Lab-on-a-Chip Metabolic Analysis Using Benchtop NMR Technology

Abstract

Organ-on-chip (OoC) systems are advancing rapidly as physiologically relevant in vitro models. However, real-time, noninvasive metabolic monitoring tools remain lacking. While NMR offers a powerful solution, current setups are not compatible with the planar format of microfluidic chips. Here we introduce the first benchtop NMR spectrometer for real-time metabolic monitoring of cell cultures on microfluidic platforms, utilizing hyperpolarization via dissolution dynamic nuclear polarization. This work details modifications made to a commercial benchtop NMR spectrometer, including the design and fabrication of a microfluidic platform that enables precise injection of hyperpolarized substrates and continuous renewal of cell culture media. The platform integrates a radiofrequency coil for signal transmission and reception and incorporates a custom-built sample carrier. Preliminary NMR data acquired with this system demonstrate its feasibility for dynamic metabolic studies.

1

Overview of the design of benchtop NMR spectrometer for the detection of hyperpolarization-enhanced NMR metabolism in 3D tissue engineered cell models. (A) NMR spectrometer. (B) Chip carrier with the microfluidics chip and the embedded RF coil. (C) 3D liver cancer spheroids positioned within the chip’s detection region. (D) Example of ¹³C NMR spectrum observed upon injection of hyperpolarized [1-¹³C]­pyruvate and its conversion to lactate by the cancer cells in C.

2

Microfluidic device and fabrication mold. (A) (i) Tilted top view of the microfluidic device highlighting key features, including the membrane pump, microfluidic channels, saddle coil embedded around the sample well, the scientific chamber (well), and the expected engineered 3D cell model inserted into the well. (ii) Bottom view of the microfluidic device showing the hyperpolarization channel and the bottom glass. (B) Poly­(methyl methacrylate) (PMMA) mold used to fabricate the microfluidic device. (i) Top view of the mold replicating the bottom side of the chip. (ii) Lateral view of the mold replicating the bottom side of the chip. (iii) Top view of the mold replicating the top side of the chip. (iv) Lateral view of the mold replicating the top side of the chip.

3

Simulation results for the optimized saddle coil: (A) schematics of the 2 mm height coil, 6 turns, 0.6 mm diamameter, relative permittivity 80 and conductivity 1.2 S m^–1, (B) B ₁ field magnitude in the transversal plane, (C) B ₁ field magnitude in the coronal plane, (D) B ₁ field magnitude in the sagittal plane. The black thin line represents the cylindrical phantom position.

4

Schematics of the tuning and matching circuit used to interface the embedded RF coil with the spectrometer.

5

Schematic representation of the setup showing the injection of a hyperpolarized solution through the channels of the microfluidic device, which is positioned inside the spectrometer.

6

Longitudinal relaxation measurements of [1-¹³C]­pyruvate in the BLOC spectrometer setup. (A) Schematic of the procedure used to inject the hyperpolarized pyruvate dissolution into the microfluidic device through the hyperpolarization channel, filling the well. Two channels were blocked, leaving one inlet and one outlet. The dashed line in the right figure indicates the physical position of the saddle coil. (B) Plot illustrating the exponential decay of the [1-¹³C]­pyruvate signal due to T ₁ relaxation and radiofrequency pulses. The T ₁ value was calculated by fitting the exponential decay function, yielding a value of 72 ± 3.4 s. The stacked spectra represent the dynamic ¹³C measurement of hyperpolarized [1-¹³C]­pyruvate over time.

7

Kinetic analysis of HepG2 cell metabolism. Time 0 s corresponds to the start of NMR acquisition. The delay between the initial contact of the cells with the hyperpolarized substrate and the first acquired spectrum was accounted for in the kinetic rates calculations. (A) Monoexponential decay function fit to the [1-¹³C]­pyruvate polarization decay due to T ₁ relaxation, radiofrequency pulses, and metabolism for the first replicate. (B) Kinetic function fit to the production of [1-¹³C]­lactate for the first replicate. (C) Same as (A) for the second replicate. (D) Same as (B) for the second replicate. (E) Same as (A) for the third replicate. (F) Same as (B) for the third replicate.