Diffusion tensor imaging analysis to assess stem cell therapy efficacy in traumatic brain injury
Nastaren Abad1,2, Abdol Aziz O. Ould Ismail1,2, Ali Darkazali3, Jens T. Rosenberg1, Cathy Levenson3, and Samuel Colles Grant1,2

1Center for Interdisciplinary MR, National High Magnetic Field Laboratory, Tallahassee, FL, United States, 2Chemical & Biomedical Engineering, Florida State University, Tallahassee, FL, United States, 3Biomedical Sciences and Program in Neuroscience, Florida State University, Tallahassee, FL, United States

Synopsis

Traumatic Brain Injury (TBI) interferes with the functionality of the brain due to heterogeneous complications that continue after the initial trauma. As a result, stem cell therapy is viewed as a potential treatment approach that can be instituted during the chronic phase of TBI. This study employs Diffusion Tensor Imaging (DTI) to non-destructively probe the pathological impacts of endogenous Neural Progenitor Cells (NPC) and exogenous Mesenchymal Stem Cells (MSC) in a rodent model of TBI. This study represents the first investigation of the efficacy of MSC treatment in TBI and the potential synergistic effect of MSC and NPC.

Purpose

To analyze quantitatively white matter recovery and reorganization following Traumatic Brain Injury (TBI) due to the action of endogenous neural progenitor cells (NPC) and exogenous human mesenchymal stem cells (hMSC). The potential synergistic effect of hMSC and NPC was evaluated by utilizing irradiation of the sub-ventricular zone to eliminate proliferation of the endogenous NPC.

Methods

TBI Model and Delivery of hMSC: The TBI rodent model has been described previously (Darkazali et al., 2015), making use of a controlled cortical impact (CCI). Eight-week old Sprague-Dawley rats (N=25, N=5 per treatment group) were anesthetized and fixed within a stereotaxic frame to perform a 6-mm craniotomy just rostral to Bregma. A programmable CCI device was used to deliver a 5-mm diameter bilateral cortical injury by means of an impact piston delivered at a velocity of 2.25 m/s to a depth of 3.0 mm and with a dwell time of 500 ms. Six hours after TBI, the rats were administered either: an intravenous injection of passage-3 hMSC (106 cells) via the tail vein or a saline injection. Sham surgeries, with saline injection, also were instituted.

Subventricular Zone (SVZ) Irradiation: Targeted irradiation (Irr) was performed under anesthesia on the rats directly after the TBI by making use of a custom-built lead shield, which protected the rest of the brain and was positioned 7.0 mm rostral to the interaural line and thereby exposed only the SVZ. Irradiated rats received a dose of 8 Gy over a time of 300 s (Tada et al., 1999; Shinohara et al., 1997) to eliminate the proliferation of SVZ-located NPC. Shams were anesthetized for a comparable duration.

Treatment groups: The following treatment groups were generated and analyzed: (1) TBI/no Irr/saline; (2) TBI/no Irr/hMSC; (3) TBI/Irr/saline; (4) TBI/Irr/hMSC; and (5) Sham.

MRI Acquisition: Three weeks post-TBI and following transcardial perfusion and fixation with 4% paraformaldehyde, N=5 rats in each treatment group were imaged ex vivo at 11.75 T. This vertical magnet is equipped with a Bruker Avance console (Bruker-Biospin, Billerica, MA), and all samples were imaged with a 25-mm radio frequency (RF) coil resonating at 500 MHz. All samples were scanned with brains within the skull to minimize distortions and changes in neuroanatomical structures and volumes.

Lesion Volumetrics: To evaluate lesion volume, a 3D gradient recalled echo (GRE) sequence (Figure 1) was acquired at an isotropic resolution of 100 μm with TR = 150 ms and TE = 7.5 ms and FOV = 2.8x2.5x2.5 cm. Volumetric analysis was performed using AMIRA 5.4.3 (FEI, Hillsborough, OR). Segmentation of the brains was manually conducted by tracing the injury periphery.

Diffusion Tensor Imaging: A high resolution, six-direction DTI evaluated impacts to white matter tracts (b value = 1000 s/mm2). To achieve an in-plane resolution of 100x100-μm, a 2.5x2.5-cm FOV was sampled with a 250x250 matrix and TE/TR=25/2750 ms over 11 h (Figure 2). DTI analysis was performed using a multiple region of interest approach in DSI Studio (Yeh et al., 2013). The brains were segmented anatomically by manually tracing different regions in the brain: Corpus Callosum (whole, body, genu and splenium), Thalamus, Hippocampus, Internal Capsule, Putamen and Neo-Cortex. Fractional Anisotropy (FA, Table 1) and Apparent Diffusion Coefficient (ADC, Table 2) were compared statistically between treatment groups (One-way ANOVA with Tukey’s post-hoc comparison test, p<0.05).

Results and Discussion

Tractography analysis showed differences between the TBI groups and shams in the Corpus Callosum in the form of reduced tracts as shown in Figure 3 in the area adjacent to the injury and surprisingly near the splenium. Segmentation of the Corpus Callosum showed a loss in tract volume in TBI models compared to sham. Therefore, DTI was sensitive to the presence of TBI.

More importantly, data suggest a strong correlation between the action of the endogenous and exogenous stem cells. A significant difference (Tables 1 and 2) was evident in the corpus callosum for the two extreme groups investigated: TBI/Irr/Saline and TBI/no Irr/hMSC. Further, differences were evident in the Genu between the groups TBI/Irr/Saline (no stem cell action) and TBI/no Irr/hMSC (synergistic action). Interestingly, DTI was not as sensitive as behavioral assays (Darkazali et al., 2015) of anhedonia/depression and novel object recognition/short term memory. Though FA and ADC values show promise as therapeutic biomarkers, ex vivo DTI analysis detected only limited reorganization, largely due to the suspected presence of migratory cells within white matter tracts.

Acknowledgements

This work was performed partially at the National High Magnetic Field Laboratory (NHMFL), which is supported by NSF DMR-1157490 and the State of Florida. Funding also was provided by the NHMFL User Collaboration Grants Program (to SCG) and through graduate stipend support from the FSU Department of Chemical & Biomedical Engineering.

References

1) Darkazalli A, Ismail AA, Abad N, Grant SC, Levenson C. 2015. Human Mesennchymal Stem Cell Treatment after Traumatic Brain Injury Prevents Trauma-Induced Depression and is Dependent on Subventricular Zone Proliferation. J. Neurotrauma. In review.

2) Shinohara C, et al. 1997. Apoptosis in the subependyma of young adult rats after single and fractionated doses of X-rays. Cancer Research, 57(13), pp.2694–2702.

3) Tada E, et al. 1999. Long-term impairment of subependymal repopulation following damage by ionizing irradiation. Experimental Neurology, 160(1), pp.66–77.

4) Yeh F-C, et al. 2013. Deterministic Diffusion Fiber Tracking Improved by Quantitative Anisotropy. PLoS ONE 8(11): e80713.

Figures

Fig 1. (A) TBI fluid-filled cystic lesion (green) within a full brain segmentation; (B) Brain segmentation relative to ex vivo 3D GRE; (C) volumetric analysis of lesion at week-3 post TBI. Irr/nohMSC lesion volume was 73.2±13.0 mm3; noIrr/hMSC treatment decreased lesion by 29.5%. Significant differences between treatments were not observed.


Fig 2. Representative images: (A) one DWI from the six-direction DTI acquisition, (B) Apparent Diffusion Coefficient (ADC) map and (C) Fractional Anisotropy (FA) map.

Table 1. Fractional anisotropy (FA) values for TBI treatment groups (n=5) and Sham (n=3) rats obtained from ex vivo DTI. Statistical significances are assessed at *p<0.05 (Tukey's test). Abbreviations: CC whole = Corpus Callosum and associated white matter tracts; CC = Corpus Callosum; IC = Internal Capsule. Values are means±SD.

Table 2. Apparent Diffusion Coefficient (ADC) for TBI treatment groups (N=5) and Sham (N=3) rats obtained from ex vivo DTI. Statistical significances are assessed at *p<0.05 (Tukey's test). Abbreviations: CC whole = Corpus Callosum and associated white matter tracts; CC = Corpus Callosum; IC = Internal Capsule. Values are means±SD.

Fig 3. Representative fiber tractography of the entire Corpus Callosum (thresholds: FA=0.581 and angular=45°) to display differences in treatment groups based on tract density. (A and B) Sham; (C and D) TBI/no Irr/hMSC; (E and F) TBI/no Irr/Saline; (G and H) TBI/Irr/hMSC; and (I and J) TBI/Irr/Saline



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
3364