Nonhomogeneous Gadolinium Retention in the Cerebral Cortex after Intravenous Administration of Gadolinium-based Contrast Agent in Rats and Humans

Abstract

Background

Gadolinium retention after repeated gadolinium-based contrast agent (GBCA) exposure has been reported in subcortical gray matter. However, gadolinium retention in the cerebral cortex has not been systematically investigated.

Purpose

To determine whether and where gadolinium is retained in rat and human cerebral cortex.

Materials and Methods

The cerebral cortex in Sprague-Dawley rats treated with gadopentetate dimeglumine (three doses over 4 weeks; cumulative gadolinium dose, 7.2 mmol per kilogram of body weight; n = 6) or saline (n = 6) was examined with antemortem MRI. Two human donors with repeated GBCA exposure (three and 15 doses; 1 and 5 months after exposure), including gadopentetate dimeglumine, and two GBCA-naive donors were also evaluated. Elemental brain maps (gadolinium, phosphorus, zinc, copper, iron) for rat and human brains were constructed by using laser ablation inductively coupled plasma mass spectrometry.

Results

Gadopentetate dimeglumine–treated rats showed region-, subregion-, and layer-specific gadolinium retention in the neocortex (anterior cingulate cortex: mean gadolinium concentration, 0.28 µg ∙ g⁻¹ ± 0.04 [standard error of the mean]) that was comparable (P > .05) to retention in the allocortex (mean gadolinium concentration, 0.33 µg ∙ g⁻¹ ± 0.04 in piriform cortex, 0.24 µg ∙ g⁻¹ ± 0.04 in dentate gyrus, 0.17 µg ∙ g⁻¹ ± 0.04 in hippocampus) and subcortical structures (0.47 µg ∙ g⁻¹ ± 0.10 in facial nucleus, 0.39 µg ∙ g⁻¹ ± 0.10 in choroid plexus, 0.29 µg ∙ g⁻¹ ± 0.05 in caudate-putamen, 0.26 µg ∙ g⁻¹ ± 0.05 in reticular nucleus of the thalamus, 0.24 µg ∙ g⁻¹ ± 0.04 in vestibular nucleus) and significantly greater than that in the cerebellum (0.17 µg ∙ g⁻¹ ± 0.03, P = .01) and white matter tracts (anterior commissure: 0.05 µg ∙ g⁻¹ ± 0.01, P = .002; corpus callosum: 0.05 µg ∙ g⁻¹ ± 0.02, P = .001; cranial nerve: 0.02 µg ∙ g⁻¹ ± 0.01, P = .004). Retained gadolinium colocalized with parenchymal iron. T1-weighted MRI signal intensification was not observed. Gadolinium retention was detected in the cerebral cortex, pia mater, and pia-ensheathed leptomeningeal vessels in two GBCA-exposed human brains but not in two GBCA-naive human brains.

Conclusion

Repeated gadopentetate dimeglumine exposure is associated with gadolinium retention in specific regions, subregions, and layers of cerebral cortex that are critical for higher cognition, affect, and behavior regulation, sensorimotor coordination, and executive function.

© RSNA, 2019

Online supplemental material is available for this article.

See also the editorial by Kanal in this issue.

Summary

Repeated exposure to intravenous gadopentetate dimeglumine, a linear gadolinium-based contrast agent, is associated with nonhomogeneous retention of gadolinium in rat and human cerebral cortex.

Key Results

• ■ Gadolinium retention in rat cerebral cortex varied by region, subregion, and layer, with peak concentration in the neocortex (anterior cingulate cortex: 0.28 µg ∙ g⁻¹) that was comparable (P > .05) to concentration in the allocortex (piriform cortex: 0.33 µg ∙ g⁻¹; hippocampus dentate gyrus: 0.24 µg ∙ g⁻¹) and subcortical structures (caudate-putamen: 0.29 µg ∙ g⁻¹; reticular nucleus of the thalamus: 0.26 µg ∙ g⁻¹).

• ■ Retained gadolinium in the rat neocortex and subcortical gray matter colocalized with parenchymal iron.

• ■ In two human brains, gadolinium retention was detected in the cerebral cortex, pia mater, and pial-ensheathed leptomeningeal vessels.

Introduction

There is mounting evidence that exposure to intravenous gadolinium-based contrast agents (GBCAs) is associated with gadolinium retention in subcortical regions of the brain. However, there has been much less attention to gadolinium retention in the cerebral cortex. The cerebral cortex has critical functions, including higher cognition, sensorimotor coordination, affect and behavior regulation, executive function, and consciousness (1). Moreover, cerebral cortical function is sensitive to disruption by diverse neurotoxicants, including lanthanide series elements and other heavy metals (lead, mercury, manganese) (2,3).

GBCAs are widely used in conjunction with MRI and have highly favorable safety profiles (4,5). Recent MRI studies have focused attention on gadolinium retention in subcortical gray matter nuclei in adult and pediatric patients with repeated intravenous GBCA exposure and normal renal function (4–12). Alterations in T1-weighted MRI signal intensity have been reported in various subcortical structures (globus pallidus, putamen, thalamus, pons, dentate nuclei) after intravenous GBCA exposure, especially following repeated administration of linear class GBCAs (5,13,14). Gadolinium retention in subcortical brain regions has been confirmed in humans and laboratory animals by elemental solution analysis, transmission electron microscopy, energy-dispersive x-ray spectroscopy, and laser ablation inductively coupled plasma mass spectrometry (LA-I;lkjhgtbnm,./CP-MS) (10,12,15–19).

On the basis of recent research from our laboratory, we hypothesized that gadolinium is retained and nonhomogeneously distributed in the cerebral cortex after repeated intravenous GBCA exposure (20). Furthermore, we hypothesized that cortical gadolinium retention may vary by region, subregion, layer, and local pathology, and additionally, colocalizes with parenchymal iron enrichment. The aim of this study was to determine whether and where gadolinium is retained in rat and human cerebral cortex using LA-ICP-MS imaging, an ultrasensitive elemental-isotopic tissue mapping technique.

Materials and Methods

Animal Subjects

Adult female Sprague-Dawley rats (n = 12; Charles River Laboratories, Wilmington, Mass), 9–10 weeks of age, were group-housed at the Animal Science Center, Boston University School of Medicine (BUSM), and examined by antemortem MRI and postmortem elemental brain mapping (Fig 1). Strain, sex, age, and GBCA exposure were selected to harmonize with recent research (21). Rats were randomly assigned to one of two groups: GBCA-exposed (n = 6) treated with three intravenous injections of gadopentetate dimeglumine (Bayer Vital GmbH, Germany; 2.4 mmol/kg per dose, three doses administered over 4 weeks, total gadolinium: 7.2 mmol/kg). The daily dose was selected to harmonize with recent preclinical studies (21) and represents four human equivalent GBCA doses after adjustment for total body surface (22). GBCA-naive (n = 6) treated with three volume-matched saline injections over 4 weeks. One week after the final gadopentetate injection, rats were sacrificed by CO₂ asphyxiation, transcardially perfused with 200 mL phosphate-buffered saline to remove blood from the cerebrovasculature by fluid-exchange exsanguination, followed by perfusion fixation (200 mL 4% (w/v) paraformaldehyde [PFA]). Harvested brains were submerged in 4% (w/v) PFA for 24 hours, then processed for neuropathologic examination and elemental brain mapping by LA-ICP-MS imaging. The study was approved by the Institutional Animal Care and Use Committee, BUSM.

Figure 1:

Flowchart shows experimental design. Two groups of healthy young adult female Sprague-Dawley rats were treated with three intravenous injections of gadopentetate dimeglumine (2.4 mmol per kilogram of body weight intravenously for each of three doses; cumulative total dose: 7.2 mmol/kg) or an equal volume of saline. All rats were examined with 3.0-T MRI before the first injection (pre-exposure) and 1 week after the last injection (postexposure). Rats were sacrificed and exsanguinated by means of open-thorax transcardial saline perfusion to remove blood from the cerebrovasculature. Brains were subsequently harvested and processed for neuropathologic examination and elemental-isotopic brain mapping. Cu = copper, Fe = iron, Gd = gadolinium, P = phosphorus, T1W = T1 weighted, vol = volume, Zn = zinc.

Human Subjects

Postmortem superior frontal cortex was obtained from two brain donors with repeated exposure to intravenous GBCAs and normal renal function (Table E1 [online]). Case 1 was that of a 49-year-old man with a diagnosis of a single traumatic brain injury sustained 245 days before death who received three 10-mmol doses of gadopentetate dimeglumine at days 243, 238, and 145 before death (total gadolinium dose, 30 mmol). Case 2 was that of a 35-year-old man with a diagnosis of glioblastoma multiforme who received four 5-mmol doses of gadopentetate dimeglumine 3 years prior to death; four 5-mmol doses of gadopentetate dimeglumine and one 7-mmol dose of gadobutrol 2 years prior to death; one 10-mmol dose of gadopentetate dimeglumine, two doses of gadobutrol (7.5, 10 mmol), and three doses of gadobenate dimeglumine (5, 5, and 10 mmol) in the year preceding death (total gadolinium dose, 94.5 mmol) (last dose, 1 month before death). Control superior frontal cortex tissue was obtained from two brain donors without GBCA exposure (case 3: 29-year-old man with early-stage chronic traumatic encephalopathy [CTE]; case 4: 18-year-old man without a neuropathologic diagnosis; Table E1 [online]). Human brain specimens were procured from National Institutes of Health (case 1) and Boston University Alzheimer’s Disease and CTE Center brain bank (cases 2–4). Clinical histories and GBCA exposure were determined by review of available medical records. The study was approved by Institutional Review Boards at and compliant with Material Transfer Agreements between BUSM and the National Institutes of Health.

MRI Evaluation

Rat MRI examinations (3.0-T Achieva Philips System, Cleveland, Ohio) were acquired (N.H., 15+ years experience) prior to the first injection (pre-exposure) and 1 week after the third of three injections (postexposure) of gadopentetate dimeglumine (n = 6) or saline (n = 6) (Fig 1). During MRI, rats were anesthetized with isoflurane (induction, 3.5% (v/v); maintenance, 1.5%–2% (v/v)) and inserted into a 16-element knee coil. T1-weighted images were acquired by using a multishot turbo spin-echo sequence: repetition time msec/echo time msec, 700/12; echo train length, three; number of averages, 12; acquisition matrix, 200 × 198; field of view, 40 × 40 mm²; zero-filling in-plane resolution, 0.1 × 0.1 mm²; and slice thickness, 2 mm. Human brain MRI parameters are listed in Table E2 (online). MRI for case 1 was performed on ex vivo brain blocks (A.D.G., L.L.L.).

Neuropathology and Elemental Brain Mapping (LA-ICP-MS Imaging)

Human brains were processed for comprehensive neuropathology (23) (V.E.A., B.R.H., 20+ years experience; A.C.M., 40+ years experience). Adjacent sections of paraffin-embedded rat brain were processed for conventional histochemistry (20,21). Rat brain anatomy and nomenclature followed the Paxinos and Watson atlas (24). Elemental and isotopic brain mapping (n = 5 per group) was performed by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) imaging, Boston University Center for Biometallomics (O.M., 15+ years; N.L., 25+ years), using a laser ablation system (LSX-213, Teledyne-CETAC, Omaha, Neb) hyphenated to a quadrupole ICP-mass spectrometer (iCAP-Q, Thermo Fisher Scientific, Waltham, Mass): spot size, 50 µm; line separation, 0 µm; scan speed, 100 µm·sec⁻¹. Analytical calibrations were performed with reference standards (SRM-612, NIST, Gaithersburg, MD; gadolinium-spiked gelatin 10% (w/v)). The ¹⁵⁸Gd-LA-ICP-MS imaging analyses were calibrated for total gadolinium concentration. The gelatin standards were slice with a cryotome (CM1850, Leica Biosystems) to a thickness of 10 μm and were imaged by using the same parameters for laser ablation platform and mass spectrometer as for all human and rat brain sections. Calibration curves: seven points, 0.0–12.8 μg · g⁻¹. LA-ICP-MS data sets were exported to a customized MATLAB program (Mathworks, Natick, Mass) and ImageJ (National Institutes of Health open-source software) for anatomic mapping and analytical quantitation.

Statistical Analysis

Manual contours were drawn by two independent reviewers on elemental brain maps to calculate gadolinium tissue concentrations for 15 regions of interest (L.E.G., with more than 35 years of experience in experimental pathology; N.H., with more than 15 years in preclinical MRI) based on the rat brain atlas of Paxinos and Watson (24). Gadolinium concentration was calculated for each region. Regional group averages were reported as mean concentrations ± standard errors of the mean. Paired t tests were used to compare different brain regions. P < .05 was considered to represent a significant difference. The intraclass correlation coefficient between two readers was 0.99, with a confidence interval of 0.984, 0.994, which indicated excellent agreement between the two readers.

Results

Evaluation of Gadolinium Retention in Rat Cerebral Cortex

MRI examinations performed before the first injection and 1 week after the last of three injections of gadopentetate dimeglumine (dose: 2.4 mmol/kg; total gadolinium dose, 7.2 mmol/kg; 12 human dose equivalents; tissue elimination half-life, rats: approximately 17 minutes; Fig 1) or saline revealed normal structural anatomy throughout the brain (Fig 2). In contrast to previous rat studies involving larger cumulative doses of gadopentetate (21,25–27), we did not detect T1-weighted MRI signal intensification in any brain region. Postmortem neuropathologic examination showed normal brain structure and cytologic organization in both groups (Fig 3, left panels). We conducted LA-ICP-MS elemental mapping analysis of postmortem brains from the same rats (Figs 3, 4, Figs E1–E4 [online]). We detected distinct elemental distribution patterns for phosphorus, zinc, and copper (Fig E1 [online]) that differentiated gray matter from white matter and revealed element-specific enrichment that varied by region and subregion (Appendix E1 [online]).

Figure 2:

Representative same-subject images from T1-weighted MRI (at 3.0 T) of the brain before the first injection (pre-exposure, left column) and 1 week after the third injection (postexposure, right column) of intravenous gadopentetate dimeglumine. Coronal sections are at the level of, A, the anterior cingulate cortex, B, the rostral hippocampus, and, C, the cerebellum. No evidence of T1-weighted signal intensification was noted in the cerebral cortex or subcortical structures, including the dentate nucleus.

Figure 3:

Images show gadolinium retention in rat cerebral cortex after repeated gadopentetate dimeglumine administration. Coronal sections at the level of, A, the anterior cingulate cortex, and, B, the rostral hippocampus. Left: Luxol fast blue hematoxylin-eosin staining. Right: Laser ablation inductively coupled mass spectrometry imaging brain maps for gadolinium. Symbols in, A: hashed arcs = olfactory area, hashed oval = caudate-putamen, hashed area (dorsomedial) = rostral cingulate cortex (A32; dashed arrow [filled] = layer II), arrow = insular cortex, * = corpus callosum; star = anterior commissure. Symbols in, B: hashed arc = piriform cortex, cortical amygdala (layer I); hashed circle = hypothalamus, arcuate nucleus; hashed oval = entopeduncular nucleus; hashed area (dorsomedial) = midcingulate cortex (A30, A29); hashed area (lateral) = caudate-putamen, globus pallidus; arrow = thalamus, reticular nucleus; arrowhead (unfilled) = hippocampus CA1; dashed arrow (unfilled) = dentate gyrus; * = corpus callosum; darts = choroid plexus; diamond = internal capsule. Abbreviations (in italics for white matter tracts): 1,2 = cortex layers; 3,4v = ventricles; 7n = facial nucleus; ac = anterior cingulate cortex (dorsal, ventral); acm = anterior commissure; ar = arcuate nucleus; cc = corpus callosum; cp = caudate-putamen; dg = dentate gyrus; ec = external capsule; en = entopeduncular nucleus; fi = fimbria; fm = forceps minor; g = globus pallidus; hb = habenula; hf = hippocampus; in = insular cortex (agranular, dysgranular, granular); ic = internal capsule; is = interhemispheric sulcus; m₁,₂ = motor cortex; na = nucleus accumbens; ot = optic tract; pca = posterior cortical amygdala; pi = piriform cortex; rs = rhinal sulcus; s₁,₂,_bf = somatosensory cortex; thal = thalamus; tu = olfactory tubercle.

Figure 4:

Images show that gadolinium retention in rat cerebral cortex colocalizes with parenchymal iron enrichment. Coronal section levels at the level of, A, the anterior cingulate cortex, and, B, the rostral hippocampus. Laser ablation inductively coupled mass spectrometry imaging brain maps for parenchymal iron (Fe, left) and gadolinium (Gd, right). Symbols in, A: hashed arc = olfactory area; arrow (filled) = insular cortex (agranular); arrow (unfilled) = piriform cortex (layer I); dashed arrow (filled) = anterior cingulate cortex (layer II); dashed arrows (unfilled) = motor, somatosensory cortices; * = corpus callosum; star = anterior commissure. Symbols in, B: hashed arc = piriform cortex, cortical amygdala (layer I); hashed circle = arcuate nucleus; hashed oval = entopeduncular nucleus; hashed area (dorsomedial) = midcingulate cortex (A30, A29); circle = habenula; arrows = ventral posterior thalamus; dashed arrows (filled) = hippocampus CA1; dashed arrows (unfilled) = dentate gyrus; * = corpus callosum; darts = choroid plexus; diamond = internal capsule. Abbreviations (in italics for white matter tracts): 1,2 = cortex layers; 3,4v = ventricles; 7n = facial nucleus; ac = anterior cingulate cortex (dorsal, ventral); acm = anterior commissure; ar = arcuate nucleus; cc = corpus callosum; cp = caudate-putamen; dg = dentate gyrus; ec = external capsule; en = entopeduncular nucleus; fi = fimbria; fm = forceps minor; g = globus pallidus; hb = habenula; hf = hippocampus; in = insular cortex (agranular, dysgranular, granular); ic = internal capsule; is = interhemispheric sulcus; m₁,₂ = motor cortex; na = nucleus accumbens; ot = optic tract; pca = posterior cortical amygdala; pi = piriform cortex; rs = rhinal sulcus; s₁,₂,_bf = somatosensory cortex; thal = thalamus; tu = olfactory tubercle.

We detected substantial gadolinium retention in neocortex, allocortex, and subcortical structures in GBCA-treated rats (Figs 3, 4; Table 1; Figs E2, E3 [online]). We confirmed the elemental identity of the LA-ICP-MS signal attributed to gadolinium by isotopic mapping and stable isotope ratio analysis (Fig E2 [online]). In addition, we detected identical neuroanatomic localization of three naturally occurring stable gadolinium isotopes (¹⁵⁶Gd, ¹⁵⁷Gd, ¹⁵⁸Gd) by LA-ICP-MS brain mapping, and confirmed by the same technique that gadolinium isotopic ratios reflected natural abundance (Fig E2 [online]). LA-ICP-MS imaging analysis showed that gadolinium signal in brains from saline-treated control rats (GBCA-naive) was below background concentration (Fig E4 [online]).

Regional Gadolinium Concentrations in Brains of GBCA-treated Rats

Retained gadolinium distributed bilaterally and nonhomogeneously in discrete regions, subregions, and layers of cerebral cortex and subcortical gray matter (Figs 3, 4; Table; Figs E2, E3 [online]). Retained gadolinium was enriched in agranular cortex (anterior cingulate cortex, ACC: A32d/v; insular cortex) and ventral paleocortex (piriform cortex, olfactory tubercle) and showed striking laminer-specificity (ACC layer II: mean gadolinium concentration, 0.28 µg ∙ g⁻¹ ± 0.04 [standard error of the mean]; piriform cortex layer I: 0.33 µg ∙ g⁻¹ ± 0.04) with peak gadolinium concentrations that were as high as those in the caudate-putamen (0.29 µg ∙ g⁻¹ ± 0.05). By comparison, gadolinium retention in motor and somatosensory cortices was modest and diffuse. At the rostral hippocampus level, cingulate cortex gadolinium retention (A29c: 0.32 µg ∙ g⁻¹ ± 0.10) was comparable to that in the hippocampus (CA1: 0.17 µg ∙ g⁻¹ ± 0.04; dentate gyrus: 0.24 µg ∙ g⁻¹ ± 0.04), thalamic reticular nucleus (0.26 µg ∙ g⁻¹ ± 0.05), and choroid plexus (0.39 µg ∙ g⁻¹ ± 0.10). Retained gadolinium in the brainstem facial nucleus (0.47 µg ∙ g⁻¹ ± 0.10) was the highest in the brain and significantly greater than that in the cerebellar cortex (0.17 µg ∙ g⁻¹ ± 0.03; P = .03) or brainstem white matter (spinotrigeminal tract: 0.02 µg ∙ g⁻¹ ± 0.01; P = .01). Major white matter tracts exhibited low gadolinium retention throughout the brain (anterior commissure: 0.05 µg ∙ g⁻¹ ± 0.01; corpus callosum: 0.05 µg ∙ g⁻¹ ± 0.02; internal capsule: 0.04 µg ∙ g⁻¹ ± 0.01).

Retained Gadolinium Colocalizes with Parenchymal Iron Enrichment

Cortical and subcortical gadolinium retention paralleled parenchymal iron distribution (Fig 4; Fig E3 [online]). Parenchymal iron was enriched in discrete layers of agranular cortex (ACC, layer II), periallocortex (insular cortex, agranular subdivision), paleocortex (olfactory tubercle, piriform cortex, posterolateral cortical amygdala; layer I), and archicortex (hippocampus stratum pyramidale, CA1 subregion; dentate gyrus, granule cell layer). Subcortical structures exhibiting parenchymal iron enrichment including caudate-putamen, globus pallidus (and entopeduncular nucleus), somatosensory thalamus, habenula (epithalamus), hypothalamic arcuate nucleus, and choroid plexus. We detected parenchymal iron enrichment in the cerebellum, dentate nucleus, brainstem nuclei (cochlear, vestibular, facial nuclei), and choroid plexus. White matter was relatively devoid of iron. LA-ICP-MS mapping for gadolinium in the same brain sections from gadopentetate-treated rats revealed a gadolinium retention pattern that closely paralleled cortical and subcortical parenchymal iron enrichment.

Evaluation of Gadolinium Retention in Human Cerebral Cortex

Postmortem MRI of superior frontal cortex from case 1, a 49-year-old man in chronic recovery after a traumatic brain injury who underwent three MRI examinations with intravenous gadopentetate dimeglumine (total gadolinium, 30 mmol; postexposure interval, 144 days), showed normal gross and microscopic anatomy of the cerebral cortex and underlying subcortical white matter without evidence of hemorrhage, hematoma, contusion, or tissue necrosis in the region of interest (Fig 5, A, B). LA-ICP-MS imaging analysis of adjacent sections of superior frontal cortex revealed distinct elemental distribution patterns (iron: Fig 5, C); phosphorus, zinc, copper: Fig E5 [online]) that differentiated cortical gray matter from subcortical and deep white matter. Zinc showed enrichment in superficial cortical laminae and subcortical white matter. By contrast, analytical LA-ICP-MS mapping of gadolinium in the same section revealed a markedly different distribution pattern (Fig 5, D). Retained gadolinium was more pronounced in cortical gray matter compared with underlying white matter (subcortical U-fibers, deep white matter). Within the cortical mantle, gadolinium retention was relatively sparse in molecular layer I and enriched in deeper layers. Gadolinium retention was prominent in the sulcal fundus adjacent to the gray-white matter interface. Retained gadolinium in the sulcal fundus did not correlate with evidence of focal hemorrhage or hemosiderin-laden macrophages. However, we cannot exclude the possibility of microhemorrhage residuum.

Figure 5:

Images show gadolinium retention in the human cerebral cortex in case 1, a 49-year-old man with history of traumatic brain injury and three intravenous injections of gadopentetate dimeglumine (total, 30 mmol; postexposure interval, 145 days). A, Image from postmortem MRI of blocked brain specimen shows region of interest (box), superior frontal cortex. B, Low-power photomicrograph. Luxol fast blue hematoxylin-eosin staining. C, Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) imaging brain map for iron (Fe) in adjacent brain section reflects hematogenous (intravascular) and nonhematogenous (parenchymal) iron. D, LA-ICP-MS imaging brain map for gadolinium (Gd) in the same section of superior frontal cortex as in C. Abbreviations: ccx = cerebral cortex; dwm = deep white matter; G_c = gyral crest; S_b,f = sulcal bank, fundus; suf = subcortical U fibers; I–VI = cortex layers.

Antemortem MRI and postmortem neuropathology of superior frontal cortex in case 2, a 35-year-old man with glioblastoma multiforme who underwent 15 MRI examinations with linear and macrocyclic glioblastoma multiforme (gadopentetate dimeglumine, 50 mmol; gadobenate dimeglumine, 20 mmol; gadobutrol, 24.5 mmol; total gadolinium, 94.5 mmol; postexposure interval, 33 days), showed normal anatomy without evidence of hemorrhage, hematoma, or necrosis in the region of interest (Fig E6 [online]). LA-ICP-MS imaging analysis revealed distinct distribution patterns of iron as well as phosphorus, zinc, and copper (Fig E5 [online]). As in case 1, gadolinium retention was relatively sparse in molecular layer I and relatively enriched in deeper cortical layers, though with less obvious laminar differentiation and no apparent concentration at the sulcal fundus. Retained gadolinium was also detected in pia mater and associated with pial-ensheathed leptomeningeal blood vessels (Fig E6 [online]).

Case 3, a 29-year-old man with early-stage chronic traumatic encephalopathy who underwent a single MRI examination without intravenous GBCA (Fig E7 [online]). Antemortem MRI examination and postmortem neuropathological analysis of the superior frontal cortex showed normal macroscopic anatomy without evidence of focal pathology in the region of interest. LA-ICP-MS imaging analysis revealed distinct elemental distribution patterns for phosphorus, iron, zinc, and copper. We did not detect gadolinium in the superior frontal cortex by LA-ICP-MS imaging analysis in this case nor in a second GBCA-naive control subject (case 4).

Discussion

Mounting evidence indicates that intravenous gadolinium-based contrast agent (GBCA) exposure may be associated with gadolinium retention in subcortical regions of the brain. However, gadolinium retention in the cerebral cortex has not been systematically investigated. We sought to address this knowledge gap by conducting laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) elemental brain mapping in rat and humans with repeated exposure to gadopentetate dimeglumine, a linear-ionic class GBCA. Our results show that gadolinium is retained in the cerebral cortex. We found that peak gadolinium concentrations in specific regions of rat neocortex (anterior cingulate cortex: 0.28 µg ∙ g⁻¹ ± 0.04) and allocortex (piriform cortex: 0.33 µg ∙ g⁻¹ ± 0.04; hippocampus dentate gyrus: 0.24 µg ∙ g⁻¹ ± 0.04) were as high as (P > .05) gadolinium-retaining subcortical structures (caudate-putamen: 0.29 µg ∙ g⁻¹ ± 0.05; thalamus, reticular nucleus: 0.26 µg ∙ g⁻¹ ± 0.05; brainstem vestibular nucleus: 0.24 µg ∙ g⁻¹ ± 0.04; choroid plexus: 0.39 µg ∙ g⁻¹ ± 0.10). However, gadolinium retention in the cerebral cortex was not homogeneous and varied by cortical region, subregion, and layer. These results provide clues regarding gadolinium uptake and retention in the brain. Sparse gadolinium retention in cortical layer I may reflect the high neuropil density and paucity of neuronal soma that characterize this layer (Appendix E2 [online]). This interpretation is consistent with low levels of gadolinium retention in white matter (corpus callosum: 0.05 µg ∙ g⁻¹ ± 0.02; anterior commissure: 0.05 µg ∙ g⁻¹ ± 0.01). These results indicate that the cerebral cortex serves as a brain reservoir of retained gadolinium.

We hypothesized that brain iron metabolism underpins cortical gadolinium retention. This hypothesis is difficult to test in human brain as parenchymal iron is confounded by retained blood in the cerebrovasculature. To address this issue, we conducted LA-ICP-MS imaging analysis on brain sections from rats that were transcardially perfused with saline, a procedure that removes blood from the cerebrovasculature. Our results provide evidence that gadolinium retention is greatest in brain regions that exhibit parenchymal iron enrichment. GBCA-related gadolinium retention has been reported in subcortical structures that accumulate nonheme, protein-bound ferric iron (3,21,28,29). These same iron-enriched subcortical regions also demonstrate abnormal MRI signals in rare inherited iron-related neurologic disorders (neurodegeneration with brain iron accumulation) (30) and neurodegenerative diseases such as Alzheimer disease (31). The overlapping distribution pattern of gadolinium retention and iron enrichment is difficult to explain based on ventricular proximity, passive diffusion, or glymphatic trafficking. We speculate that gadolinium-iron overlap may reflect regional, subregional, and layer specialization that bears on specific carriers, transporters, and trafficking mechanisms by which these two metals cross the blood-brain barrier and distribute within the brain. Recent research indicates that gadolinium may also coprecipitate with calcium or phosphorus (18,32–34). Gadolinium retention in choroid plexus implicates alternative entry via the blood-cerebrospinal fluid barrier and ventricular transit, while gadolinium retention in pia mater and pial-ensheathed leptomeningeal blood vessels suggests a role for paravascular transit and glymphatic distribution.

Several points deserve mention. First, our finding of nonhomogeneous gadolinium retention in rat cerebral cortex is based on exposure to gadopentetate, a linear-ionic class GBCA. Caution is warranted with regard to generalizing our results to other GBCAs. Second, our finding that retained gadolinium localizes to specific cortical layers provides a powerful strategy for investigating potential neurotoxicity. For example, the neurobiologic impact of retained gadolinium can be directly evaluated by testing neurophysiologic functions in ex vivo slice culture (Appendix E3 [online]). Third, we did not detect T1-weighted MRI signal intensification in any brain region, including cerebral cortex or dentate nucleus. Establishing a quantitative, mechanistic basis for understanding how gadolinium tissue concentrations relate to MRI signal changes in specific brain regions is needed to support meaningful interpretation of preclinical and clinical imaging results. Fourth, we found that peak gadolinium concentration in the rat cerebral cortex was comparable to subcortical nuclei. It is important to underscore that peak gadolinium levels detected in human and rat brains were low (<0.5 µg · g⁻¹) and were not associated with neuropathology (21,25,27,33). However, the absence of overt neuropathology does not provide proof of normal neurophysiologic function nor preclude potential for latent pathology, genotoxicity, neurodevelopmental abnormalities, or neuropsychiatric sequelae. Our results support multiple gadolinium trafficking mechanisms and modulation by local and systemic factors.

Study limitations of the rat study include the following: (a) single species, strain, sex, age, agent, high-dose regimen, and single postexposure interval; (b) sample preparation, processing, and analysis; and (c) slice thickness and absence of quantitative MRI data. Limitations of the human study include the following: (a) small number and heterogeneity of cases; (b) male sex bias; (c) limited sampling; (d) potential confounders resulting from brain pathology, systemic comorbidities, genetic factors, and treatment effects; (e) variation in contrast agent, class, dosing regimen, cumulative exposure, and postexposure interval; (f) inability to differentiate contributions from multiagent exposures (in case 2); (g) variation in MRI timing, quality, and protocols; (h) methodologic issues (specimen preparation, processing, and analysis); (i) absence of information regarding gadolinium chemical state, binding profile, chelation status, subcellular localization, and physicochemical dynamics; and (j) inherent limitations of clinicopathologic correlation to establish mechanistic causality.

In conclusion, gadolinium-based contrast agent (GBCA) exposure, specifically gadopentetate dimeglumine, is associated with nonhomogeneous gadolinium retention in discrete regions, subregions, and layers of the cerebral cortex that are critical for cognition, affect, and behavior regulation, sensorimotor integration, and executive function. Our study suggests multiple mechanisms for gadolinium transport, distribution, and retention in the brain that may challenge current concepts, research priorities, and clinical considerations regarding GBCA-related gadolinium exposure and potential neuropsychiatric sequelae.