Estimating blood and brain concentrations and blood-to-brain influx by magnetic resonance imaging with step-down infusion of Gd-DTPA in focal transient cerebral ischemia and confirmation by quantitative autoradiography with Gd-[(14)C]DTPA

Abstract

An intravenous step-down infusion procedure that maintained a constant gadolinium-diethylenetriaminepentaacetic acid (Gd-DTPA) blood concentration and magnetic resonance imaging (MRI) were used to localize and quantify the blood-brain barrier (BBB) opening in a rat model of transient cerebral ischemia (n=7). Blood-to-brain influx rate constant (K(i)) values of Gd-DTPA from such regions were estimated using MRI-Patlak plots and compared with the K(i) values of Gd-[(14)C]DTPA, determined minutes later in the same rats with an identical step-down infusion, quantitative autoradiography (QAR), and single-time equation. The normalized plasma concentration-time integrals were identical for Gd-DTPA and Gd-[(14)C]DTPA, indicating that the MRI protocol yielded reliable estimates of plasma Gd-DTPA levels. In six rats with a BBB opening, 14 spatially similar regions of extravascular Gd-DTPA enhancement and Gd-[(14)C]DTPA leakage, including one very small area, were observed. The terminal tissue-plasma ratios of Gd-[(14)C]DTPA tended to be slightly higher than those of Gd-DTPA in these regions, but the differences were not significant. The MRI-derived K(i) values for Gd-DTPA closely agreed and correlated well with those obtained for Gd-[(14)C]DTPA. In summary, MRI estimates of Gd-DTPA concentration in the plasma and brain and the influx rate are quantitatively and spatially accurate with step-down infusions.

Figure 1

Blood and plasma concentration time courses of Gd-DTPA (left ordinate) and Gd-[¹⁴C]DTPA (right ordinate), respectively. Data are given as mean±s.e.m. (n=7). The points for the Gd-DTPA represent the mean ΔR₁ and are plotted at the midpoint of the 2.5-min MRI sampling period. The Gd-[¹⁴C]DTPA points are for 2 to 3 secs sampling periods and are plotted accordingly. The changes in Gd-DTPA relaxation rates (ΔR₁), were measured in the superior sagittal sinus and approximate concentration. The Gd-[¹⁴C]DTPA data are from plasma samples obtained from blood drawn from the femoral artery. The scaling was such that the two sets of points fairly closely overlay each other. The mean levels were nearly constant from ~400 secs to the end of the sampling periods for both tracers.

Figure 2

Examples of blood–brain barrier (BBB) opening as visualized by the MRI central slice (A and D) and QAR (B and E) and the corresponding area of stroke injury as demarcated by histologic pallor (i.e., light staining; C and F). The data are from two different rats. The QAR images were inverted to make the dark leakage regions appear bright to match the hyperintensities that indicate contrast enhancement on magnetic resonance images. Panels A (K_i map) and B (corresponding ARG, central of the five that were cut from the tissue spanned by the 2-mm-thick MRI slice) show a BBB lesion that spreads across the preoptic area (PoA) into the striatum (Str) in one of the two rat brains. Large, white arrows point to the areas of abnormal brightness. Although not apparent from the figure, the degree of brightness on the ARG was not uniform and implies some variation in the BBB opening within this tissue. Panel C shows a CV-stained section obtained adjacent to the ARG of panel B with a large area of pallor. The latter was similar in location to the BBB lesion in panel B, but somewhat larger in size. Regions that normally lack a BBB (circumventricular organs or CVOs) are marked by thin arrows. Panels D and E are from a different animal and display a large area of BBB opening that comprised virtually contiguous parts of the PoA, Str, and cortex (PCx); they are marked by the lower, middle, and upper broad arrows, respectively. Panel F is the corresponding histology with a very large area of pallor that extended across all three ROIs. The latter also showed considerable swelling of the ischemic (right) hemisphere and midline displacement.

Figure 3

Examples of blood–brain barrier (BBB) opening as observed by an MRI central slice (A) and the five ARGs (B, D, F, H, and J) that spanned the 2-mm-thick MRI slice and the adjacent histologies (C, E, G, I, and K). Panel A is a K_i map. The two thick arrows indicate a sizable BBB lesion in a part of the PoA and Str (lower arrow) and a small area of BBB opening in the parietal cortex (upper arrow). The ARG images show that the PoA–Str area of Gd-¹⁴C-DTPA leakage extended through ~80% of the MRI slice (panels B, D, F, and H). The small cortical BBB lesion was observed only in ARGs, B and D (large arrow on the right), most prominently in panel D. The histologies show that the PoA–Str area of tissue pallor was large and visible through all five sections, whereas just a small, <1.0 mm-wide area of pallor was found within PCx (panels C and E, which correspond to ARGs, B and D, respectively). The latter implies that the sensitivity of MRI to small areas of the BBB opening was very good, with the step-down infusion, approaching that of QAR.

Figure 4

Example of a Patlak plot from a representative animal. The ordinate is the Gd-DTPA concentration in the tissue at time t, (C_t) divided by the plasma concentration at that time [C_a/(1–Hct)], where C_a is the concentration in whole blood within the superior sagittal sinus and Hct the hematocrit. The abscissa is labeled t_plot and represents the plasma concentration–time integral up to t divided by [C_a/(1–Hct)] at t. The contralateral points form a straight line with zero slope, indicating no detectable Gd-DTPA leakage. The points for the ipsilateral PoA, PCx, and Str formed straight lines with virtually identical K_i values, around 0.0024 mL/g per min. The y-intercepts represent the volume of rapid distribution of the MRCA-affected protons in plasma (V_p) and in other intravascular compartments (V_o). In this example, V_p+V_o was lowest for the contralateral brain (0.017 mL/g brain) and for Str (0.020 mL/g) and highest for PoA (0.032 mL/g). The plotted points extend along the x-axis, t_plot, up to nearly 1,000 secs, which was appreciably less than the real time of the experiment, ~1,200 secs.