) | 1.00a | ± | 0.19 | 1.09b | ± | 0.05 | 1.89a | ± | 0.07 | 2.73c | ± | 0.16 | 0.006 | <0.0001 | 0.02 |
| Choline (nmol/g) | 0.87 | ± | 0.18 | 0.84 | ± | 0.10 | 1.22 | ± | 0.09 | 1.08 | ± | 0.11 | 0.50 | 0.029 | 0.68 |
| Betaine (nmol/g) | 1.00 | ± | 0.13 | 1.01 | ± | 0.07 | 0.70 | ± | 0.04 | 0.73 | ± | 0.02 | 0.80 | 0.003 | 0.85 |
Values are mean ± SEM, n = 3–4 biological replicates of pooled zebrafish/group. Each pool was composed of n = 50 larvae for folates and n = 30 larvae for methyl metabolites. No differences in pooled weight were observed. Analyzed by two-way ANOVA with FA Dose (0 versus 100 mm) and mthfr Genotype (WT versus mthfr −/−) as main factors and a Dose × Genotype (D × G) interaction term. Significant at P < 0.05. Superscript letters indicate significant differences by Tukey post-hoc analysis. Values in bold indicate significant effects. Abbreviations: mthfr −/−, mthfr −/− mutant zebrafish (HSC194); WT, wild-type zebrafish.
Whole-body methyl metabolite concentrations in larvae exposed to 0 or 100 μm FA up to 5 dpf
| WT | mthfr −/− | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 μm FA | 100 μm FA | 0 μm FA | 100 μm FA | Two-way ANOVA P-value | |||||||||||
| Measure | Mean | ± | SEM | Mean | ± | SEM | Mean | ± | SEM | Mean | ± | SEM | D | G | D × G |
| 5 dpf | |||||||||||||||
| FA (nmol/g) | <LOD | 0.95 | ± | 0.33 | <LOD | 1.19 | ± | 0.40 | |||||||
| Methionine (nmol/g) | 1.00 | ± | 0.15 | 1.01 | ± | 0.09 | 0.70 | ± | 0.07 | 0.77 | ± | 0.07 | 0.68 | 0.03 | 0.79 |
| SAM (nmol/g) | 1.00 | ± | 0.26 | 1.54 | ± | 0.12 | 0.86 | ± | 0.26 | 1.08 | ± | 0.16 | 0.09 | 0.45 | 0.17 |
| SAH (nmol/g) | 1.00 | ± | 0.04 | 1.32 | ± | 0.15 | 4.56 | ± | 0.33 | 4.42 | ± | 0.43 | 0.78 | <0.0001 | 0.47 |
| SAM:SAH (nmol/g) | 1.00 | ± | 0.16 | 1.35 | ± | 0.12 | 0.26 | ± | 0.07 | 0.34 | ± | 0.07 | 0.06 | <0.0001 | 0.21 |
| Cystathionine (nmol/g) | 1.00a | ± | 0.19 | 1.09b | ± | 0.05 | 1.89a | ± | 0.07 | 2.73c | ± | 0.16 | 0.006 | <0.0001 | 0.02 |
| Choline (nmol/g) | 0.87 | ± | 0.18 | 0.84 | ± | 0.10 | 1.22 | ± | 0.09 | 1.08 | ± | 0.11 | 0.50 | 0.029 | 0.68 |
| Betaine (nmol/g) | 1.00 | ± | 0.13 | 1.01 | ± | 0.07 | 0.70 | ± | 0.04 | 0.73 | ± | 0.02 | 0.80 | 0.003 | 0.85 |
| WT | mthfr −/− | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 μm FA | 100 μm FA | 0 μm FA | 100 μm FA | Two-way ANOVA P-value | |||||||||||
| Measure | Mean | ± | SEM | Mean | ± | SEM | Mean | ± | SEM | Mean | ± | SEM | D | G | D × G |
| 5 dpf | |||||||||||||||
| FA (nmol/g) | <LOD | 0.95 | ± | 0.33 | <LOD | 1.19 | ± | 0.40 | |||||||
| Methionine (nmol/g) | 1.00 | ± | 0.15 | 1.01 | ± | 0.09 | 0.70 | ± | 0.07 | 0.77 | ± | 0.07 | 0.68 | 0.03 | 0.79 |
| SAM (nmol/g) | 1.00 | ± | 0.26 | 1.54 | ± | 0.12 | 0.86 | ± | 0.26 | 1.08 | ± | 0.16 | 0.09 | 0.45 | 0.17 |
| SAH (nmol/g) | 1.00 | ± | 0.04 | 1.32 | ± | 0.15 | 4.56 | ± | 0.33 | 4.42 | ± | 0.43 | 0.78 | <0.0001 | 0.47 |
| SAM:SAH (nmol/g) | 1.00 | ± | 0.16 | 1.35 | ± | 0.12 | 0.26 | ± | 0.07 | 0.34 | ± | 0.07 | 0.06 | <0.0001 | 0.21 |
| Cystathionine (nmol/g) | 1.00a | ± | 0.19 | 1.09b | ± | 0.05 | 1.89a | ± | 0.07 | 2.73c | ± | 0.16 | 0.006 | <0.0001 | 0.02 |
| Choline (nmol/g) | 0.87 | ± | 0.18 | 0.84 | ± | 0.10 | 1.22 | ± | 0.09 | 1.08 | ± | 0.11 | 0.50 | 0.029 | 0.68 |
| Betaine (nmol/g) | 1.00 | ± | 0.13 | 1.01 | ± | 0.07 | 0.70 | ± | 0.04 | 0.73 | ± | 0.02 | 0.80 | 0.003 | 0.85 |
Values are mean ± SEM, n = 3–4 biological replicates of pooled zebrafish/group. Each pool was composed of n = 50 larvae for folates and n = 30 larvae for methyl metabolites. No differences in pooled weight were observed. Analyzed by two-way ANOVA with FA Dose (0 versus 100 mm) and mthfr Genotype (WT versus mthfr −/−) as main factors and a Dose × Genotype (D × G) interaction term. Significant at P < 0.05. Superscript letters indicate significant differences by Tukey post-hoc analysis. Values in bold indicate significant effects. Abbreviations: mthfr −/−, mthfr −/− mutant zebrafish (HSC194); WT, wild-type zebrafish.
Discussion
In this study, we generated and characterized a novel model of MTHFR deficiency using the zebrafish and evaluated the effects of mthfr deficiency and external FA supplementation during embryonic development on energy homeostasis and metabolism. In mthfr zebrafish, energy homeostasis was impaired as indicated by altered food intake, reduced metabolic rate and altered expression of genes involved in regulating food intake in the central nervous system. Microglia abundance was also reduced. Compared with controls, mthfr zebrafish had greater hepatic and aortic lipid deposition. Folates, methionine, betaine and SAM:SAH ratio were lower and total Hcy was elevated in mthfr zebrafish indicative of reduced methylation potential and consistent with metabolite changes seen in patients with MTHFR mutations. In total, these data support the utility of zebrafish as a model for studying MTHFR deficiency and its interplay with early life nutritional environment.
A novel finding of this study is the observed changes in energy homeostasis throughout development in mthfr −/−zebrafish up to 8 dpf. We found altered food intake and decreased metabolic rate in mthfr −/− zebrafish, which we hypothesized to be driven, in part, by disruption of central energy regulatory pathways. Central regulatory genes that have been previously shown to be targets of maternal folate inadequacies in rodents and/or postulated to be epigenetically modified in patients with MTHFR deficiency (33,34,43) were affected. Overall reductions in central genes involved in dopaminergic (i.e. drd2a, comta, th), GABAergic (i.e. gabra4) and glutamatergic (i.e. grin1a) signaling pathways, as well as those involved in neuronal survival and growth (i.e. bdnf) were observed in 5 dpf mthfr −/− zebrafish. While the effects of these genes are pleiotropic, they have key roles in regulating the neurotransmission efficiency of central homeostatic and hedonic systems that mediate food intake and energy expenditure (44,45). Downregulation of these signaling genes, as observed in mthfr −/− zebrafish, suggests that mthfr may act as a regulator of their transcription, and mthfr loss may repress related signaling pathways to result in the disease phenotype. Future studies are required to investigate these mechanisms.
mthfr KO in zebrafish also caused a reduction in microglia (likely via increased efferocytic clearance of apoptotic cells) (46) during development, which has important implications in energy balance. Loss of these cells may thus have contributed to widespread energy dysregulation. In zebrafish, microglia are derived from embryonic hematopoietic tissue, colonize the developing brain starting at 2.5 dpf and rapidly populate the midbrain where they are involved in clearance of apoptotic cells, neuronal activity, cellular plasticity and energy metabolism (47). Notably, microglia have been previously shown to be sensitive to changes in folate status as they require 5-MTHF for survival (48) and are directly inhibited by an excess of the neurotoxins SAH and Hcy (49), increases in which were identified in our mthfr −/− zebrafish. While the effects of folate fluctuations on microglia abundance during embryonic development in relation to energy metabolism have not been investigated in the murine model, one study reported that the elimination of microglia in the fetal mouse brain resulted in lower Pomc neurons and concomitant increases in body weight gain (50). Taken together, our findings support a programming effect of mthfr deficiency contributing to disrupted central energy balance regulatory pathways in zebrafish during early life.
Our results also show that mthfr deficiency promotes hepatic and aortic accumulation of lipids, and particularly, cholesterol. This is in line with preclinical findings using the Mthfr murine model (39,40,51) as well in humans wherein pathogenic variants in MTHFR correlate with obesity and cardiometabolic disease risk factors, including dyslipidemia (52,53). During early zebrafish development, the yolk-sac provides a rich lipid reserve to sustain growth up to the exogenous feeding phase (54). As such, lipid accumulation observed in mthfr −/− zebrafish during early development likely reflects the onset of atherogenesis in zebrafish (55), and indicates reduced endotrophic lipid consumption (56); findings that may also explain the reduced food intake observed in mthfr −/− zebrafish fed exogenously at this time.
The observed energy and metabolic dysregulation in mthfr zebrafish may reflect an imbalance in methyl substrates that are key in regulating lipid and energy homeostasis. Our mthfr zebrafish reproduced key clinical biochemical hallmarks of severe MTHFR deficiency including increased plasma Hcy and cystathionine and lower plasma methionine (38). This biochemical profile is positively correlated with aberrant lipid profiles and coronary artery disease (57). We also observed several downstream disruptions in the 1-carbon cycle. Increased SAH and reductions in the SAM:SAH ratio and betaine concentrations in mthfr zebrafish are in line with reduced methylation potential and increased utilization of the betaine-dependent methyl pathway upon functional loss of mthfr. This is consistent with findings from human, rodent and other zebrafish studies (58–61) and supports widespread perturbations in methyl metabolism. This altered methyl metabolism may cause dysregulation of the epigenome, and thus aberrant epigenetic modifications, such as via DNA methylation, that may underlie the observed changes in central regulatory gene expression in mthfr −/− zebrafish and early and later-life disease consequences of MTHFR deficiency.
An important finding of our study is the phenocopying effects of excess supplemental FA to mthfr deficiency in zebrafish. These results hold several important implications. First, while preclinical studies have alluded to the potential adverse effects of excess FA exposure in utero, there is limited understanding of the consequences throughout embryogenesis. Our study shows that excess FA may adversely affect cellular energy metabolism and microglia abundance in healthy developing embryos. Second, our results suggest a direct effect of UMFA on the development of adverse metabolic outcomes during embryogenesis, independent of the maternal contribution to folate status. Third, the adverse effect of UMFA may involve noncanonical folate pathways since FA exposure did not disrupt methyl metabolism. The involvement of UMFA on alternative signaling pathways has been recently hypothesized as an important mechanism of FA-induced pathogenesis, but has not been validated in vivo (25).
In this study, we also demonstrate that both mthfr crispants and germline mthfr −/− mutants show phenotypic and molecular convergence. As crispant analyses are performed directly in the mosaic F0 zebrafish, future experimentation with this model would enable for a high-throughput, rapid and cost-effective screening option to begin to elucidate nutrient–gene interactions and underlying mechanisms mediating mthfr-related pathologies. However, differences between models with respect to key methyl metabolites (e.g. choline, Hcy) highlight the importance of corroborating results in stable biallelic mthfr −/− germline zebrafish. Biochemical variability between mthfr crispant zebrafish and germline mutants may be explained in part by a maternal effect in the former, and/or the activation of genetic compensation, which was predicted to be reduced in mutants created in this study harboring a TSS deletion (62).
Together, the strength of our study lies in the utility of the mthfr zebrafish to address FA knowledge gaps and to advance the understanding of genetic mutation of mthfr and the interaction between genotype and the amount and form of supplemental folate during embryogenesis. In addition, the current results have potential application for regulatory decisions and public health advice. Potential adverse outcomes attributed to excess supplemental FA exposures were recently addressed (25) and the findings from our study lend further support. We show adverse effects of UMFA, which may indicate that the use of the natural form of folate, 5-MTHF, should be a priority in maternal supplements, especially during early development of the fetus. In addition, the overriding mthfr genotype effect upon exposure to FA, as observed on metabolic rate and methyl metabolites, for example, suggests that treatments with excess FA may have limited therapeutic benefit for mthfr-related pathologies, as recently postulated (63). Future dose–response experiments are warranted to elucidate the threshold of possible adverse effects and/or benefits on mthfr genotype.
Nevertheless, we also acknowledge that this study has limitations. A main weakness arises from the lack of follow-up on target gene expression within specific brain nuclei, as well as regulatory mechanisms to explain the observed phenotypic changes and pathways related to UMFA actions. Moreover, the lack of standardization of the folate and methyl contents of common zebrafish feeds (e.g. brine shrimp) and limited knowledge of nutrient requirements in zebrafish makes it difficult to perform dietary manipulation studies and infer the contribution of other methyl nutrients to the observed phenotypes. While the FA dose used in this study was informed by past literature (64,65), the standardization of zebrafish nutrient requirements is of high priority to support the use of the zebrafish model for nutrient depletion and exposure studies in the near future.
Both genetic mutation of mthfr and FA supplementation affect the development of energy and metabolic homeostasis in zebrafish. These preclinical findings support the translatability of the mthfr zebrafish as a preclinical model for folate research and public health applications.
Materials and Methods
Generation of mthfr-deficient zebrafish
In collaboration with the Zebrafish Genetics and Disease Models Core Facility at The Hospital for Sick Children (Toronto, Canada), loss of mthfr in zebrafish was generated using two CRISPR mutagenesis techniques. To begin assaying for the function of mthfr in vivo, a rapid transient KO approach reported to recapitulate germline loss-of-function phenotypes was employed (‘Crispants’, Fig. 1A) (66). Specifically, 1-cell stage AB WT embryos were injected with 400 pg of Cas9 protein alone (cas9, control, Alt-R® S.p. Cas9 Nuclease V3, IDT, Cat#1081058) or 400 pg of Cas9 protein and a set of 4gRNA at 250 pg each (herein referred to as 4gRNA mthfr crispants). 4gRNA targets were adapted from the published genome-scale lookup table for four-guide sets and targeted exons 3, 5, 6 and 9 of the zebrafish mthfr gene (Fig. 2A; Supplementary Material, Table S1). For functional validation and characterization of an mthfr germline transmitted mutation (mthfr-atgdel, herein referred to as mthfr −/− mutant zebrafish, HSC194, Fig. 1B), 1-cell stage WT embryos were co-injected with 150 pg of Cas9 mRNA and a set of two gRNAs at 100 pg each that were designed to target the 5′ UTR and TSS of the gene located in exon 2 (Fig. 2A). Deletion of the TSS has been previously shown to limit the genetic compensation response induced by the introduction of a premature stop codon by single gRNA injections (62). The gRNA target sites for generation of germline mutants were designed to contain a 20-base pair target sequence using the CRISPR design program Chopchop (http://chopchop.cbu.uib.no/) and are shown in Supplementary Material, Table S1. All gRNAs were synthesized by in vitro transcription as previously described (67). Cas9 mRNA was synthesized via in vitro transcription from a PCS2 Cas9 vector using the mMessage mMachine SP6 kit (Invitrogen). To evaluate gRNA cutting efficiency, high-resolution melt analysis (HRMA) was performed on a Roche Lightcycler 96 machine using the HRMA master mix kit (Roche) and custom designed primers using PRIMER-BLAST software (Supplementary Material, Table S2). To identify founders (F0s) with germline mutations, sexually mature F0s were outcrossed to WT zebrafish and genomic DNA was isolated from single F1 embryos, PCR amplified and sequence confirmed to determine heterozygous carriers (mthfr +/−). As mthfr is a maternally inherited transcript, F1 sequence confirmed that mthfr +/− mutants were in-crossed to generate F2 mthfr homozygous (−/−) mutants, which were confirmed by PCR amplification. PCR sequencing primers (Fig. 2A; Supplementary Material, Fig. S1; Supplementary Material, Table S2) were designed to amplify a 2308 bp product in WT zebrafish, a 250 bp product in heterozygous zebrafish and to result in no amplification in homozygous zebrafish. All mthfr disruptions were also validated at the transcript level by qRT-PCR using pooled samples of whole larvae (n = 20/pool × 3 pools/group) at 5 dpf (Supplementary Material, Table S3). Levels of 5-MTHF, which requires a functional mthfr transcript, were quantified by liquid chromatography tandem mass spectrometry (LC-MS/MS) as described below. Progenies from the F2 or F3 adult mthfr −/− in-cross (INX) were used for characterization in this study. Progenies generated from age-matched WT zebrafish maintained under identical conditions were used as controls. All mating pairs were between the age of 6 and 8 months.
Study design
A simplified schematic of the study design is depicted in Figure 1C. 4gRNA mthfr crispants and/or mthfr −/− mutant zebrafish and their respective controls (cas9 and WT) were exposed to either no FA (0 μm, control) or a supplemental dose of 100 μm FA in their water during their endogenous feeding phase (up to 5 dpf) and phenotypic and molecular analyses were conducted. For analyses during the exogenous feeding phase (>5 dpf), zebrafish larvae were either fasted or fed a high cholesterol diet (HCD) up to 15 dpf with or without fluorescent cholesteryl (CE)-BODIPY-C12. A power analysis was performed a priori to determine the sample size required for food intake (n = 20–24/batch), metabolic rate (n = 40–45/batch) and remaining analyses (n = 5–8/batch). All experiments were replicated at least twice or more unless indicated and methods were conducted as detailed below. Because of the small size of the zebrafish larvae, pooled samples were used for qRT-PCR and biochemical analyses. All analyses were conducted in a blinded manner.
Zebrafish husbandry
All adult zebrafish were maintained at 28.5°C under a 14:10-h light/dark cycle at the Zebrafish Facility at SickKids Peter Giligan Center for Research and Learning (Toronto, CA), and experiments were approved by the SickKids Animal Care Committee (AUP Protocol #56033). Adult zebrafish (>60 days) used for breeding were fed live brine shrimp (Artemia salina; containing ~90% moisture, 50% protein, 15% lipids, 15% carbohydrates, 3.5% fiber and 15–20% ash, % dry weight) twice daily (68,69). Of note, the folate content of brine shrimp has previously been reported to be ~0.0007 mg/g. Breeding pairs (2 female: 2 male) were set up with dividers following the last feed of the day. From 0 to 5 dpf, embryos were maintained in a 28.5°C incubator in embryo water (60 μg/ml Instant Ocean® sea salt in system water, pH = 7.7) containing 0.1% methylene blue or when indicated, maintained in 0.003% 1-phenyl 2-thiourea (PTU) in autoclaved system water to inhibit pigmentation. Water was changed daily to ensure sterile conditions. The 0–5 dpf early developmental period of the zebrafish represents the endogenous feeding phase wherein nutrients are obtained from their maternally loaded yolk-sac (54). Notably by 1 dpf, the zebrafish embryo had no detectable levels of synthetic FA as determined by LC-MS/MS (data not shown), limiting confounding effects of the maternal nutrient supply on outcome measures.
FA exposure
Embryos (n = 20 embryos/group) were collected 3 h after the onset of mating, staged according to Kimmel et al. (27) and exposed to FA (Sigma) at a dose of either 0 μm (0FA, control exposure) or 100 μm (100FA, high) up to 5 dpf. The 3-h time point was chosen as the zebrafish genome is globally re-methylated following fertilization at this time (70). FA solutions were prepared fresh daily in the dark using autoclaved system water. All petri dishes were sealed with parafilm, wrapped in aluminum foil and left in the incubator undisturbed for each consecutive 24-h period. Of note, the 100 μm dose of FA was chosen as it has been shown to impair the activity of the rate limiting enzyme for FA metabolism, dihydrofolate reductase in zebrafish, similar to humans (64). No teratogenicity has been reported in zebrafish exposed to this dose during development (65), which was confirmed by our group following a preliminary dose response experiment.
Intestinal food intake quantification
Intestinal food intake was measured in first-time fed 5 and 8 dpf zebrafish larvae raised in PTU according to the protocol by Wee et al. (71) with slight modifications. Paramecia were incubated with a lipophilic tracer (DiD’ solid, D-7757, Thermo Fisher Scientific, dissolved in ethanol) for 2 h with mild agitation. Zebrafish were habituated in small petri dishes (35 × 15 mm) for at least 2 h (n = 20 larvae/dish) and subsequently exposed to an abundance of labeled paramecia for 3 h in the incubator. Following feeding, larvae were thoroughly washed to remove excess paramecia, transferred to a clean petri dish with PTU water and anesthetized with 0.003% tricaine. Samples were mounted in 3% methylcellulose in autoclaved system water on a single cavity depression slide and imaged using a Zeiss Axio Zoom macroscope. ImageJ was used to calculate corrected total cell fluorescence (CTCF): CTCF = Integrated Density − (Area of selected cell × Mean fluorescence of background readings). Animals that did not feed (as indicated by no fluorescence) were not included in final analyses.
Metabolic rate
Metabolic rate assessment was performed according to a protocol by Williams and Renquist (72) wherein NADH2 production was measured as a proxy for energy expenditure. Briefly, 5 and 8 dpf zebrafish larvae were plated individually into a 96-well plate with 300 μL of assay solution containing alamarBlue™ Cell Viability Reagent (ThermoFisher Scientific Cat #: A50100). Blank wells containing only assay solution served as controls. Baseline fluorescence was measured using a fluorescent plate reader with excitation and emission wavelengths at 530 and 590 nm, respectively. The plate was wrapped in aluminum foil to avoid photobleaching and placed in a 28.5°C incubator overnight. Fluorescence was measured 24 h later. Change in fluorescence was calculated as follows: fluorescence (24 h) − fluorescence (baseline). The average change in fluorescence of the blank wells was subtracted from each well of interest. Lastly, the relative change in fluorescence was calculated as follows: Change in fluorescence (well of interest)/Average change in fluorescence (WT wells).
Central energy regulatory gene expression
To begin to explore the effects of mthfr disruption on central energy regulation, we performed qRT-PCR for select energy regulatory transcripts using pooled samples of dissected larval heads with eyes removed (n = 5 heads/pool × 5 pools/group). The targets chosen were adapted from our previous study in rats (33,34) and others using the Mthfr mouse model (35) whereby they have been shown to be sensitive to folate fluctuations. Targets included: the orexigenic npy and anorectic cart4; and genes involved in neurotransmission of energy regulating neurons: th, drd2a, comta, gabra3, gabra4, grin1a and bdnf. Total RNA was extracted using Trizol reagent and quantified using a NanoDrop™ 2000 spectrophotometer. cDNA synthesis was conducted using the High Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems Inc., catalog #: 4368813). qRT-PCR was performed using SYBR Green (ThermoFisher, catalog #: A25742) on the Roche Lightcycler 96 machine. mRNA levels were normalized to levels of b-actin as the housekeeping gene selected based on lowest variation as the control. Primers were designed using PRIMER-BLAST software as listed in Supplementary Material, Table S3. The 2−∆∆CT method was used for all analyses (73).
In vivo myeloid cell analyses and tail wounding assay
Immune dysfunction has been suggested to be an underlying pathophysiology for diseases associated with pathogenic variants in MTHFR (36). As such, we explored whether mthfr disruption in zebrafish affected myeloid regulatory cells during embryonic development. 4gRNA + Cas9 protein or Cas9-only was co-injected into 1-cell stage zebrafish embryos generated from the INX of a double transgenic line expressing green fluorescent protein (GFP) under the control of a myeloid-specific peroxidase promoter (mpx) to highlight neutrophils [Tg(mpx:GFP)i114] (74), and red fluorescent protein (mCherry) under the macrophage expressed gene 1 promoter (mpeg1) to visualize macrophages [Tg(mpeg1:mCherry)gl23] (75). To test the effects of neutrophil chemotaxis upon tissue injury, a tail fin wounding assays was performed as previously described (76). Z-stacks were captured through the entire dorsal head region or the left–right axis of the zebrafish trunk using a Nikon A1 confocal microscope.
In addition, NR vital staining was performed for visualization of microglia at 3, 4, and 6 dpf using freshly prepared NR solution with PTU autoclaved system water (5 μg/mL, Sigma, catalog #: N-4638) as previously described (46). For each experiment, equal larvae per petri dish containing 25 mL of NR solution were incubated in the dark at 28.5°C for 2 h. Following incubation, larvae were rinsed in PTU autoclaved system water for 20 min, anesthetized with tricaine (0.16 mg/ml) and mounted in 3% methylcellulose on a single cavity depression slide for analysis using a Zeiss Axio Zoom. ImageJ was used to count the number of NR+ cells.
AO staining
As an index of apoptosis, the vital dye AO (Invitrogen™, catalog #: A1301) dye was used (77). Briefly, larvae were raised to 3 dpf in PTU autoclaved system water. At 3 dpf, larvae (n = 10/group) were stained with 2.5 μg/mL of AO in PTU system water for 20 min, thoroughly washed and imaged in the GFP channel using the Zeiss Axio Zoom. ImageJ was used to count the number of AO+ cells.
In vivo cholesterol tracking
In vivo analysis of cholesterol metabolism was conducted in 4gRNA mthfr crispants using a CE-BODIPY-C12® fluorophore (ThermoFisher, catalog #: D3822) as previously described (78–80). Briefly, embryos (n = 13/group) were transferred to a 1.6 L tank at 5 dpf and fed equal quantities (equating to 0.3 g of food/larvae/day) of the AP-100 larvae zebrafish diet (crude protein min 50.0%, crude fat min 12.0%, crude fiber max 2.5%, moisture max 10.0%, ash max 15.0%, phosphorus min 1.3%, Ziegler, 50 μm size) with added 4% w/w cholesterol (HCD) up to 13 dpf. Larvae were then fed the HCD supplemented with 1 μg/g of CE-BODIPY-C12® until 15 dpf. The HCD helped stimulate native enzymes required for cholesterol metabolism to enable the incorporation of the CE-BODIPY-C12® fluorophore. Live 15 dpf zebrafish were then mounted in 1.5% low-melting agarose containing tricaine on a 3 cm glass-bottom petri dish and z-stacks were taken using a Nikon Eclipse Ti confocal microscope using a 10× objective lens. Maximum Z-stack projections of equal size were analyzed on ImageJ to measure total liver area (mm2) and fluorescent intensities.
Neutral lipid analyses
Total neutral lipid accumulation was investigated via ORO staining starting at 5 dpf, when yolk-derived lipids are being absorbed (54), followed by 8 dpf, when yolk-sac lipid absorption is predicted to be completed (81) and then at 15 dpf (n = 5 larvae/group) following a HCD challenge. Briefly, 5 and 8 dpf larvae raised in PTU were fixed in 4% paraformaldehyde overnight at 4°C and washed 3x for 5 min with 1X-PBS-0.5% Tween (1X-PBST). Larvae were then stained with 300 μL of filtered 0.5% ORO dissolved in 100% isopropanol and 200 μL of distilled water for 15 min. Larvae were washed with 1X-PBST 3x for 5 min, 2x in 60% isopropanol for 5 min and 1x in 1X-PBST for 5 min. For 15pf larvae, bleach was used to remove pigmentation. A bleaching solution was made using equal volumes of 3% H2O2 and 2% KOH for a final concentration of 1.5% H2O2 and 1% KOH. Upon removal of stain, 1 mL of freshly prepared bleach solution was added to samples in microcentrifuge tubes with lids opened at room temperature for 10 min, followed by 3x washes in 1X-PBST. Samples were transferred into 80% glycerol and mounted onto a single cavity depression slide for analysis using Zeiss Axio Zoom. The mean gray area of a region of interest was quantified on ImageJ.
Folates and methyl metabolites
Concentrations of folate derivatives and methyl metabolites were measured at 5 dpf when yolk-sac nutrient stores have begun depletion. For folates, liquid chromatography coupled with electrospray positive ionization tandem mass spectrometry (LC-ESI-MS/MS) was performed as previously described with minor modifications (82). Briefly, equal pools of zebrafish larvae (n = 50 per pool × 4 pools/group) were sonicated in 250 μL of extraction buffer (HEPES buffer pH 8) to which charcoal treated rat plasma was added as a source of pteroylpolyγglutamate hydrolase for the deconjugation of the polyglutamate forms to monoglutamate forms of folates. Following centrifugation, filtering and addition of 13C5-methyl-THF internal standard (IS), clear extracts were injected in the UHPLC-MS/MS system consisting of a Waters Acquity UPLC coupled to a Xevo-TQS triple quadrupole mass spectrometer equipped with an electrospray ionization probe (Waters Corporation, Milford, MA, USA). Quantification of folate forms was achieved using the peak area ratio (area of metabolite/area of IS). Concentrations of methionine, s-adenosylmethionine (SAM), SAH, Hcy, cystathionine, choline and betaine in whole-zebrafish larvae were determined by LC-ESI-MS/MS. Pooled zebrafish larvae (n = 30/pool × 4 pools/group) were sonicated with 100 μL of methanol, spun for 10 min and 40 μL of the supernatant was dried down under nitrogen for 5 min. Samples were then reconstituted in 40 μL of mobile phase A and diluted 1:5 with IS. Extracts were loaded into the LC-MS/MS system (QTrap 5500, Sciex, Framingham, MA) and metabolites quantitated using Analyst 6.0 (Sciex, Framingham, MA). Data points that were below the lower limit of detection were not included in statistical analysis.
Statistical analysis
All statistical analyses were conducted using SAS version 9.4. A Student’s t-test was used to determine differences between control and mthfr disrupted zebrafish with stratification based on method of mutagenesis. When experiments were conducted in replicates, a one-way ANCOVA using the PROC GLIMMIX procedure was used to control for variance between days of sample collections. To compare the effects of mthfr genotype and FA dose, a two-way ANOVA was used with Genotype and Dose as main effects and a Genotype × Dose interaction effect. Significant interaction effects were followed by Tukey’s post-hoc test for pairwise comparisons. All significance is reported at P < 0.05 and all values are expressed as mean ± SEM. All sample data points were included in each analysis unless stated otherwise.
Conflicts of Interest statement: None declared.
Funding
The Canadian Institutes of Health Research, Institute of Nutrition, Metabolism and Diabetes (Grant No. MOP-130286); the Natural Sciences and Engineering Research Council of Canada (Grant No. RGPIN-2016–06639).
Data Availability
All data supporting the findings of this study are available within the article and its supplementary materials. The data that support the findings of this study are also available upon request from the corresponding author, GHA.
Authors’ Contributions
R.S., E.P., R.H., R.K., R.R.N. and G.H.A. contributed to study design; R.S., E.P. and X.C. performed zebrafish experiments; R.S., E.P., R.H. and E.K. analyzed data; P.A., B.W. and T.B. were responsible for LCMS analyses; J.J.D. and G.H.A. provided essential reagents and materials; R.S., E.P., R.H., J.J.D. and G.H.A. contributed to manuscript preparation; R.S., E.P., J.J.D. and G.H.A. had primary responsibility for final content. All authors read and approved the final manuscript.
References
Author notes
Rebecca Simonian and Emanuela Pannia Co-first author.
