Longitudinal MRI for the detection of a typical pattern of cerebral blood flow and tissue integrity changes after stroke in ischaemia tolerant rats

Introduction: The individual vulnerability to ischaemia has a significant impact on outcome in stroke patients. Therapeutic interventions could be much better tailored to the patient if information about the individual resistance to ischemic injury was available. Tolerance to ischaemia can be induced experimentally by „preconditioning“. One preconditioning model is the injection of 3-nitroproprionic acid (3NPA), a respiratory chain inhibitor. We used MRI in adult rats to characterize how preconditioning with 3NPA affects the development of a stroke lesion and if a typical „MRI signature“ of preconditioned tissue can be extracted from these data.

Methods: Adult rats received i.p. injections of 3NPA or saline and were subjected to transient middle cerebral artery occlusion (MCAO) three days later. MR images were acquired during the occlusion and after reperfusion (days 1, 4, 14). Quantitative CBF maps were obtained using flow-sensitive alternating inversion recovery (FAIR) with the QUIPSSII modification. In addition, diffusion-weighted (DWI) images, T1 and T2 maps and anatomical T2w images were obtained. Functional deficits were assessed using a modified neurologic deficit scale.

Results: At 15min after insertion of the occluding device, the extent of the initial lesion on apparent diffusion coefficient (ADC) maps was similar between groups. Over the next 30min of MCAO, however, there was a remarkable ADC recovery in the 3NPA group (Figure 1). During MCAO, preconditioned animals had fewer voxels with severely compromised CBF than controls (Figure 2). CBF on the unaffected side was lower in the 3NPA group during ischaemia. A similarly low CBF was also found in 3NPA- preconditioned animals without stroke (data not shown). An area of hyperperfusion could be observed in controls between 4 and 14 days after stroke, with a maximum in the infarct border zone. Such hyperperfusion was much less pronounced, infarcts were smaller and functional deficits less severe in the 3NPA group.

Conclusions: Even with a large initial lesion on ADC maps, tissue recovery was substantial in 3NPA-preconditioned animals, most likely due to better residual perfusion in the ischaemic area, and was associated with a better functional outcome. Lower CBF values on the unaffected hemisphere were observed after 3NPA preconditioning and could indicate vascular/metabolic adjustments of ischemia tolerance. Thus, MRI might have the potential to discriminate ischaemia tolerant from vulnerable brains.

Figure 1:A) Masks of the ischemic lesion from individual animals of the control or NPA-preconditioned group generated from ADC maps acquired 15 (upper panel) or 45 (Iower panel) minutes after the onset of MCAO. Masks are overlaid onto anatomical (T2w) images.Regions of overlap between individual masks are color coded. While the ischemic area is similar 15 and 45 min after onset of the vascular occlusion in controls, there is a dramatic recovery in the NPA group.B and C) Number of voxels with a significantly reduced ADC in both groups (mean +/- SE of mean). While the lesion depicted on ADC maps is of similar size 15 min after start of the occlusion (B), it is much smaller in NPAtreated animals at 45 min.

Figure 2A) Averaged CBF maps from control (con) and NPA-treated (npa) animals during MCAO (Day 0) and on days 1, 4 and 14. B) Voxel wise analysis of differences between the CBF maps of the two groups reveals significant clusters of higher (yellow) or Iower (red) CBF in the NPA group. The CBF reduction during ischemia is less pronounced with npa, while CBF is Iower on the contralateral side. There is hyperperfusion in the area of ischemia from days 1-14 in controls, but barely any with npa.C) averaged FA maps overlaid onto averaged ADC maps show loss of fiber tracts in controls but integrity of tracts in npa animals.D) Averaged T1 and T2 maps on days 1 and 14.E) Number of voxels with a CBF value below 30 ml/100g/min on day 0 (*p<0.05).F) Time course of voxels with an initial value below 30ml/100g/min (upper panel), or a significantly reduced ADC (Iower panel) suggest hyperperfusion and cystic necrosis in controls at the chronic stage.G) Number of voxels with a significantly increased T2 on day 14 as an estimate of final infarct size confirms neuroprotection with NPA.