7‐Tesla Magnetic Resonance Imaging for Brain Iron Quantification in Homozygous and Heterozygous PANK2 Mutation Carriers

Abstract

Pantothenate‐kinase–associated neurodegeneration (PKAN) is an autosomal recessive disorder characterized by iron deposits in basal ganglia. The aim of this study was to quantify iron concentrations of deep gray matter structures in heterozygous PANK2 mutation carriers and in PKAN patients using quantitative susceptibility mapping MRI. By determining iron concentration, we intended to find mutation‐specific brain parenchymal stigmata in heterozygous PANK2 mutation carriers in comparison to age‐matched healthy volunteers. We studied 11 heterozygous PANK2 gene mutation carriers (mean age: 43.4 years; standard deviation [SD]: 10.5), who were found to be clinically asymptomatic by neurological examination. These carriers were compared to 2 clinically affected PKAN patients 21 and 32 years of age and to 13 age‐matched, healthy controls (mean age: 39.7; SD, 13.6). Scanning was performed on a 7.0‐Tesla whole‐body scanner applying three‐dimensional susceptibility‐weighted gradient echo acquisitions. Susceptibility maps were calculated by threshold‐based k‐space division with single‐orientation acquisition. Magnetic susceptibility values, relative to the occipital white matter, were determined for the following regions of interest (ROI): globus pallidus (GP), thalamus, putamen, internal capsule (IC), caudate nucleus, substantia nigra (SN), and red nucleus. Heterozygous PANK2 mutation carriers did not show increased brain iron concentrations, compared to healthy controls (P > 0.05), in any of the examined ROIs. In PKAN patients, more than 3 times higher concentrations of iron were found in the GP, SN, and IC. Our results suggest that heterozygous mutations in PANK2 gene do not cause brain iron accumulation nor do they cause movement disorders.

The syndromes of neurodegeneration with brain iron accumulation (NBIA) are defined by a progressive hypo‐ and/or hyperkinetic movement disorder and pathological excess of iron deposition in the brain, particularly affecting the basal ganglia, mainly the globus pallidus (GP). Of these, pantothenate‐kinase–associated neurodegeneration (PKAN) is the most common form caused by mutations in the PANK2 gene. However, PKAN is an orphan disease with an estimated world‐wide prevalence of approximately 1 in 1 million.1

In NBIA syndromes, iron accumulation observed on gross pathological assessment as brown discoloration can be detected noninvasively as a prominent hypointensity in T2‐ and T2*‐weighted MRI.2

In PKAN, a specific MRI pattern of iron deposition, the “eye of the tiger sign” consistent with a central hyperintensity within a surrounding area of hypointensity in the anterior‐medial part of the GP, is thought to be virtually pathognomonic. However, emergence of this MRI pattern appears to be a dynamic process.3, 4 It has been shown that alterations in GP detected by MRI may precede the development of clinical signs5 (i.e., in asymptomatic carriers of homozygous mutations), but may also be absent in early disease stages.3, 6, 7 Additionally, the bright spot within the surrounding area of hypointensity may vanish over time.3

Recently, a 7‐Tesla (T) MRI study in 3 PKAN patients was performed using the method of field‐dependent R2 relaxivity increase.8 Iron concentration in GP of PKAN patients was more than doubled, compared to controls. Moreover, based on the differences in the field dependency of the relaxation rates, the researchers suggested that, in PKAN patients, iron deposited in forms other than ferritin also contributes to the MRI signal.

On the other hand, little is known about brain iron content in heterozygotes. Results of recent 3‐T MRI studies suggested no signs of increased brain iron content in the 13 examined heterozygous PANK2 mutation carriers.3 However, the latter study showed subtle differences in mean water diffusivity (MD) and fractional water diffusion anisotropy (FA) in GP, cerebral peduncles, and internal capsule (IC) between heterozygous carriers and controls, which may be indirect effects reflecting altered iron content. Indeed, correlations between MD, FA, and ferritin‐iron content were observed in an in vitro phantom study.9

Taking these results into account, the question remains of whether iron accumulation occurs in heterozygous PANK2 mutation carriers. To answer this question, iron quantification techniques using ultra‐high‐field MRI can be very helpful. Changes in magnetic susceptibility of tissue resulting from iron accumulation cause magnetic field shifts on the microscopic scale. These field shifts are more closely related to iron concentration than spin‐spin relaxation processes and can be detected in phase‐sensitized MRIs.2 Quantitative susceptibility mapping (QSM) techniques were developed to determine tissue susceptibility.10, 11, 12, 13 Ultra‐high‐field imaging using magnetic fields of B₀ ≥7 T is particularly sensitive to very small susceptibility changes because of the linear relationship between magnetic field strength and the field shift.14, 15, 16, 17 We used this advantage of ultra‐high‐field MRI to study iron concentration in PANK2 mutation carriers.

The aims of our study were to (1) quantify iron concentration in basal ganglia, IC, and thalami (THAs) of PKAN patients and heterozygous PANK2 mutation carriers by employing QSM with ultra‐high‐field MRI and (2) seek for differences between heterozygous PANK2 mutation carriers and healthy volunteers.

Methods

Participants

The study was approved by the relevant local ethic committees, and participants gave written informed consent before the study.

Two PKAN patients (both female, 21 and 32 years of age) and 11 clinically unaffected relatives of these and other PKAN patients (4 female and 7 male; mean age: 43.4 years; strandard deviation [SD]: 10.5) were recruited. Patients were examined in order to determine reference susceptibility values of pronounced brain iron deposits at ultra‐high field. Both patients had the atypical form of PKAN with disease onset at 12 to 14 years and duration of 7 to 20 years. Genetic testing with full sequencing of the PANK2 gene revealed homozygous and compound heterozygous mutations, respectively, in the 2 patients and heterozygous mutations in all nonaffected family members. Thirteen age‐ and sex‐matched healthy individuals (5 female and 8 male; mean age: 39.7 years; SD, 13.6) served as controls. All participants were neurologically examined with a focus on presence (patients) or absence (nonaffected mutation carriers and controls) of movement disorder abnormalities. Demographic and clinical characteristics are shown in Table 1.

Table 1.

Demographic and clinical characteristics of patients and heterozygous mutation carriers

Proband	Nationality	Age	Clinical Status (Age at Onset)	Disease Severity (BFM Score)	Genetic Status
PKAN‐1	Czech	21	Generalized dystonia, spasticity (R>L), micrographia, dysarthria, dysphagia, drooling, personality changes (14 years)	27	c.1253C>T (homozygous)
PKAN‐2	Czech	32	Generalized dystonia, Babinski sign positive (R>L), gait freezing, dysarthria, dysphagia, drooling, personality changes (12 years)	62	c.1369G>T and c.1561G>A
Non‐Manif‐1	Polish	43	Unaffected	n/a	c.573delC
Non‐Manif‐2	Polish	41	Unaffected	n/a	c.1583C>T
Non‐Manif‐3	Czech	29	Unaffected	n/a	c.1253C>T
Non‐Manif‐4	Czech	26	Unaffected	n/a	c.1253C>T
Non‐Manif‐5	Czech	56	Unaffected	n/a	c.1253C>T
Non‐Manif‐6	Czech	47	Unaffected	n/a	c.1253C>T
Non‐Manif‐7	Argentinian	47	Unaffected	n/a	c.821‐822del
Non‐Manif‐8	Czech	54	Unaffected	n/a	c.1369G>T
Non‐Manif‐9	Czech	56	Unaffected	n/a	c.1561G>A
Non‐Manif‐10	German	40	Unaffected	n/a	Deletion exon 5
Non‐Manif‐11	German	39	Unaffected	n/a	c.1663‐1G>C

BFM, Burke Fahn Marsden Dystonia Rating Scale; Non‐Manif, nonmanifesting heterozygous gene mutation carrier; n/a, not applicable.

MRI Data Acquisition

MRIs were acquired on a Magnetom 7‐T whole‐body MR scanner (Siemens Healthcare, Erlangen, Germany), using a 24‐channel receive head coil (Nova Medical, Wilmington, MA). The subject's head was carefully padded with foam cushions during the entire measurement to avoid head movements. A three‐dimensional (3D) flow‐compensated susceptibility‐weighted imaging (SWI) gradient echo technique was employed (echo time [TE] = 15 ms; repetition time [TR] = 25 ms; flip angle = 12 degrees; 72 slices; matrix = 448 × 352; spatial resolution: 0.5 × 0.5 × 1.0 mm³). In PKAN patients, turbo inversion recovery magnitude (TIRM; TE = 90 ms; TR = 10,070 ms; inversion time [TI] = 2,614 ms; flip angle = 128 degrees; 35 slices; matrix = 512 × 448; spatial resolution: 0.5 × 0.5 × 3.0 mm³) and 3D magnetization‐prepared rapid acquisition with gradient echo (MP‐RAGE; TE = 2.98 ms; TR = 2,300 ms; TI = 900 ms; flip angle = 5 degrees; matrix = 256 × 240; spatial resolution: 1.0 × 1.0 × 1.0 mm³) techniques were also employed.

Data Processing and Susceptibility Calculation

For the calculation of the susceptibility maps, the phase images derived from the high‐resolution SWI data sets were high‐pass filtered and divided into the original complex image to remove effects of large‐scale background magnetic fields (e.g., air‐tissue boundaries at the sinuses). This processing yields a homodyne high‐pass‐filtered phase image, with most low spatial frequency phases removed.18

To keep the field of view with an aspect ratio of 1:1:4, k‐space was interpolated by zero filling the phase images to a 512 × 512 × 128 three‐dimensional matrix. Susceptibility calculations using a regularized inverse filter were performed in the selected 3D regions of the Fourier transform of the high‐pass‐filtered phase image (Fig. 1).19

Figure 1.

Output of image‐processing steps. (A) Magnitude image, (B) unwrapped filtered phase image, (C) susceptibility map calculated by thresholded k‐space division with single‐orientation acquisition, and (D) susceptibility map with ROIs highlighted in color: blue, caudate nucleus; orange, putamen; dark green, globus pallidus; red, internal capsule; green, thalamus.

Image and Region of Interest Analysis

Seven regions of interest (ROIs) were placed into the deep gray matter structures bilaterally to evaluate the dependence of the calculated susceptibility values in the iron‐rich deep gray matter structures on the iron content, namely, the GP, THA, putamen (PUT), caudate nucleus (CN), substantia nigra (SN), and red nucleus (RN). An additional ROI was placed into the IC. ROIs were manually plotted on the susceptibility maps.

We calculated the difference in susceptibility values of each structure relative to the occipital white matter; values of both hemispheres were averaged. The occipital white matter was used as a reference, considering that this region exhibits the lowest intersubject variance in iron concentration and shows no iron deposits in pathological states.20

Statistical Analysis

Statistical analyses were conducted using GraphPad Prism software (GraphPad Software Inc., San Diego, CA). A normality test was performed using Kolmogorov‐Smirnov's test, and the phase and susceptibility data showed a normal distribution. Two‐tailed t tests with false discovery rate correction were performed to compare ROI susceptibility values of relatives and healthy controls. To calculate iron content, the following steps were performed: In healthy volunteers, the mean susceptibility values in every ROI were plotted against the iron content of each corresponding region, as reported by Hallgren and Sourander, in a postmortem analysis.21 Pearson's correlation coefficient and linear regression fit were calculated. Using the linear regression fit, iron content was then calculated based on measured susceptibility values for all study participants.

Results

In healthy volunteers, a strong positive linear correlation between regional bulk magnetic susceptibility and reported postmortem iron concentration21 was found (Fig. 2A). Including all ROIs, the linear regression fit yielded: χ = 0.57 ppb × [Fe] + 43.625, where [Fe] is the iron concentration in μg/g wet tissue mass and the susceptibility value χ is referenced to the mean susceptibility of occipital white matter (r² = 0.9321). The results of the regional QSM analysis to calculate iron content using the linear regression fit are summarized in Table 2.

Figure 2.

(A) Average susceptibility values in healthy controls plotted against estimated nonheme iron content based on the work of Hallgren and Sourander21 for the following brain regions (iron concentration values presented as mean ± SD μg/g): TH (47.6 ± 1); CN (92.8 ± 2); PUT (133.2 ± 3); SN (184.6 ± 7); RN (194.8 ± 7); and GP (213.0 ± 3). The calculated slope of the linear regression fit 0.57 ± 0.1 ppb per μg iron per g of wet tissue, which is in good agreement with the values of 0.56 and 0.6 previously reported by Shmueli et al.31, and Wharton and Bowtell32, respectively. (B) Differences in the magnetic susceptibility of the ROIs, relative to the occipital white matter, for all subjects in both study groups. Heterozygotes showed no significant increased magnetic susceptibility versus the healthy controls (P > 0.05) in any of the ROIs. Y‐axis scale is equal for part (A) and (B).

Table 2.

Mean bulk tissue magnetic susceptibilities (in ppb) and the corresponding iron concentrations (μg/g wet tissue) in healthy volunteers, heterozygous relatives, and patients

ROI	Controls		Relatives		Patient 1		Patient 2
	Bulk Susceptibility (ppb)	Iron Concentration (μg/g wet tissue)	Bulk Susceptibility (ppb)	Iron Concentration (μg/g wet tissue)	Bulk Susceptibility (ppb)	Iron Concentration (μg/g wet tissue)	Bulk Susceptibility (ppb)	Iron Concentration (μg/g wet tissue)
GP	173 ± 45	228	169 ± 47	220	605	986	465	740
RN	137 ± 39	164	136 ± 40	162	156	198	142	173
SN	153 ± 46	193	145 ± 45	178	628	1026	593	965
PUT	122 ± 42	138	144 ± 32	176	135	161	138	166
CN	104 ± 51	107	98 ± 55	96	110	117	106	110
THA	65 ± 27	37	75 ± 26	55	56	21	73	52
IC	4 ± 2	N.D.	4 ± 2	N.D.	19	N.D.	22	N.D.

Susceptibility values are given relative to the occipital white matter region. Iron concentration was not calculated for IC because the relationship between susceptibility and iron concentration is not linear in white matter and cannot be reliably derived from the linear regression equation calibrated on deep gray matter structures. N.D., not done.

Figure 2B shows the mean susceptibility values in each ROI for healthy controls and PANK2 mutation carriers. PANK2 mutation heterozygotes did not exhibit significantly increased concentration of iron, compared to healthy controls (P > 0.05), in any ROI nor did they exhibit any signs of movement disorders in neurological examination.

We found the typical eye of the tiger sign in patients showing iron‐deposits–related hypointensity in T2‐weighted (T2w) MRIs with a central T2w hyperintensity, presumably representing gliotic changes and edema, as indicated by the magnitude images derived from TIRM and SWI imaging (Fig. 3). In PKAN patients, considerably higher susceptibility values were found in the GP, SN, and IC (Table 2). In the GP of the 2 PKAN patients, susceptibility values were 465 to 605 parts per billion (ppb), in comparison to the mean value 173 ± (SD) 45 ppb in healthy controls. In the SN, the 2 patients had 4 times higher susceptibility values (593–628 ppb), compared to healthy controls (153 ± 46 ppb); in the IC, susceptibility values were 5 times increased in patients (19–22 ppb; healthy controls: 4 ± 2 ppb). No differences in iron concentrations were detected in other ROIs (THA, PUT, CN, and RN). When the calculated iron concentrations in GP for PANK2 mutation heterozygotes and healthy controls were plotted against age, we did not reveal differences in the slope of age‐related iron content changes between these groups (data not shown).

Figure 3.

PKAN patient: the eye of the tiger sign, that is, hypointense signal in the GP with a central region of hyperintensity clearly visible in the TIRM image (A) that is barely apparent on the SWI image (B). MP‐RAGE image shows only hypointense signal in the anterio‐medial part of the GP (C).

Discussion

We employed QSM using 7‐T MRI to characterize and quantify the iron distribution in the basal ganglia of PKAN patients and asymptomatic heterozygous PANK2 gene mutation carriers and compared these to age‐matched healthy controls. QSM was used to make sure that the susceptibility weighting would be dictated by microscopic B₀ susceptibility gradients, rather than by macroscopic B₀ field inhomogeneities. For this purpose, a threshold k‐space division with single‐orientation acquisition was employed. Our iron concentration estimates derived from susceptibility values are highly correlated with postmortem iron concentration values,21 confirming the validity of this approach.

The major aim of this study was to ascertain whether asymptomatic relatives with heterozygous gene mutations show subtle signs of iron accumulation that could be used as an endophenotypic marker, similar to mild abnormalities observed in heterozygous gene mutation carriers of other recessive movement disorders, such as Parkin‐related parkinsonism.22, 23 However, we did not detect signs of abnormal iron deposits in healthy heterozygotes with high spatial resolution 7‐T MRI. This finding argues against abnormalities in brain iron metabolism in subjects carrying a single abnormal PANK2 allele. Consequently, our results do not support the idea that subtle changes in FA and MD in heterozygous gene carriers detected by Delgado et al. are reflections of iron deposits.

The group of examined heterozygotes covers the age spectrum of the third to sixth decade, and the slope of age‐related iron accumulation in GP in this age group was not significantly different, compared to healthy controls. However, because we did not examine heterozygous subjects in the seventh decade or older, we cannot confidently exclude that abnormal iron accumulation occurs in this age group. It is also important to note that the PANK2 enzyme is not involved in iron metabolism, but in coenzyme A synthesis. Thus, iron accumulation may be only a secondary factor in the pathophysiology of PKAN and may not be best suited as an endophenotypic marker of heterozygous PANK2 mutation. Instead, heterozygotes may manifest other subtle metabolic abnormalities (e.g., in lipid or energy metabolism).24 Absence of movement disorders in examined heterozygotes suggests, however, that any abnormality, if potentially present, would not be clinically significant.

It was suggested that, besides ferritin, part of the iron deposits in PKAN patients is composed of a ferromagnetic compound with similar properties as superparamagnetic iron oxide nanoparticles.8 Our approach does not distinguish between these iron forms. Nevertheless, the QSM method is capable of detecting iron in any form with magnetic properties. Thus, even if superparamagnetic iron accumulation occurred in heterozygous gene carriers, it would have been detected by our approach.

As expected, we detected a significant increase of iron in the GP of our patients of more than 3 times, compared to controls. The iron concentration in GP measured by QSM (mean, 863 μg/g) was higher than previous results obtained by R2 relaxivity measurements (mean, 443 and 480 μg/g, respectively).8, 25 Because we scanned only 2 PKAN patients, our sample is not representative, and no solid conclusion can be drawn regarding over‐ or underestimation of the iron concentration while using the QSM or R2 relaxivity approach, respectively. Notably, no MRI method is sufficiently validated for measurement of brain iron concentration in pathological states with profound iron deposits, and calculated values should be thus treated as rough estimates.

The eye of the tiger sign was best appreciated using the TIRM sequence, whereas in the SWI sequence, the sign was not as clearly visible as a result of the strong T2w signal extinction resulting from the deposited iron (Fig. 3). The presence of the eye of the tiger sign thus strongly depends on the MRI parameters used.

In line with recent literature,3, 26, 27 iron deposition extended to adjacent areas in our patients, namely, to the SN and IC, which was reflected in clinical symptoms of mild parkinsonism and pyramidal signs in addition to generalized dystonia. Indeed, dopamine transporter single‐photon emission CT, as a marker of dopamine loss in Parkinson's disease, may be abnormal in PKAN.28 Similarly, GP and SN signal abnormalities in PKAN patients were reported for diffusion tensor imaging,3, 29 presenting with pseudoincreased FA and abnormal MD. These abnormalities are, however, likely to be consequences of iron deposits disturbing the local magnetic field.9 Transcranial ultrasound investigations30 also support the involvement of GP and SN, showing hyperechogenicity of the SN and nucleus lentiformis hypothesized to represent the correlate of iron deposits.

In conclusion, we did neither observe signs of movement disorders nor increased iron accumulation in heterozygous carriers of PANK2 gene mutation. This implies that a single functional PANK2 allele is sufficient for normal function.

Author Roles

(1) Research Project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript: A. Writing of the First Draft, B. Review and Critique.

P.D.: 1A, 1B, 1C, 3A, 3B

E.M.T.M.: 1A, 1B, 1C, 2B, 3B

V.I.M.: 1A, 1B, 1C, 3B

R.J.: 1A, 1B, 1C, 3B

J.S.: 1A, 1B, 1C, 3B

F.P.: 1A, 1B, 1C, 3B

T.N.: 1A, 1B, 1C, 3B

J.W.: 1A, 1B, 1C, 3B

S.A.S.: 1A, 1B, 1C, 3A, 3B

Disclosures

Funding Sources and Conflicts of Interest: Petr Dusek received funding from the Charles University in Prague with the grant PRVOUK‐P26/LF1/4. Vince Madai and Jan Sobesky received funding from the German Federal Ministry of Education and Research with the grant “Center for Stroke Research Berlin” (01 EO 0801; http://www.bmbf.de). Robert Jech received funding from the Czech Ministry of Health with the grant IGA MZ ČR NT12282‐5/2011 and from the Charles University in Prague with the grant PRVOUK‐P26/LF1/4. Friedemann Paul is supported by the German Research Foundation (Exc 257). Thoralf Niendorf has nothing to disclose related to the project. Susanne A. Schneider received funding from Else Kroener‐Fresenius‐Stiftung, Eva Luise und Horst Köhler‐Stiftung, and the Novartis Pharma GmbH.

Financial Disclosures for previous 12 months: The study was supported by the Else Kröner‐Fresenius‐Stiftung, the Eva Luise und Horst Köhler Stiftung, the Novartis Pharma GmbH, the German Federal Ministry of Education and Research, and

by the Czech Ministry of Education (research project PRVOUKP26/LF1/4). None of the funding bodies had any input into the study design or analysis.