To use the intrinsic volume selectivity of miniaturized coils to perform highly-localized spectroscopy within seconds in an animal scanner.

A 2mm-diameter circular coil (Fig.1), tuned, matched, and configured as transceiver, is used in a Bruker Biospec spectrometer (Bruker BioSpin GmbH, Ettlingen, Germany) interfaced to a 14.1-T/26-cm horizontal magnet (Magnex Scientific, Abingdon, Oxfordshire, UK). The coil was built with a 400µm-diameter copper wire and is surrounded by plastic for susceptibility-matching. Spectra of a phantom metabolite solution (Table 1) and A9L mouse fibroblasts cells were acquired without using gradients. The same type of culture chamber (Ibidi®, Germany) was used for growing the cells and for measuring the phantom metabolite solution.

Two types of culture chambers were used: The first has a polymer foil as bottom (thickness ~180µm), the second has a glass bottom (thickness ~100µm), both with 400µm of chamber height. Several experiments with FIDs acquisitions were performed with 128 (Fig.2) and 8 averages (Fig.3), with TR=3sec (total scan times of 6:24min and 24sec, respectively) for the phantom solution and cells. All FIDs were obtained with 4096 data points (no water suppression). An exponential window function was applied prior to Fourier transformation. After thorough phase correction, manual baseline correction was performed.

The measured linewidth for the phantom solution water-peak was ~39Hz with a SNR=5536 (calculated as in Meier et al²). Spectra from culture medium-only and the A9L cells showed linewidths between 32 to 58 Hz, and SNRs from 3117 to 5144. The estimated sensitivity volume of the coil in the culture chamber was ~1.5µL. Figures 2-4 display resulting spectra of the metabolite solution, culture medium-only, and A9L cells grown in culture medium.

Figures 2-3 shows the possibility to detect the resonances of all metabolites from a small volume (~1.5µL) in an acceptable scan time 6:24min (Fig.2) for the phantom solution, where even the multiplet resonances of choline methylene groups (2H, 5mM) are well visible. Even with reduced scan time (24sec) we are still able to obtain the larger signals from the major metabolites with still acceptable SNR (Fig.3).

Figure 4 displays the results from the cell measurements. In the spectrum of pure cell culture medium-only the resonances of glucose (4.5mM) are detectable when measured in the 180µm-thick-polymer-bottom culture chamber. All other medium ingredients (next highest concentrated is glutamine with 2mM) are too lowly concentrated to be reliably detected. Comparing the spectra obtained in the presence of cells with the one of medium-only, a reduction of the glucose signals can be observed. In addition, the lactate signal appears, which is a metabolic product of the cells, is excreted by the cells and thus accumulates gradually in the medium. However, no metabolites from inside the cells could be detected. Since the monolayer of cells has only a height of ~5-15µm, the contribution of intracellular metabolites to the signal from the sensitive volume is most likely too low under current conditions. As expected, the sensitivity increased when switching from normal culture chambers with 180µm thick bottom to the 100µm of glass-bottom chambers allowing the potential detection of additional metabolites (e.g. glutamine) in the medium.

The major metabolites in the phantom solution can be detected at physiological concentrations in few seconds to minutes with the current setup when reaching these concentrations in the whole sensitive volume. For metabolites in cells that are grown in a monolayer, this prerequisite is not fulfilled. Therefore, at the moment the presence of cells is only indirectly detectable by changes in the culture medium, for instance the accumulation of lactate or the decrease of glucose. However, using these microcoils on/in tissue so that preferentially the whole sensitive volume above the coil is filled by cells/tissue should permit the detection of the higher-concentrated intracellular metabolites. Experiments on ex-vivo tissue samples are under progress. More improvements are needed in the coil construction, electronics, and pulse-sequence design, in order to achieve faster and highly-localized spectra. The linewidth will be improved by using better susceptibility-matched materials for the coil and its surroundings. Further miniaturization and microelectronics integration³ will allow insertion and in-vivo studies⁴ of localized activity within single layers with high temporal resolution.