Nuclear Magnetic Resonance Metabolomic Analysis of Spent Human Embryo Culture Media: Method Validation and Technical Considerations

Summary

Here, we present a protocol for collecting, preparing, and analyzing spent human embryo culture medium using nuclear magnetic resonance (NMR) spectroscopy, enabling reproducible detection of low-molecular-weight metabolites from microliter-scale samples under routine clinical culture conditions.

Abstract

Spent human embryo culture media contains low-molecular-weight metabolites that can provide a non-invasive readout of preimplantation embryo physiology. However, nuclear magnetic resonance (NMR)-based metabolomic analysis of these samples is technically challenging because routine embryo culture conditions involve microliter volumes, protein supplementation, and paraffin oil overlays that generate background signals and compromise spectral quality. Here, we present an optimized NMR metabolomics workflow for spent embryo culture media incorporating controlled droplet collection to minimize oil contamination, fluorinated ethylene propylene (FEP) tube liners to chemically isolate the sample from the external reference (TMSP) while accommodating microliter volumes, and acquisition using a Carr-Purcell-Meiboom-Gill pulse sequence to suppress macromolecular background signals. The protocol enables consistent detection and quantification of low-molecular-weight metabolites from individually cultured human embryos and supports reproducible analysis under routine laboratory conditions. Due to numerous technical challenges, the method must be properly validated before it can be used for research or clinical purposes.

Introduction

During preimplantation development in vitro., embryos continuously consume and release low-molecular-weight metabolites into the surrounding medium. The composition of spent embryo culture medium (SECM), therefore, reflects embryo metabolic activity and provides a non-invasive functional readout of embryo physiology and developmental competence^1,2. Analysis of this metabolic footprint may complement conventional assessment methods and improve embryo selection strategies in assisted reproductive technology^3,4. In particular, alterations in amino acid and carbohydrate turnover have been linked to embryo developmental potential and viability^5,6,7.

Nuclear magnetic resonance (NMR) spectroscopy is widely used for untargeted metabolite profiling of diverse biological samples, including spent embryo culture media^8,9,10. However, analysis of these samples remains analytically challenging¹¹. In routine clinical practice, embryos are cultured in defined microenvironments designed to mimic physiological conditions. Culture media are supplemented with proteins, typically albumin, to stabilize osmotic pressure, bind toxic compounds, and provide carrier molecules for lipids and growth factors that support normal development¹². To prevent evaporation and maintain constant solute concentrations, culture droplets are overlaid with sterile paraffin oil¹³.

Lipids originating from overlay oil and protein components generate broad background signals that obscure narrow resonances from low-molecular-weight metabolites such as glucose, pyruvate, and amino acids. Furthermore, analyzing microliter sample volumes in standard 5 mm NMR tubes requires significant dilution to reach the active volume of the RF coil. This reduction in analyte concentration severely diminishes the signal-to-noise ratio (S/N), masking low-abundance metabolites and hindering reproducible quantification. Although numerous studies have applied NMR-based metabolomics to spent human embryo culture media, the lack of standardized and reproducible collection and preparation protocols has limited reproducibility and inter-study comparability^8,9,11.

Here, we present an NMR-based metabolomics workflow (Figure 1) designed to overcome technical challenges associated with sampling and analyzing human embryo culture media. The approach combines controlled droplet collection to reduce oil carryover, microliter-volume acquisition using fluorinated ethylene propylene (FEP) liners to avoid dilution artifacts, and macromolecular signal attenuation using a Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence.

Protocol

This study was conducted between January and April 2023 at the Department of Reproductive Medicine and Gynaecological Endocrinology, University Medical Centre Maribor, Slovenia. NMR analysis was performed at the National Institute of Chemistry in Ljubljana, Slovenia. Written informed consent was obtained from all participants. It received approval from both the Institutional Ethics Committee (No: UKC-MB-KME-1/21) and the National Medical Ethics Committee (No: 0120-37/2021/15).

1. Study design

NOTE: The study analyzed human spent embryo culture media using NMR-based metabolomics.

1. Collect samples after extended embryo culture following intracytoplasmic sperm injection (ICSI). Collect an equal number of unspent incubated control samples.

2. Sample collection

1. Collect spent embryo culture media (SECM) samples following embryo incubation, as illustrated in Figure 2.
2. After removal of the embryo from the culture droplet, carefully aspirate excess overlay oil from the surface of the culture dish without disturbing the medium (Figure 2, step 1).
3. Using a sterile pipette, aspirate approximately 35 µL of culture medium from the original droplet, leaving approximately 5 µL behind to avoid drawing in residual surface oil. Transfer this volume into a sterile Petri dish to form a new droplet (Figure 2, step 2).
4. Under a stereomicroscope, visually inspect the newly formed droplet. Carefully aspirate approximately 30 µL from this droplet, again leaving approximately 5 µL behind to minimize oil carryover. Transfer this volume to a second sterile location to form a cleaner droplet (Figure 2, step 3).
5. Visually inspect the second droplet under magnification. Identify regions where oil droplets are least pronounced. Insert the pipette tip into this region and aspirate approximately 25 µL of medium, again leaving approximately 5 µL behind to minimize oil carryover (Figure 2, step 4).
6. Immediately transfer the collected 25 µL into a sterile, RNase-free microcentrifuge tube. Avoid touching the inner walls of the tube while dispensing the sample to prevent contamination and sample loss. Immediately snap freeze all collected samples in liquid nitrogen and store at -80 °C until NMR analysis (Figure 2, step 5).
7. Collect matched unspent incubated control droplets (medium incubated without embryos) prepared, incubated, and processed under identical culture conditions using the same culture dishes, overlay oil, incubation duration, incubator environment, and handling procedure described above to minimize environmental variability and potential systematic bias.
   NOTE: Strict adherence to sequential droplet transfer and visual inspection significantly reduces oil contamination. At this stage, the experiment may be paused and resumed prior to further sample processing for NMR acquisition.

3. Sample preparation for NMR analysis

1. Thaw samples at room temperature for ~10 min.
2. Prepare NMR mixture:
   1. Transfer 25 µL of SECM into a 1.5 mL microcentrifuge tube. Add 200 µL of D₂O. Vortex it for 10 s and then centrifuge at 1000 × g. for 10 s at room temperature.
3. Load insert: Transfer 225 µL of the mixture into a FEP liner.
4. Prepare the external reference tube.
   1. Fill a 5 mm NMR tube with 150 µL of D₂O containing 1 mM 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium salt (TMSP-d4).
5. Assemble the sample: Insert the FEP insert into the NMR tube and seal with parafilm.

4. NMR acquisition and spectral processing

1. Acquire ¹H-CPMG acquisition using the Bruker cpmgpr1d pulse sequence with: 64 scans; 65,536 data points; Relaxation delay: 8 s; Spectral width: 11.9 kHz; Echo time: 400 µs × 80 loops (effective T₂ filter 32 ms); Total acquisition time: ~15 min.
   NOTE: The effective CPMG T₂ filter duration of 64 ms was selected based on established metabolomics protocols for protein-containing biological samples, where this value provides efficient suppression of macromolecular signals while preserving low-molecular-weight metabolite resonances¹⁴.
2. Perform spectral processing:
   1. Obtain Fourier transform spectra.
   2. Apply 0.3 Hz exponential line broadening.
   3. Perform phase and baseline correction.
   4. Reference spectra to TMSP-d4 at 0 ppm.
   5. Reference chemical shifts (δ) to the external TMSP-d₄ signal at 0 ppm.
      NOTE: Because the sample mixture inside the FEP liner (225 µL) and the external TMSP-d₄ reference solution in the outer coaxial gap (150 µL) occupy distinct physical geometries and volumes within the active NMR coil detection window, their absolute signal intensities do not scale 1:1. Direct absolute quantification using the external reference requires a system-specific geometric calibration curve. Therefore, given the strictly constant sample volumes and highly reproducible pipetting precision of this protocol, raw bucket integrals were utilized directly to track relative metabolic variations between samples, omitting the need for external reference or total sum intensity normalization.
3. Perform spectral binning:
   1. Exclude the water resonance region (e.g., 4.65–5.00 ppm) and the TMSP-d4 reference region (e.g. -0.05–+0.05 ppm) prior to binning.
   2. Divide spectra into variable-width buckets to avoid splitting peaks.
   3. Use raw bucket integrals for statistical analysis rather than normalizing to the external TMSP-d4 signal, as the reference is physically isolated in the outer tube volume. This approach avoids quantification errors stemming from the different active volumes of the insert versus the outer tube.
   4. Export bucket integrals as a numerical data matrix for statistical analysis.
      NOTE: Each "bucket" (or bin) represents the summed numerical integral of all NMR signals within a defined ppm range. While these integrals reflect the relative concentration of the chemical groups present in that range, a single bucket may contain signals from multiple overlapping metabolites or only a portion of a specific metabolite's total signal.

5. Statistical analysis

1. Import processed ¹H-CPMG spectra into Python (v3.12) using nmrglue¹⁵.
2. Remove water and TMSP-d4 spectral regions.
3. Export bucket integrals for analysis.
4. Perform statistical testing using SciPy (v1.14.1):
   1. Assess normality of each spectral bucket using the Shapiro–Wilk test.
   2. Apply one-way ANOVA (for normally distributed data) or Kruskal-Wallis (for non-parametric data) to identify buckets with statistically significant differences between unspent incubated control and spent media.
   3. For buckets showing significant differences, perform structural assignment using a combination of 1D and 2D NMR experiments: ¹H-¹³C HSQC, ¹H-¹³C HMBC, and ¹H-¹H COSY.
   4. Compare the observed 2D correlations against standard databases (e.g., HMDB) to verify the identity of the metabolites contributing to the discriminatory buckets.

Results

The analytical workflow for untargeted NMR metabolomic profiling of spent embryo culture medium is summarized in Figure 1.

Removal of oil contamination during sampling
Figure 2 illustrates the sequential droplet transfer procedure used to reduce overlay oil carryover. Samples collected without careful pipetting exhibited broad hydrocarbon resonances spanning the aliphatic region of the spectrum (δ 0.5–2.3 ppm), resulting in substantial baselin...

Discussion

In this study, we addressed key analytical challenges associated with identifying and quantifying low-molecular-weight metabolites in spent embryo culture media using untargeted NMR metabolomics. Our results show that the optimized workflow enables acquisition of high-quality, information-rich spectra suitable for reliable detection of low-molecular-weight metabolites.

Careful sequential droplet transfer proved essential for reducing oil contamination, which otherwise generated broad resonance...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
5 mm FEP tube liner	SP Wilmad-LabGlass	4682701	Inner insert for coaxial NMR
5 mm NMR tube	SP Wilmad-LabGlass	535-PP-7-SB	Outer tube
6-well culture plate	Oosafe	OOPW-SW02	Embryo culture dish
Cryogenic liquid nitrogen	Messer	Contact manufacturer	Sample snap-freezing
Deuterium oxide (D2O)	Eurisotop	D214L	NMR solvent
Micropipette (10–100 µL)	Eppendorf	4924000053	Adjustable volume pipette
Microcentrifuge tubes, 1.5 mL, RNase-free	Eppendorf	N216966N	Sample collection and storage
Mineral oil	Vitrolife	VTL-10029	Culture droplet overlay
NMR spectrometer, 600 MHz	Bruker	https://www.bruker.com/en/products-and-solutions/mr/nmr/avance-nmr-spectrometer.html	Spectral acquisition
Pasteur pipette	Falcon	FAL-357575	Droplet transfer
Pipette, single-channel mechanical	Sartorius	728060	Sample handling
Pipette tips, sterile	Eppendorf	22491148	Compatible with micropipette
Pipette tips, sterile filter	Sartorius	790201F	Aerosol barrier
Sequential G-series culture media	Vitrolife	VTL-10143	Embryo culture medium
TMSP-d4 (sodium 3-trimethylsilylpropionate-d4)	Cambridge Isotope Laboratories	DLM-48-1	Chemical shift reference
Tubes, Safe-Lock	Eppendorf	0030121589	Storage tubes