Introduction

A sodium triple-quantum(TQ) signal is created when sodium ions interact with negatively charged groups of macromolecules, mainly proteins. An increased sodium TQ signal was observed during different Na⁺/K⁺-ATPase inhibitions, which increased the intracellular sodium concentration, in isolated rat hearts¹. A linear correlation of the TQ signal with the intracellular sodium concentration is an attractive feature to quantify cellular responses¹. An MRI-compatible microcavity array-based bioreactor can be used to study different mammalian cells². Such cell experiments allow a large flexibility with cell interventions and combined with MR is a promising tool for anti-cancer drug development.

In this study, we modified the TQ time proportional phase incrementation(TQTPPI) sequence^3,4. This modification allowed us to achieve several times gain in TQ signal sensitivity while preserving the simultaneous measurement of the TQ and single-quantum(SQ) signal at distinct frequency offsets. In a second step, the capability of the sodium TQ signal to monitor intracellular sodium changes during Na⁺/K⁺-ATPase inhibition of HepG2 cells was evaluated.

Material and Methods

Data was acquired at a 9.4T preclinical MRI(Bruker Biospec 94/20, Ettlingen, Germany). A ¹H/²³Na/³⁹K Bruker volume coil was used for the phantom measurements, while a ²³Na in-house built surface coil was used for the bioreactor experiments.

Hepatocellular carcinoma cells(HepG2) cultivated three-dimensionally within two microcavity arrays(MCAs) were actively perfused(400 μl/min) under normoxic conditions at 37°C inside a bioreactor system²(Fig.1). HepG2 cells(ATCC HB-8065, USA, Manassas) were prepared according to^2,5,6.

The TQTPPI sequence was modified so that phase α was still incremented while the evolution time was optimized and fixed throughout the experiment(Fig.2). This approach increases the TQ SNR as the TQ signal is always measured at its maximum value when τ_evo=τ_opt. In comparison to TQ filtration, the fixed TQTPPI still separates the different coherences at different frequency offsets. The sensitivity gain of the fixed TQTPPI sequence(TR=300ms, 720 phase steps) in comparison to the standard TQTPPI sequence(TR=300ms, 720 time steps) was evaluated in three phantoms with [2, 4, 6]% agarose in 134.75mM NaCl. The standard TQTPPI FID was fitted according to⁴ and the sodium TQ and SQ signal amplitude were corrected to the fixed evolution time according to their fit function. The fixed TQTPPI FID was fitted by:
$$Y(t) = A_{SQ} sin(α+θ_1) + A_{TQ} sin(3α+θ_2) + DC$$
Where Y(t) is the FID amplitude; A_SQ and A_TQ are the sodium TQ and SQ amplitude; θ_i are phase offsets and DC is a baseline offset.

In cell experiments, the fixed TQTPPI sequence(TR=250ms, 420 phase steps, TA=4min) was used for sodium TQ measurements. To evaluate the TQ signal origin, the TQ signal from collagen on two MCAs in medium was compared to HepG2 cells on two collagenized MCAs in medium. In a second step, the Na⁺/K⁺-ATPase was blocked in two 3D cultures with 16*10⁶ and 14*10⁶ HepG2 cells for 60min using either 1mM ouabain or a K⁺-free medium. The sodium TQ signal was monitored for 50 min before intervention and 190min after reperfusion with normal medium. The time course of A_TQ/A_SQ was normalized to the first 50min. The K⁺-free medium consisted of 1.8mM CaCl₂, 0.8mM MgSO₄, 26.2mM NaHCO₃, 117.2mM NaCl, 1.0mM NaH₂PO₄-H₂O and 5.6mM glucose. The TQ signal of cells in K⁺-free medium was corrected by the background TQ signal since the K⁺-free medium did not contain proteins.

Results and Discussion

In the standard TQTPPI spectrum the different coherences can be described by Lorentzian functions⁴(Fig3a), while in the fixed TQTPPI spectrum the coherences are close to delta functions(Fig.3b). The evaluation of both sequences in [2, 4, 6]% agarose phantoms yielded the same values within the 95% confidence intervals for A_TQ, A_SQ and A_TQ/A_SQ(Fig.3c). However, for the fixed TQTPPI sequence, a measured TQ SNR gain of 3-4 was even higher than expected(Fig.3c). This higher SNR gain resulted from fewer fitting parameters for the fixed TQTPPI FID fit compared to the standard TQTPPI FID fit.

The measurement of two MCAs in medium resulted in a small TQ signal A_TQ/A_SQ=0.4±0.6‰(Fig.4). The TQ signal from 14*10⁶ HepG2 cells on two MCAs in medium increased to A_TQ/A_SQ=1.2±0.7‰(Fig.4).

Blocking the Na⁺/K⁺-ATPase by 1mM ouabain for 60min led to an increase of the TQ signal to 128.9±3.9% of the pre-intervention level(Fig.5a). Reperfusion with normal medium did not reduce the TQ signal, but a further TQ signal increase was observed. This continuation of sodium influx into the cells is indication of irreversible cell damage during ouabain perfusion.

Perfusion of cells for 60min with a K⁺-free medium produces a specific Na⁺/K⁺-ATPase inhibition which corresponds to a maximum TQ signal increase of 165.0±7.4%(Fig.5b). After reperfusion, the TQ signal recovered to 129.5±4.5%. This recovery indicated that cells remain viable and partly preserve Na⁺/K⁺-ATPase pump activity.

Our results correlate with earlier experiments in isolated rat hearts¹ where an increase of the TQ signal was observed by 190±9% and ~265% for 30min of 1mM ouabain and 60min of K⁺-free medium, respectively. The TQ signal increases in our study were lower, which may indicate a higher intracellular sodium concentration in cancer cells.

Conclusion

Changes in the intracellular sodium of cancer cells in a microcavity array-based bioreactor were successfully detected by sodium TQ signal. This bioreactor design with the improved TQ signal detection provides a capability to investigate a variety of cells using sodium TQ signal.

Acknowledgements

One of the authors (Schepkin, V.D.) would like to acknowledge the support from the National Science Foundation through NSF/DMR-1644779 and State of Florida.