Apparent diffusion coefficients in spinal cord transplants and surrounding white matter correlate with degree of axonal dieback after injury in rats

Abstract

Background and purpose: Abnormal apparent diffusion coefficient (ADC) values in injured spinal cord white matter and fibroblast transplants have been shown to correspond with qualitative histologic findings of axonal loss or regeneration. We proposed that ADC values would correlate with quantitative axonal tracing in the transected rubrospinal tract (RST).

Methods: Eleven rats received right-sided lateral funiculus lesions at C3-4 (disrupting the RST) and transplantation of fibroblasts that were unmodified or modified to secrete brain-derived neurotrophic factor (BDNF). Behavioral tests measured hindlimb function at 1, 2, 4, 6, 8, 10, and 12 weeks after injury. At 12 weeks after injury, the antegrade axon tracer biotinylated dextran amine was stereotactically injected into the red nucleus to label the injured RST axons. Animals were sacrificed 2 weeks later. Diffusion-weighted MR imaging of the excised, fixed spinal cord specimens was then performed at 9.4 T.

Results: In white matter surrounding transplants, ADC values transverse to axons were elevated and ADC values longitudinal to axons were decreased. These ADC values were more abnormal closer to the transplant, and this correlated with decreases in numbers of labeled RST axons. ADC values in BDNF-expressing fibroblast transplants were significantly lower than those in unmodified fibroblast transplants, and these lower values correlated with decreased axonal dieback. Behaviorally, all animals showed partial recovery, but animals with BDNF-expressing fibroblast transplants had slightly improved hindlimb function compared to those with unmodified fibroblast transplants.

Conclusion: ADC values may be able to evaluate graft function after spinal cord injury by demonstrating the degree of axonal dieback and preservation.

Fig 1.

Schematic of rat cervical spinal cord. The labeled tracts in the white matter are RST (1), dorsal corticospinal tract (2), fasciculus gracilis (3), fasciculus cuneatus (4), reticulospinal tract (5), vestibulospinal tract (6), spinothalamic tract (7), dorsal spinocerebellar tract (8). GM indicates gray matter.

Fig 2.

Picture of cervical spinal cord after partial right-sided hemisection and transplantation with unmodified fibroblasts (arrow, transplant outlined in red).

Fig 3.

Spinal cord images obtained 2 mm rostral to a right-sided partial hemisection and transplant of unmodified fibroblasts. A, T2-weighted (b = 0) MR image used as a guide to draw an ROI in the right DLF. An ROI in the surrounding buffer was also created to account for slight differences in temperatures. B and C, Corresponding tADC (B) and lADC (C) maps show atrophy in the right DLF (arrow) compared with the contralateral side (arrowhead). Spinal cord injury is expected to increase transverse water diffusion and decrease longitudinal water diffusion in injured white matter; these findings are reflected by a subtle increase in signal intensity in the right DLF on the tADC map (B) and a decrease in signal intensity in the right DLF on the lADC map (C) compared with the signal intensity of the undamaged left DLF.

Fig 4.

Spinal cord images obtained at the level of right-sided partial hemisection and transplant of BDNF-expressing fibroblasts. A, T2-weighted (b = 0) image used as a guide to draw ROIs in the transplant (TP), bilateral corticospinal tracts (CST), and left DLF. B and C, Corresponding tADC (B) and lADC (C) maps. Analysis of the transplant showed diffusion to be close to isotropic.

Fig 5.

MR image and corresponding histologic image of spinal cord at level of right-sided partial hemisection and unmodified fibroblast transplant. A, T2-weighted (b = 0) image with remaining right-sided white matter (WM) and remaining gray matter (GM) outlined. TP indicates transplant. B, Photomicrograph of corresponding histologic section (toluidine blue stain, original magnification ×16, bar = 1.0 mm), also with white matter (WM) and gray matter (GM) outlined, demonstrates close correspondence to MR image in A. TP indicates transplant

Fig 6.

Histologic images of BDA anterograde of labeled axons in the right RST. A and B, Normal images at magnification X16 (A, bar = 1.0 mm) and at magnification X50 (B, bar = 0.1 mm) show numerous labeled axons in the right RST (arrow). There are a few labeled axons on the contralateral side (arrowhead in A), as approximately 10% of RST axons do not cross the midline. For orientation, a slit was made in the ventral left white matter before sectioning. C–E, Histologic images of the right RST in an animal with an unmodified fibroblast graft at magnification X50 (bar = 0.1 mm) obtained at 3.5 mm rostral (C), 1.5 mm rostral (D), and 1.0 mm rostral (E) to the transplant. Note that the number of labeled axons (arrow) decreases as images are obtained closer to the transplant.

Fig 7.

RST axonal dieback. Graph indicates increasing percentage loss of axons as counts are obtained closer to the grafts. Axon count at 4.0 mm rostral to graft is reference value (0% loss). Note that the normal animal (▴) has a small loss of axons due to the normal exiting of RST axons in the cervical spinal cord.

Fig 8.

A–C, Graphs of ADC values and anisotropy index (AI) in the right DLF both rostral and caudal to the transplants. * indicates both animals with unmodified fibroblasts and those with BDNF-expressing fibroblasts were significantly different from normal controls; #, unmodified fibroblasts significantly different from normal controls; ∼, BDNF-expressing fibroblasts significantly different from normal controls; +, unmodified fibroblasts significantly different from BDNF-expressing fibroblasts. A, The tADC values in the right DLF become significantly elevated in white matter adjacent to the grafts, both rostral and caudal. These values become closer to normal as measurements are obtained further from injury site. B, The lADC values in the right DLF become significantly decreased in white matter adjacent to the grafts compared with those of normal controls, both rostral and caudal. These values become closer to normal as measurements are obtained further from injury site. C, Anisotropy index (AI) values become significantly elevated (decreased anisotropy) in white matter adjacent to the grafts compared with those in normal controls, both rostral and caudal. These values become closer to normal as measurements are obtained further from injury site.

Fig 9.

Scatterplots of ADC values and anisotropy index versus axonal dieback A and B, Scatterplots of tADC and lADC versus percentage axonal dieback show that as axonal loss increases closer to the injury site, tADC increases and lADC decreases. C, Scatterplot of anisotropy index (tADC/lADC ratio) versus percentage axonal dieback shows that the changes in ADC values seen with increasing axonal loss produce an increasing anisotropy index, indicating decreasing anisotropic water diffusion.

Fig 10.

Recovery of function in the affected right hindlimb occurred in all animals over time as measured with a modified BBB scale. At 12 weeks, the animals with BDNF-expressing fibroblast grafts had significantly better scores than those of animals with unmodified fibroblast grafts (+ indicates P < .05). The difference at 12 weeks is small, however, and likely not functionally significant.