Nuclear magnetic resonance footprint of Wharton Jelly mesenchymal stem cells death mechanisms and distinctive in‐cell biophysical properties in vitro

Abstract The importance of the biophysical characterization of mesenchymal stem cells (MSCs) was recently pointed out for supporting the development of MSC‐based therapies. Among others, tracking MSCs in vivo and a quantitative characterization of their regenerative impact by nuclear magnetic resonance (NMR) demands a full description of MSCs’ MR properties. In the work, Wharton Jelly MSCs are characterized in a low magnetic field (LF) in vitro by using different approaches. They encompass various settings: MSCs cultured in a Petri dish and cell suspensions; experiments‐ 1D‐T 1, 1D‐T 2, 1D diffusion, 2D T 1‐T 2 and D‐T 2; devices‐ with a bore aperture and single‐sided one. Complex NMR analysis with the aid of random walk simulations allows the determination of MSCs T 1 and T 2 relaxation times, cells and nuclei sizes, self‐diffusion coefficients of the nucleus and cytoplasm. In addition, the influence of a single layer of cells on the effective diffusion coefficient of water is detected with the application of a single‐sided NMR device. It also enables the identification of apoptotic and necrotic cell death and changed diffusional properties of cells suspension caused by compressing forces induced by the subsequent cell layers. The study delivers MSCs‐specific MR parameters that may help tracking MSCs in vivo.

activating the regeneration processes in injured tissues. 6 For that purpose, MSCs will have to stay alive in the regenerating tissues for a prolonged period of time. However, we are still unable to efficiently trace MSCs in patients after transplantation. Due to the lack of full knowledge about their biology and behaviour after injection, the MSCs cannot be fully utilized in regenerative medicine. Thus, new methods that help to solve these problems are urgently needed.
NMR enables the study of porous material or biological systems in a non-invasive manner and in vitro or in vivo conditions. 7,8 Tracking the migration of transplanted stem cells with the use of NMR techniques has several years of practice. However, most of the recent research is based on contrast agents labelling the cells. 9-11 T 2 relaxation times or diffusion coefficients, D, as biomarkers have only been used in a few papers. 12,13 The long-term purpose of the study of live stem cells by means of truly non-invasive NMR, that is also without contrast agents, is twofold. First of all, it concerns the determination of specific parameters 'seen' by low field NMR (LF-NMR), such as relaxation times T 1 , T 2 , T 1 -T 2 and D-T 2 maps, or diffusion coefficients, which are characteristic for Wharton Jelly MSCs. The proposed multi-parametric characterization is also implemented to obtain a set of MR parameters in order to minimize the possibility of overlapping signals from other cells.
These parameters may be useful in-cell detection when studying animal models or patients by means of MRI in vivo. A similar approach was developed and implemented for porous and heterogeneous systems. 14,15 Secondly, NMR parameters characterizing in vitro cell suspensions can be used to determine their quantitative and qualitative characteristics, such as size, self-diffusion coefficient and viability. For this purpose, besides results from the characterization of MSCs by LF-NMR, a single-sided Mobile Universal Surface Explorer (MoUSE) was used. MoUSE allows the study of a sample using an extremely strong magnetic field gradient (~24 T m −1 ) and short diffusion times, which leads to higher diffusion weighting without coming into motional averaging between compartments. New promising cell studies, carried out under these conditions and considering several signal components from cell samples, have appeared recently. 16,17 Another advantage is the ability to test samples in open geometry with the use of mobile apparatus, 8 which increases the potential of future uses in the case of finding optimal measurement protocols and parameters dependent on cells characteristics.

| Experimental model
The umbilical cords were collected after Caesarean sections.
Written consents were obtained from parents. The umbilical cords were washed with phosphate-buffered saline supplemented with antibiotic-antimycotic solution, cut into small explants and plated into a plastic flask. Explants were cultured with a growth medium for MSCs (DMEM Low Glucose, Biowest), supplemented with the platelet lysate in standard culture conditions under 21% of O 2 and 5% of CO 2 at 37°C. Next, the explants were removed, and the cells were passaged using the Accutase cell detachment solution (BioLegend).
After reaching the appropriate number of cells, WJMSCs were used for further experiments.

| Experiments in a LF-NMR system with a bore aperture
A suspension of MSCs from Wharton Jelly in a PBS buffer was put into glass pipette and centrifuged. Then, the glass pipette was closed and so the prepared samples were examined on a Magritek Rock Core Analyzer at a magnetic field of 0.05 T. Samples with 5 and 15 million cells in a volume of 0.5-1 ml were tested (suspensions a-d, see Table 1). The Inversion Recovery (IR) and Carr-Purcell-Meiboom-Gill (CPMG) sequences were used for 1D-T 1 (inter-experiment delay, ID = 5 s, T 1 delay range: 0.1-5 s) and T 2 (ID = 7.5 s, echo time, TE = 200-400 μs, number of echoes in CPMG encoding train, NoE = 50,000) measurements, respectively. 2D T 1 -T 2 correlation maps were obtained with IR-CPMG sequence ID = 3 s for buffer, ID = 350 ms for cells, T 1 delay range: 0.1-5 s, TE = 400 μs, NoE = 20,000). In order to enhance the signal from cells, shorter inter-experiment delays were applied for T 1 -T 2 (350 ms) than in the case of 1D experiments. For 2D complementary diffusion experiments, a diffusion-weighted pulsed-field gradient spin-echo (PGSE) sequence was applied with an increasing gradient amplitude to 0.5 T m −1 and CPMG sequence for detection (ID = 350 ms, TE = 400 μs, NoE = 10,000, gradient pulse length, δ = 6 ms for suspensions a and c, δ = 8 ms for suspension d, interval between two gradient pulses, Δ = 20 ms). The maximum b-value achieved for suspensions a and c was equal to 11.6 × 10 9 sm −2 and 19.8 × 10 9 s m −2 for suspension d. All the experiments were conducted in seven separate experimental series, and for each cell concentration measurement with the same parameters was repeated at least once. In the work, representative data were shown. for cylindrical container were registered. Then, the obtained data were calculated using the Inverse Laplace Transform (ILT) (L&H algorithm, Prospa software) and fitted independently using a oneor bi-exponential model.

| Quantification and statistical analysis
The registered data were analysed using ILT with Lawson&Hanson and FISTA algorithms, 18 allowing us to obtain 1D distributions and 2D maps, respectively (Prospa software, Magritek). Data from single-sided NMR-MoUSE were additionally processed by fitting independently a mono-or bi-exponential diffusion model (for descriptions please see for example in the work of Mazur and Krzyżak 16

| WJMSCs characterization
The WJMSCs show the minimal criteria outlined for MSCs by the International Society of Cellular Therapy. They adhere to plastic surface in standard culture conditions and display fibroblast-like morphology ( Figure 1A). Cytometric analysis revealed high expression of specific mesenchymal markers. More than 90% of cells were CD73, CD90 and CD105 positive, whereas they do not express hematopoietic antigens (CD45, CD14, CD19, CD34 and CD3) ( Figure 1B).
We have also confirmed multipotent differentiation potential of TA B L E 1 Peak positions from 1D-T 1 , T 2 distributions (A) and 2D T 1 -T 2 (B) and D-T 2 (C) correlation maps WJMSCs. These cells demonstrated strong capacities for differentiation towards adipogenic ( Figure 1C), osteogenic ( Figure 1D) and chondrogenic ( Figure 1E) lineages.

| T 1 and T 2 relaxation
In Figure 2, the T 1 and T 2 distributions for MSCs samples with various amounts of cells in a specified volume are presented, and in Table 1, the relaxation times at maximum and T 1,2 log-mean values are collected. In the case of T 1 distributions only one peak is visible, for both the buffer and for the cell samples (see Figure 2

| T 1 -T 2 and D-T 2 correlation maps
In Figure 3A-C, T 1 -T 2 maps are presented corresponding to the 1D distributions from Figure 2 for suspension a, c and d. A peak with an increasing intensity and area for the cell samples, located at T 2 about 130-350 ms and not present for the pure buffer sample, is the main observation for these measurements. Its T 1 /T 2 values were a few times higher than for a free water, which is another confirmation of the assumption that the signal originates from the restricted region of the sample.
A comparison of D-T 2 maps for the pure buffer and MSCs samples is shown in Figure 3D-F. It can be observed that for the used PGSE parameters signal with T 2 from the range of 130-350 ms was

| Diffusion measurements of cells cultured in a Petri dish
In Figure Table 2) and presented in Figure 4E and J. Effective diffusion coefficients for the bottom slices of the stem cell samples (slice 1, Figure 4D and I) were

| Diffusion measurements of cell suspension in a cylindrical container
In order to obtain the results of diffusion coefficients for cells less

| DISCUSS ION
In the work, different NMR approaches were applied in order to characterize MSCs. Each of them delivered distinct information which was complementary to the others, and all are discussed below.

| Evidence of diffusion in the in-cell structures
The MSCs nucleus is a large and round cellular structure 20 and was suspected in the first place to contribute to the second diffusion component. For example, in yeast, nuclear to cellular volume ratio is equal to about 8%, 21 which is associated with a nuclear radius of ~1 μm. 22 In the TA B L E 2 Results of fitting of mono-exponential functions for stem cells and water sample in a Petri dish for slices of 10 µm (A) and biexponential functions for stem cells and mono-exponential for water sample in a cylindrical container for slices of 50 µm (B) mouse MSCs, the ratio of cell and nucleus diameters was reported to be equal to about 63%. 23 In human MSCs, the ratio of nuclear to cellular diameters is equal to 26-31%. 24 Assuming the cell radius of 7.5 μm estimated in Section 4.1, nucleus radius, R nucl , is equal to ~2 μm. This size and

| Determination of MSCs' self-diffusion coefficient
In the previous work, 16  It is important that for t d < τ nucl , signal attenuation due to diffusion in nucleus would be so small that would require a very high signal-to-noise ratio to be distinguished as a separate component, especially in small samples for which f nucl is low (in this study it is equal to ~0.24%-1.5% of a total signal). Practically, if pulsed-field gradient (PFG) techniques are used, most of the attenuation will come from diffusion in cytoplasm. Therefore, for t d ≫ τ nucl effective/intracellular diffusion coefficients D intra will be observed, while for t d →0 D 0,intra~D0,cyto . Based on the fact that intracellular self-diffusion coefficients of 0.68 × 10 −9 m 2 s −1 25 and 0.65 × 10 −9 m 2 s −1 26 were obtained for yeast cells, while self-diffusion coefficient of 0.69 × 10 −9 m 2 s −1 16 was obtained for yeast's cytoplasm, the D 0,intra~D0,cyto approximation seems to be justified for PFG in moderate gradient strengths. Since selfdiffusion coefficient of nucleus is known, D cyto (t d ) was extracted from D intra (t d ) and used for determination of D 0,cyto . Estimated

| SUMMARY
The study revealed the capability of a low field system to detect sig- In further research, the known specific NMR parameters will be This indicates that diffusion can be proposed as a natural biomarker of a cell viability. Based on the obtained results, it seems that necrosis and apoptosis can be distinguished, which can be achieved thanks to the ability of NMR-MoUSE device to detect low diffusivity components, similar to a nucleus. This provides the opportunity to trace tissue destruction or tissue remodelling through the evidence of elements of cell dissolution. However, reference studies, such as microscopy, are required.

ACK N OWLED G EM ENTS
The work was financed by the National Centre for Research

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are openly avail-