Introduction

Combining functional and structural MRI enables a full-fledged view of diverse complex neurological mechanisms and response after therapeutic interventions. In this study we aim to characterize the regenerative capacity of intracerebrally injected human neural stem cells after ischemic stroke in the mouse brain. Repetitive resting-state functional MRI (rs-fMRI) and Diffusion MRI allow longitudinal in vivo characterization of spontaneous or therapeutically induced recovery on a structural and functional level. With our long observation time of 12 weeks after stroke induction we unravel for the first time the long-term effects of cell implantation on structural and functional brain networks.

Methods

Rs-fMRI and Q-Ball images (QBI) were acquired before, as well as in week 1, 2, 4, 8, and 12 after ischemic stroke induction. Two days after stroke, stem cells (n=11) / sham (n=6) implantation was carried out in Nu/Nu mice (male, 3-4 months) adjacent to the cortico-striatal lesions. Cell vitality was repetitively measured with Bioluminescence Imaging (BLI). Both MRI scans were always acquired in one session on a preclinical 9.4 T (Bruker, BioSpin) using a cryogenic surface coil with the following parameters and preprocessing steps: rs-fMRI: GRE-EPI sequence; TR/TE=2,840/18ms; 16 slices; slice thickness=0.5mm; matrix=9696; 182x182 µm2 res.; NR = 100. Functional datasets underwent slice-wise motion correction, high-pass filtering (>0.01 Hz), in-plane smoothing (Gaussian kernel=0.3mm) and physiological noise regression of recorded respiratory signals, motion parameters and drifts. QBI: 4-shot SE-EPI; TR/TE = 3,500/20ms; 128x128 matrix; 139x139 µm2 res.; 126 directions + 8 A0 images; Δ = 10ms, δ = 4 ms; b-value= 2,000s/mm2. A0 images were motion corrected and the linear interpolation was applied to the diffusion images. Postprocessing: Seeds of the sensorimotor cortex, as well as the thalamus and caudate putamen (CPu) were extracted after a multi-step process of linear and non-linear co-registrations of the AMBMC mouse brain atlas to the functional and diffusion data. For rs-fMRI, we calculated seed-based full cross-correlation maps with FSLNets (FMRIB Software Library). Diffusion anisotropy maps were generated and fiber tracking was performed with DSI-Studio.

Results

Mean z-scored correlation maps of unilateral and bilateral functional connectivity from the stem-cell treated group show a delayed decrease in connectivity only after week 2 of stroke induction (Figure 1). Interestingly, after week 2, the homotopic contralesional side is severely affected up until the end of the experiment, after week 12, for both groups. Maximal loss of inter-seed correlation is observed in week 8 for the cell treated group and in week 4 for the sham group. The time courses of the intra-seed correlations show a gradual decline for all seeds until 8 weeks after stroke for the cell group, whereas in the sham animals the correlation already starts to increase again after 2 weeks (not shown). Diffusion parameters representing the degree of anisotropy in a specified seed stay on a constant level for the caudate putamen and the somatosensory cortex excluding the limb area (Figure 2). However, the anisotropy curve of the thalamus seems to achieve a plateau from week 4 on after stroke for the animals implanted with a sham solution.

Discussion

Our results show a delayed decrease of functional connectivity in the sensorimotor network for the stem-cell group, suggesting a stabilizing function of the cells in the first weeks after stroke. This effect did not sustain longer than four weeks and correlates with the loss of cells, monitored with BLI. The persisting, strongly impacted functional network connectivity on the contralesional hemisphere is partially reflected in the similar results of the time courses from the mean intra-seed correlation. This observation indicates a relation between the strength of an intra-seed correlation and the correlation between seeds. The results from the functional datasets illustrate that a homotopic change in functional networks not only affects short-range interactions, but has a large impact on long-range brain correlations as well. The thalamic time courses of the functional results and the anisotropy indices from the diffusion data reveal an inversely proportional behavior, with a functional delay, demonstrating a relation between structural and functional recovery.

Conclusion

In this study we present a new mode of action of the therapeutic capacity of stem cell implantation after stroke in vivo on the neuronal circuits, merging the structural and functional point of view. With our results, we show the importance of a global analysis from different angles, and the necessity for long observation times after stroke. Furthermore, we provide new insight into the structure-function relationship in neurodegenerative diseases.

Acknowledgements

This work was supported by grants from the EU-FP7 programs TargetBrain (HEALTH-F2-2012-279017) and BrainPath (PIAPP-GA-2013-612360).

References

No reference found.