Materials were obtained as follows: [U-13C]glucose (Cambridge Isotopes, CLM-1396), [U-13C]glutamine (Cambridge Isotopes, CLM-1822), C57BL/6J (UTSW Mouse Breeding Core or Jackson Labs) and Lipt1N44S knock-in mice (developed in-house)5.
Subject information and clinical data
The LIPT1-deficient individual who provided clinical data in Fig. 4f, Extended Data Fig. 6e, f was described previously5. This patient was enrolled in a prospective, non-randomized, non-blinded observational study whose overarching goal is to discover new metabolic disease-associated genes in patients of any age, and to characterize the metabolic phenotype in these patients (NCT02650622). The study was approved by the Institutional Review Board (IRB) at University of Texas Southwestern Medical Center (UTSW), and written informed consent was obtained from the patient’s parents. Patients and family members eligible for the study are identified at UTSW, its affiliated hospitals, and other collaborating hospitals. After enrollment, study subjects provide blood for metabolomics and genomics, and a research-based integrated analysis of the data allows potentially pathogenic genomic variants to be prioritized for functional analysis in the laboratory. The study is purely observational in that no therapeutic interventions are proposed, although patients are followed longitudinally to understand each disease’s natural history and the effects of therapies instituted as a part of routine clinical care. A total enrollment of over 1,500 patients is planned with the intention of representing many rare conditions within the cohort.
Reference datasets and data processing
Data for fetal tissues during midgestation are available from the ENCODE21,35,40 project Mouse Development Matrix (https://www.encodeproject.org/mouse-development-matrix). We downloaded the tsv files from the polyA plus RNAseq assay with the following identifiers: ENCFF262TPS (E11.5 liver -1), ENCFF414APX (E11.5 liver-2), ENCFF173NFQ (E12.5 liver-1), ENCFF144DHB (E12.5 liver-2), ENCFF971KKK (E13.5 liver-1), ENCFF042DVY (E13.5 liver-2), ENCFF770SOB (E10.5 heart-1), ENCFF351QKG (E10.5 heart-2), ENCFF159DWP (E11.5 heart-1), ENCFF168UJM (E11.5 heart-2), ENCFF484QWQ (E12.5 heart-1), ENCFF329HOZ (E12.5 heart-2), ENCFF148BEQ (E13.5 heart-1), ENCFF836QQS (E13.5 heart-2), ENCFF145PTV(E10.5 forebrain-1), ENCFF476ADM (E10.5 forebrain-2), ENCFF606UHO (E11.5 forebrain-1), ENCFF434CSI (E11.5 forebrain-2), ENCFF928MQD (E12.5 forebrain-1), ENCFF046RSQ (E12.5 forebrain-2), ENCFF960KJV (E13.5 forebrain-1), ENCFF356CTG (E13.5 forebrain-2). Placenta RNA transcript abundance was obtained from Gene Expression Omnibus (GEO) accession code GSE100053. Expression data were filtered based on known metabolic genes37,38,39 and human–mouse gene mapping was based on the HomoloGene database (https://www.ncbi.nlm.nih.gov/homologene).
Placental gene-expression data were obtained from the GEO repository (https://www.ncbi.nlm.nih.gov/gds) using the GEOquery package36 (https://doi.org/10.18129/B9.bioc.GEOquery) v2.62.1 from BioConductor release (3.14) (https://www.bioconductor.org/). Data were filtered based on known metabolic genes37,38,39 and sorted by Kyoto Encyclopedia of Genes and Genomes pathway annotation in the metaboAnalyst_KEGG R package (https://github.com/xia-lab/MetaboAnalystR). Human–mouse gene mapping was based on the HomoloGene database (https://www.ncbi.nlm.nih.gov/homologene).
All procedures were approved by the UT Southwestern Animal Care and Use Committee (IACUC) in accordance with The Guide for the Care and Use of Laboratory Animals. All mice were housed in a pathogen free environment (temperature 20–26 °C, humidity 30–70%) with a 12 h:12 h light:dark cycle and fed chow diet (Teklad 2916) ad libitum. Healthy 8–15 week old, naïve pregnant females were set up for mating between 05:00 and 07:00 with proven studs of the appropriate genotype. The following morning, females displaying vaginal plugs were identified as pregnant and moved to a new cage until the indicated gestational day.
All sample collection took place between 09:00 and 11:00 with no prior fasting of the pregnant dams. Mice were initially anaesthetized using isoflurane and samples were dissected in cold sodium chloride irrigating solution (Baxter) and snap frozen in liquid nitrogen. Whole embryos and placentas were homogenized manually with a rubber dounce homogenizer in ice-cold acetonitrile:water (80:20). Samples were flash frozen 3 times in liquid nitrogen and then centrifuged at 16,000g for 10 min at 4 °C. Supernatants were subject to BCA analysis and normalized to 70 μg ml−1 and placed in LC–MS vials. Metabolite analysis used a Vanquish UHPLC coupled to a Thermo Scientific QExactive HF-X hybrid quadrupole orbitrap high-resolution mass spectrometer (HRMS) as performed previously31. Pooled samples were generated from an equal mixture of all individual samples and analysed using individual positive- and negative-polarity spectrometry ddHRMS/MS acquisition methods for high-confidence metabolite ID. Metabolite identities were confirmed in three ways: (1) precursor ion m/z was matched within 5 ppm of theoretical mass predicted by the chemical formula; (2) fragment ion spectra were matched within a 5 ppm tolerance to known metabolite fragments; and (3) the retention time of metabolites was within 5% of the retention time of a purified standard run with the same chromatographic method. LC-MS/MS data were collected using SCIEX Analyst v1.6.3 and Thermo Scientific XCalibur 4.1.50 and data analysed using SCIEX Multiquant v2.1.1, and Thermo Scientific Trace Finder v5.1. Relative metabolite abundance was determined by integrating the chromatographic peak area of the precursor ion searched within a 5 ppm tolerance and then normalized to total ion count (TIC). Statistical analysis for generation of PCA plots, heatmaps, differential abundances and MSOA were performed using MetaboAnalyst 5.0 (https://www.metaboanalyst.ca). Data were log-transformed and auto-scaled prior to the analysis. Additional heatmaps (Fig. 1e, Extended Data Fig. 2e) were generated using GraphPad Prism 9.0.1. For 13C studies, observed distributions of mass isotopologues were corrected for natural isotope abundances using a customized R script, which can be found at the GitHub repository (https://github.com/wencgu/nac). The script was written by adapting the AccuCor algorithm v0.2.432.
Pregnant mouse infusions
All infusions took place between 09:00 and 11:00 with no prior fasting of the pregnant dams. Mice were initially anaesthetized using ketamine and xylazine (120 mg kg−1 and 16 mg kg−1, respectively, intraperitoneally) and maintained under anaesthesia using subsequent doses of ketamine (20 mg kg−1, intraperitoneally) as needed. Catheters (25-gauge) were inserted into the tail vein and isotope infusions began immediately after a retro-orbital blood draw to mark time zero. In the glucose infusions, the total dose was 2.48 g kg−1 dissolved in 750 μl normal saline and administered with a bolus of 62.5 μl min−1 for 1 min followed by an infusion rate of 2.5 μl min−1 for 3–4 h. Retro-orbital blood draws were taken throughout the infusion to monitor tracer enrichment in maternal blood. Glutamine infusions used a total dose of 1.73 g kg−1 dissolved in 1,500 μl normal saline administered as a bolus of 147 μl min−1 for 1 min followed by an infusion rate of 3 μl min−1 for 5 h. Mice were euthanized at the end of the infusion, then the uterus was removed and placentas and embryos dissected in cold sodium chloride irrigating solution and frozen in liquid nitrogen. Care was taken during infusions not to increase nutrient concentrations over pre-infusion levels.
Serial caesarian-section surgery
For serial caesarian sections, the infusion parameters were the same as described above with the following alterations: (1) Serial caesarian-section infusions did not include a bolus; (2) the infusion rate was increased to 5 μl min−1 in order to obtain sufficient labelling. Although the patterns of data for serial caesarian-sections matched what we observed in the 4 h infusions, the overall labelling was somewhat lower and for this reason we did not compare serial caesarian-section data to data from longer infusions. After cannulation of the tail vein and retro-orbital blood draw for time zero, the lower abdomen of the pregnant dam was opened with a small incision. The uterus was removed from the peritoneal cavity and the conceptus nearest to one of the ovaries was dissected away from the uterus and further dissected into placenta and embryo in cold sodium chloride irrigating solution and then frozen in liquid nitrogen. The peritoneal cavity was flushed with sodium chloride irrigating solution, covered with gauze, and periodically rinsed with irrigating solution throughout the remainder of the surgery. The infusion was initiated and a single conceptus was dissected in a similar manner at the indicated time points until all embryos had been dissected or the 3 h time point was reached.
Gas chromatography mass spectrometry (GCMS)
Gas chromatography–mass spectrometry (GCMS) was used to identify glucose, pyruvate, lactate, citrate, succinate, malate and aspartate. These metabolites were also identified using liquid chromatography–mass spectrometry (LC–MS) and enrichment values were similar. Blood samples obtained during the infusion were chilled on ice for 5–10 min and then flash frozen in liquid nitrogen. Aliquots of 10–20 μl were added to 80:20 acetonitrile:water for extraction. Frozen tissues (whole embryo and whole placenta) were added to 80:20 acetonitrile:water and extracted to analyse 13C enrichment. Samples were manually disrupted using a rubber dounce homogenizer, subjected to three freeze–thaw cycles, then centrifuged at 16,000g for 15 min to precipitate macromolecules. For GCMS, 1 μl D27-myristic acid was added as an internal control, supernatants were evaporated, then re-suspended in 30 μl anhydrous pyridine with 10 mg ml−1 methoxyamine and incubated at room temperature overnight. The following morning, the samples were incubated at 70 °C for 10–15 min and then centrifuged at 16,000g for 10 min. The supernatant was transferred to a pre-prepared GC/MS autoinjector vial containing 70 μl N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide (MTBSTFA) derivatization reagent. The samples were incubated at 70 °C for 1 h after which aliquots of 1 μl were injected for analysis. Samples were analysed using either an Agilent 6890 or 7890 gas chromatograph coupled to an Agilent 5973N or 5975C Mass Selective Detector, respectively. GC–MS data were collected and analysed using Agilent ChemStation E02.02.1431. The observed distributions of mass isotopologues were corrected for natural isotope abundances using a customized R script, which can be found at the GitHub repository (https://github.com/wencgu/nac). The script was written by adapting the AccuCor algorithm v0.2.432.
Total RNA was extracted from placental tissue using TRIzol Reagent (Thermo Fisher Scientific cat. no. 15596026). RNA (3,250 ng) was used as a template for a 70 μl cDNA synthesis reaction using TaqMan Reverse Transcription Reagents (Thermo Scientific cat. no. N8080234) according to the manufacturer’s instructions. cDNA was diluted 1:1 in nuclease-free water and plated at a final volume of 4 μl in a 384-well plate. Primers for placental markers were as described33 and diluted to a final concentration of 2.5 µM. Primers were mixed with iTaq Universal SYBR Supermix (Bio-Rad cat. no. 1725121) and plated at a volume of 6 μl for a total reaction volume of 10 μl. Plates were run in a Bio-Rad CFX384 Touch Real-Time PCR Detection machine using the following protocol: (1) polymerase activation: 95 °C hold for 30 min; (2) PCR phase, 40 cycles: 95 °C hold for 5 s, 60 °C hold for 30 s; (3) melt curve, instrument default settings. Relative fold induction was computed using the ∆∆CT method, as described34.
Embryo RNA sequencing data were downloaded from the ENCODE Mouse Development Matrix35 (https://www.encodeproject.org/). PolyA plus RNA-seq data were obtained for fetal heart, forebrain and liver from GD10.5-GD12.5 (not all days are available for liver). Placenta RNA transcript abundance was obtained from GEO accession code GSE100053 using the GEOquery package36 (https://doi.org/10.18129/B9.bioc.GEOquery) v2.62.1 from BioConductor release (3.14) (https://www.bioconductor.org/). Data were filtered based on known metabolic genes37,38,39 and sorted by Kyoto Encyclopedia of Genes and Genomes pathway annotation in the metaboAnalyst_KEGG R package (https://github.com/xia-lab/MetaboAnalystR). Human–mouse gene mapping was based on the HomoloGene database (https://www.ncbi.nlm.nih.gov/homologene).
Whole embryos were collected from GD10.5 pregant mice into 1× PBS and mechanically disrupted using disposable pestles (VWR) and then filtered through a 40-µM cell strainer to remove clumps. Antibody staining was performed for 20 min on ice, followed by washing with HBSS (Invitrogen) and centrifugation at 200g for 5 min. Cells were stained with directly conjugated antibodies against mouse CD71 (FITC-R188.8.131.52 Biolegend, 1:100), mouse Ter119 (APC-TER-119 TONBO, 1:100), mouse CD41 (PE/Cy7-MWReg30 Biolegend, 1:100) and mouse CD117 (cKIT-APC-eFlour 780-Invitrogen, 1:100). All cells were gated for forward and side scatter and gated for live cells based on DAPI (1 μg ml−1; Sigma, eFlour-450A). Erythrocytes were cells that were negative for CD117 (c-KIT), and positive for CD71 and Ter119. Myeloid–erythroid progenitors were negative for CD71 and TER119 and positive for CD41 and CD117 (c-KIT). Cells were examined on an LSRFortessa cell analyser (Becton Dickinson) and figures were generated using BD FACSDiva 8.0 and FlowJo v10.
Whole-mount immunofluorescent staining
Pregnant females at the desired developmental stage were euthanized by carbon dioxide asphyxiation and the uterus and extra-embryonic tissues were removed. Yolk sacs were used for genotyping and somites were counted. Embryos were fixed in 4% paraformaldehyde for 1 h at 25 °C or 4 °C overnight. Fixed embryos were washed at least 3 times with 1× PBS and dehydrated through a series of methanol or ethanol (25%, 50%, 75% and 100%, two times), permeabilized using 1% Triton X-100 (Fisher Bioreagents, cat. no. BP151-100) in PBS for 1.5–2 h at 25 °C, then blocked using CAS Block (Life Technologies, cat. no. 008120) for 2 h. Embryos were incubated in primary antibodies diluted in CAS Block overnight at 4 °C: Rat-anti-PECAM1 (1:100, BD, Biosciences, cat. no. 553370), Rat-anti-endomucin (1:100, Santa Cruz, sc-65495) and Rabbit-anti-connexin 40 (1:100, Alpha Diagnostics International, cat. no. CX-40A). Embryos were washed with 1× PBS then incubated with secondary antibodies diluted in CAS Block at 1:250 overnight at 4 °C: donkey-anti-rat 488 (Invitrogen, cat. no. A21208), donkey-anti-rabbit 555 (Invitrogen, cat. no. A31572). Embryos were washed in 1× PBS, then dehydrated to 100% methanol through a methanol series (25%, 50%, 75%, 100% two times, 10 min each), cleared in a 1:2 benzyl alcohol:benzyl benzoate (BABB) solution, and mounted in BABB in 5 mm Thick Microscopy slides (Chang Biosciences, Rb167104D_1) and cover slipped. Images were obtained using a LSM700 Ziess confocal microscope with the Carl Zeiss ZEN 2011 software. If images of the dissected heart were desired, whole embryos were rehydrated through a methanol series into PBS, hearts were dissected and placed in a 1.5 mm 2-well concavity slide (Electron Microscopy Sciences, cat. no. 71878-03) containing PBS. Whole-heart images were obtained using a Ziess Images M2 with an Axiocam 506 mono camera attached with the Carl Zeiss ZEN 2011 software. For sectioned samples, paraffin embedded samples were transverse sectioned at 5 μm and stained with haematoxylin and eosin.
During flow cytometry, isotope tracing, metabolomics, quantitative PCR, tissue weights, somite counts and histology experiments, the data were analysed in a manner blinded to sample genotype. A.S. collected the samples and then passed them to A. Tasdogan. for flow cytometry, or to I.M.-M. and M.A.C. for histology and immunofluorescence, and A. Tarangelo. for quantitative PCR. A.S. processed samples for mass spectrometry and analysed data. After the patterns had been analysed in each of these experiments, D. Dumesnil. provided the genotype information so results could be interpreted. For experiments in wild-type mice, no blinding was performed on placentas versus embryos because A.S. performed these experiments and analysed the data. For gene-expression studies from publicly available datasets, no blinding was performed.
Mice were allocated to experiments randomly and samples were processed in an arbitrary order, but formal randomization techniques were not used. Samples sizes were not pre-determined based on statistical power calculations but were based on our experience with these assays. For most experiments, the minimum number of mice was 3, with some exceptions where the embryo/placenta numbers were n ≥ 10. No data were excluded; however, sometimes the small sample size was below the threshold for metabolomic analysis. In those instances, data that could be obtained from maternal blood or other tissues were used. These samples were not used during direct comparisons of embryo relative to its own placenta if one of the samples was absent.
Prior to analysing the statistical significance of differences among groups, we tested whether data were normally distributed and whether variance was similar among groups. To test for normality, we performed the Shapiro–Wilk tests when 3 ≤ n < 20 or D’Agostino omnibus tests when n ≥ 20. To test whether variability significantly differed among groups we performed F-tests (for experiments with two groups) or Levene’s median tests (for experiments with more than two groups). When the data significantly deviated from normality or variability significantly differed among conditions, we log2-transformed the data and tested again for normality and variability. If the transformed data no longer significantly deviated from normality and equal variability, we performed parametric tests on the transformed data. If log2-transformation was not possible or the transformed data still significantly deviated from normality or equal variability, we performed non-parametric tests on the non-transformed data.
When data or log2-transformed data were normal and equally variable, statistical analyses were performed using Student’s t-tests or paired t-tests (when there were two groups), one-way ANOVAs or repeated measures one-way ANOVAs (when there were more than two groups), two-way repeated measures ANOVAs (when there were two or more groups with multiple metabolites or time points), or mixed effects models (when there were missing values but the data otherwise met the assumptions for a one-way or two-way repeated measures ANOVA). When the data or log2-transformed data were normal but unequally variable, statistical analyses were performed using Welch’s t-tests (when there were two groups) or Welch’s one-way ANOVAs followed by the Dunnett’s T3 tests for multiple-comparisons adjustment (when there were more than two groups). When the data and log2-transformed data were abnormal or unequally variable, statistical analysis was performed using Mann–Whitney or Wilcoxon matched pairs signed rank tests (when there were two groups) or Kruskal–Wallis tests (when there were more than two groups). P-values from multiple comparisons were adjusted using Tukey’s (when there were more than two groups and all of the comparisons were of interest) or Sidak’s method (when there were more than two groups and planned comparisons) after ANOVAs or mixed effects models, or Dunn’s method after Kruskal–Wallis tests. Holm–Sidak’s method was used to adjust comparisons involving multiple metabolites between two conditions. A linear regression or nonlinear curve fitting method, plateau followed by one-phase association, was used to fit the time series data and the extra sum-of-squares F-test was used to assess if there was difference between two fitted lines/curves. Multiple line/curve fitting P-values were adjusted using the Holm–Sidak method. Statistical tests were performed using GraphPad Prism V9.0.1 or R 4.0.2.
Further information on research design is available in the Nature Research Reporting Summary linked to this paper.