Magnetic Resonance Imaging DemonstratesIncomplete Myelination in 18q- Syndrome: Evidencefor Myelin Basic Protein Haploinsufficiency

C.T. Gay,1* L.J. Hardies,2 R.A. Rauch,3 J.L. Lancaster,2 R. Plaetke,4 B.R. DuPont,4 J.D. Cody,4John E. Cornell,5 R.C. Herndon,2 P.D. Ghidoni,1 J.M. Schiff,2 C.I. Kaye,1 R.J. Leach,4 and P.T. Fox2

1Department of Pediatrics, University of Texas Health Science Center at San Antonio, San Antonio, Texas2Research Imaging Center, University of Texas Health Science Center at San Antonio, San Antonio, Texas3Department of Radiology, University of Texas Health Science Center at San Antonio, San Antonio, Texas4Department of Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, SanAntonio, Texas

5Department of Medicine, University of Texas Health Science Center at San Antonio, San Antonio, Texas

Magnetic resonance imaging (MRI) and MRIrelaxometry were used to investigate dis-turbed brain myelination in 18q- syndrome,a disorder characterized by mental retarda-tion, dysmorphic features, and growth fail-ure. T1-weighted and dual spin-echo T2-weighted MR images were obtained, and T1and T2 parametric image maps were cre-ated for 20 patients and 12 controls. MRIdemonstrated abnormal brain white matterin all patients. White matter T1 and T2 re-laxation times were significantly prolongedin patients compared to controls at all agesstudied, suggesting incomplete myelination.Chromosome analysis using fluorescence insitu hybridization techniques showed thatall patients with abnormal MRI scans andprolonged white matter T1 and T2 relax-ation times were missing one copy of themyelin basic protein (MBP) gene. The onepatient with normal-appearing white mat-ter and normal white matter T1 and T2 re-laxation times possessed two copies of theMBP gene. MRI and molecular genetic datasuggest that incomplete cerebral myelina-tion in 18q- is associated with haploinsuffi-ciency of the gene for MBP. Am. J. Med.Genet. 74:422–431, 1997. © 1997 Wiley-Liss, Inc.

KEY WORDS: 18q- syndrome; white matter;delayed myelination; myelinbasic protein

INTRODUCTION

18q- syndrome is a rare disorder characterized bymental retardation, dysmorphic features, growth fail-ure, and abnormal cerebral white matter [Werteleckiand Gerald, 1971; Miller et al., 1990; Vogel et al., 1990;Gay et al., 1994]. Table I summarizes the most com-monly reported clinical characteristics associated withthis disorder. Neurologic manifestations include sei-zures, hypotonia, nystagmus, and poor coordination[Wertelecki and Gerald, 1971; Wilson et al., 1979; Feld-ing et al., 1987]. Reported structural brain abnormali-ties include impaired migration with associated hetero-topic gray and white matter, ventricular enlargement,and cerebellar hypoplasia [Vogel et al., 1990; Felding etal., 1987]. Neuroimaging [Miller et al., 1990; Weiss etal., 1991; Rodichok and Miller, 1992; Ono et al., 1994;Lebron et al., 1994] and neuropathologic [Felding et al.,1987; Vogel et al., 1990] studies demonstrate whitematter abnormalities, typified by ‘‘delayed’’ myelina-tion, in several patients. Abnormal myelination in 18q-is most likely the result of the absence of one or moregenes that participate in myelinogenesis. Normal cere-bral myelination requires the coordinated expression ofa number of genes, including the gene for myelin basicprotein (MBP) [De Vellis, 1990]. The gene for MBP isnear the telomere of chromosome 18q, the region mostcommonly deleted in 18q- syndrome [Kamholz et al.,1987]. Most patients with 18q- are deficient for onecopy of this gene [DuPont et al., 1995]. Thus, we soughtto investigate the possibility that abnormal brain my-elination in people with 18q- syndrome correlates withMBP haploinsufficiency [Weiss et al., 1991; Miller etal., 1990; Ono et al., 1994; Wilkie, 1994].

Myelination begins in utero and continues intoyoung adulthood, but most brain myelination occursduring infancy [Yakovlev and Lecours, 1967; Brody etal., 1987; Kinney et al., 1988]. Conventional spin-echoMRI readily demonstrates this developmental process

Presented in part at the 24th national meeting of the ChildNeurology Society, Baltimore, Maryland, October 25–28, 1995.

*Correspondence to: Charles T. Gay, M.D., Department of Pe-diatrics (Neurology), UTHSCSA, 7703 Floyd Curl Drive, San An-tonio, TX 78284.

Received 3 July 1996; Revised 19 March 1997

American Journal of Medical Genetics (Neuropsychiatric Genetics) 74:422–431 (1997)

© 1997 Wiley-Liss, Inc.

[Barkovich et al., 1988; Barkovich, 1990]. As the infantbrain matures, white matter becomes progressivelybrighter on T1-weighted MRI images and darker onT2-weighted images, reflecting shortening of the T1and T2 relaxation time constants [Holland et al., 1986;Barkovich et al., 1988; Bird et al., 1989]. Myelinationcontributes to age-related T1 and T2 relaxation timeshortening in the developing brain [Barkovich et al.,1988; Holland et al., 1986; Van der Knaap and Valk,1990; Bird et al., 1989].

Post-processing of raw spin-echo MRI data yieldsquantitative information–the T1 and T2 relaxationtime constants [Bottomley et al., 1984; Holland et al.,1986; Downs et al., 1992; Koenig et al., 1990]. Calcu-lating mean T1 and T2 from selected brain regions ofinterest, a process called MRI relaxometry, permitsquantitative analysis of brain white matter develop-ment [Holland et al., 1986] and of pathological pro-cesses affecting white matter [Armspach et al., 1993].

We used MRI and MRI relaxometry to characterizecerebral white matter development in 20 children with18q- syndrome and in 12 normal children. We also de-termined the MBP genotype in the children with 18q-using fluorescence in situ hybridization (FISH) tech-niques. Correlating these results, we demonstrated arelationship between abnormal white matter develop-ment and decreased MBP gene copy number in thisdisorder.

MATERIALS AND METHODSHuman Subjects

Twenty patients (11 female and 9 male) and 12 con-trol subjects (9 female and 3 male) comprised a pilotgroup from a large multidisciplinary project investigat-ing the molecular and clinical characteristics of pa-tients with 18q- syndrome. Affected individuals werebetween 5 mo and 13.9 yr of age (mean age, 6.2 yr; SD3.3 yr). Control subjects were normal children–as de-termined by history–between 3 mo and 13.3 yr (meanage, 7.0 yr; SD 4.1 yr). They were recruited locally orwere siblings of patients and were scanned for thisstudy, not for any clinical indication.

In the patients, chromosome 18q deletions were vari-able, ranging from a proximal breakpoint at 18q21.1 toa more distal breakpoint at 18q23. Two patients, 8 and13, had cytogenetically observable interstitial dele-tions. Preliminary data regarding physical [Ghidoni etal., 1994], endocrinological [Hale et al., 1995; Ghidoniet al., 1997], neuropsychological [Thompson et al.,1995], and neuroimaging [Gay et al., 1994] character-istics of this group have been reported previously.

This study was approved by the Institutional ReviewBoard of the University of Texas Health Science Centerat San Antonio.

Neuroimaging Techniques

MRI acquisition. All MR examinations of thebrain were performed on a 1.9-T magnet (Elscint,Haifa, Israel) using conventional spin-echo techniques.Axial T1-weighted (500/20: TR/TE) and double-echoaxial T2-weighted (3,400/20–80) images with identicalslice alignment were obtained using 5-mm thick sliceswith a 1-mm gap and a 256 X 256 matrix. The T2-weighted studies used a second-order motion compen-sation pulse sequence during the second echo to com-pensate for pulsatile flow-induced artifacts. Patientswho were unable to lie still were sedated with chloralhydrate, 50-100 mg/kg of body weight (maximum dose,1,000 mg), and physiologic indices were monitored by apulse oximeter.

Qualitative MRI analysis. Two investigators in-dependently evaluated the MRI scans for degrees ofcerebral myelination. The investigators were blinded tothe patients’ MBP genotype. The images were assigneda numerical score based on the general appearance ofcerebral white matter, with higher numbers indicatingmore immature white matter. Patients’ T2-weightedMR images were compared to age-matched normal im-ages after scoring was completed. The following grad-ing scale was used: grade 1, normal-appearing whitematter (adult pattern); grade 2, white matter signalintensity higher than normal, but lower signal inten-sity than adjacent gray matter; grade 3, white matterisointense compared with gray matter; grade 4, whitematter hyperintense compared with gray matter.

Quantitative MRI analysis: MRI relaxometry.Digitized raw image data were converted to quantita-tive T1 and T2 parametric image maps on a Sun Sparc-station (Sun Microsystems, Mountain View, CA) usingMagnetic Resonance Parametric Analyzer (MRPA)software (Research Imaging Center, University ofTexas Health Science Center at San Antonio) [Downset al., 1992; Herndon, 1995]. This method for esti-mating T1 and T2 maps uses T1-weighted and dual-echo T2-weighted conventional spin-echo sequencescommonly used clinically for MR imaging of thebrain. MRPA calculates the parametric image mapsusing a look-up table. The look-up table contains pre-solved solutions to Bloch’s equations for ratios of MRsignals (images) across predefined ranges of T1 and T2in milliseconds. The outputs of these calculations areestimated T1 and T2 parametric image maps with pixel

TABLE I. Commonly Reported Features of 18q- Syndrome(From The 18q-Registry and Research Society, San

Antonio, Texas)

% %

Mental retardation 96 Congenital heart disease 33Short stature 80 Abnormal female genitalia 32Midface hypoplasia 70 Palate abnormality 30Hypotonia 63 Cleft 9Prominent antihelix 63 High 17Microcephaly 61 Bifid uvula 4Carp mouth 57 Pale optic discs 28Abnormal male genitalia 55 Skin dimples 28Foot deformity 48 Broad nasal bridge 26Increased whorl patterns 44 Abnormally placed 2nd toe 26Stenotic ear canals 41 Hearing impairment 26Long tapering digits 39 Epicanthal folds 26Proximally set thumbs 35 IgA deficiency 26Nystagmus 35

MRI and Myelin Basic Protein in 18q- 423

values in milliseconds (Fig. 1). Relaxometry resultswere calibrated by measuring T1 and T2 of standardcopper sulfate solutions using MRPA and comparingT1 and T2 on the same copper sulfate concentrationsdetermined with inversion recovery sequences.

T1 and T2 parametric image maps were transferredto a Macintosh (Apple Computer, Cupertino, CA) work-station for viewing and for region-of-interest (ROI)analysis using Digital Image Processing Station (DIPS)(HIPG, Boulder, CO). We performed analyses on sev-eral widely separated regions of the brain because thetiming and rate of cerebral myelination vary regionally[Barkovich et al., 1988; Yakovlev and Lecours, 1967;Brody et al., 1987; Kinney et al., 1988]. ROIs weredrawn in four white matter regions: minor forceps ofthe corona radiata (frontal white matter), major for-ceps of the corona radiata (occipital white matter),genu of the corpus callosum, and middle cerebellar pe-duncle. Homologous regions in the left and right hemi-spheres were combined to increase the size of each ROI,thereby reducing the impact of noise and of randomvariation. Combined homologous ROIs contained aminimum of 30 pixels. Mean T1 and T2 relaxationtimes are reported for four white matter regions.

Molecular Genetic Analysis

Fluorescence in situ hybridization (FISH).Metaphase chromosome preparations were obtainedfrom each patient either from transformed lympho-cytes or by primary blood harvest using the standardmethod of Moorehead et al. [1960] with ethidium bro-mide treatment to produce prometaphase spreads[Ikeuchi, 1984]. The MBP P1 probe was isolated byscreening a human genomic P1 library [Shepherd etal., 1994] using a set of MBP-specific PCR primers,4018 and 4017 (Primer sequence obtained from C.Chinault, Genome Database, Johns Hopkins Univer-sity, 1992). The purified P1 DNA was labeled with bio-tin by nick translation using biotin-14-dATP (BRL/Gibco, Bethesda, MD). FISH was carried out as previ-ously described [Pinkel et al., 1988] using 40 ng of P1probe. The fluorescent probes were visualized using aZeiss Axioplan Fluorescent microscope equipped withFITC, DAPI, and triple band pass filter sets. Imageswere digitized using Applied Imaging Probevision(Pittsburgh, PA), and photographs were printed on aKodak XL 7700 color image printer.

Determination of deletion extent. Molecularanalysis using genomic DNA was performed to deter-

Fig. 1. Diagrammatic representation of a rapid method for calculating T1 parametric image maps with MRPA. PD represents the image described byBloch’s equation for signal intensity of a proton density MR image. T1W represents the image described by Bloch’s equation for signal intensity of aT1-weighted MR image. X represents the value of the ratio obtained by dividing each pixel (signal intensity) in PD by each corresponding pixel (at thesame 3-dimensional coordinate) in T1W. The look-up table contains the T1 value corresponding to the ratio ‘‘X’’ for each pixel.

424 Gay et al.

mine the extent of deleted chromosome 18 material inall of the patients. Hybrid cell lines were constructedon key patients defining the critical region using pre-viously described methods [Cody et al., 1997]. The DNAwas then analyzed using polymerase chain reaction(PCR)-based markers from Genethon [Gyapay et al.,1994].

PCR was performed in a total reaction volume of 10ml using 50 ng of genomic DNA, 50 ng of each primer,200 mM dNTP’s, 0.5 U Taq polymerase (Perkin Elmer-Cetus, Norwalk, CT), and 1.5 mM MgCl2. One primerof the pair was end-labeled at the 5’ end with g32P-dATP. PCR amplification consisted of 30 cycles of 1 minat 95°C, followed by 1 min at an annealing temperatureof 60°C, and 1 min elongation at 72°C. PCR productswere separated on a 7% polyacrylamide gel run at 65 Wfor 5 hr and visualized using Kodak XAR-5 film andintensifying screens.

Statistical Analysis

Statistical analyses were performed on MRI relax-ometry data with SPSS [1993]. We analyzed the effectof age on white matter development using linear re-gression analyses of log-transformed data (all log-transformed data were normally distributed, Kol-mogorov-Smirnov test, 2p $0.058). Log-transformationfor all dependent and independent variables providedthe best fit. Intercepts of the regression lines for pa-tients and controls were compared with the t-test.Slopes of the regression lines (regression coefficients)were tested with analysis of variance (ANOVA). Whenthere are no significant differences between slopes (i.e.,when slopes are parallel), differences between the in-tercepts of the regression lines of patients and controlsmeasure the differences in T1 and T2 relaxation timesbetween groups. Slopes of the regression lines assessage-related effects on T1 and T2.

The Mann-Whitney U test was used to compare over-all differences in age between patients and controls

Combining homologous regions of interest. Wesought to increase ROI sample size–and thus increasethe accuracy of the estimate of the mean pixel value–bycombining samples from homologous left and rightcerebral hemisphere regions. Using the paired-sam-ple t-test, there were no differences in T1 (or T2) be-tween homologous left and right hemisphere regions(P > 0.05). Thus, we report mean T1 (and mean T2) assingle values for each of these paired regions.

Interrater reliability for region of interestanalyses. Interclass correlation coefficients wereused to evaluate interrater reliability for the region ofinterest measurements [Fleiss, 1986]. Three investiga-tors independently determined T1 and T2 in the frontaland occipital white matter of five patients. Variancecomponents for investigators, patients, and error werecomputed from an analysis of variance table. No sig-nificant mean differences were observed among raters(P > .05). Interclass correlations ranged from r 4 .898for T2 measurements in the occipital white matter tor 4 .996 for associated T1 measurements. Three of thefour interclass coefficients were r 4 .99 or higher. Allfour coefficients suggest excellent interrater reliabilityfor these region of interest analyses.

Interrater reliability for qualitative MRIanalyses. The simple kappa statistic was calculatedto determine interrater reliability for qualitativeanalysis of patients’ MRI scans [Fleiss, 1981]. A simplekappa score of 0.7 indicated that there was good agree-ment between the two investigators using the qualita-tive scale described above.

RESULTS

Age did not differ significantly between the patientgroup and the control group (Mann-Whitney U test, P4 0.55). MRI data from patient 8 was treated sepa-rately because this patient’s genotype differed from thegenotypes of all other patients regarding MBP genecopy number. The MBP gene copy number for all 20patients was determined using FISH. Nineteen pa-tients had only one copy of the gene for MBP (Fig. 2a).Patient 8 had two copies of the MBP gene (Figure 2b).Figure 3 demonstrates the deletion extent in three pa-tients. Patients 8 and 13 had interstitial deletions of18q. All other patients had terminal deletions, and pa-tient 33 had the smallest deletion. In patient 8, aunique region of approximately 2 megabases was con-served that contained the gene for MBP. This regionwas missing from one copy of 18q in all other patients.

Qualitative MRI assessment was made by two inde-pendent raters. Investigators concurred on the degreeof myelination for 16 patient scans; for each of the otherfour scans, all rated as abnormal by both investigators,an average grade was calculated. Nineteen of 20 pa-tients’ scans exhibited abnormal white matter com-pared to age-matched controls. The principle abnor-mality was a reduced gray-white distinction due to in-creased white matter signal intensity in patients. Thewhite matter in the 5-month-old patient and in the3-month-old control had similar signal intensities, butthis patient’s total cerebral white matter volume andcorpus callosum size were reduced and posterior hornventricular volume was increased (colpocephaly, Fig.4a). For the 18q- group, gray-white contrast increased–such that the scans appeared more mature–with in-creasing patient age (Fig. 4). However, gray-white con-trast was poor and white matter T2-weighted signalintensity was abnormally high even in the oldest pa-tient studied. The genu of the corpus callosum ap-peared normal in all but three patients. These findingscontrasted with quantitative MRI results (see MRI re-laxometry, below). Patient 8–the only patient with twocopies of the gene for MBP–had a normal MRI scan(Fig. 2d). Brain MRI was normal in all 12 control sub-jects.

Quantitative evaluation of the MRI data using relax-ometry revealed that white matter T1 and T2 relax-ation times were significantly longer in patients thanin controls. Age-related changes occurred in all regions(Fig. 5). Comparison of the intercepts for the log-transformed data demonstrated that T1 and T2 relax-ation times were significantly longer in patients thanin controls (t-test, P # 0.001), while slopes of the re-gression lines for patients and controls were equal(ANOVA, p $ 0.08). Thus, T1 and T2 remained pro-longed even in older children with 18q-. Although the

MRI and Myelin Basic Protein in 18q- 425

genu of the corpus callosum appeared normal on stan-dard MR images in all but three patients, regressionanalysis demonstrated significant differences in corpuscallosum T1 and T2 values between patients and con-trols (t-test, P < 0.001).

T1 and T2 relaxation times decreased with increas-ing age in both patients and controls. For both groups,scattergrams demonstrated nonlinear relationships be-

tween T1 (or T2) and age that were best described byexponential decay functions, such that 1) age effectson T1 and T2 were more pronounced at younger ages,and 2) T1 and T2 reached plateau values in early child-hood (Fig. 5). Linear regression analysis of the log-transformed data revealed that the effect of age on T1and T2 measurements was significant in all four re-gions studied in patients and in controls.

Fig. 2. FISH and MRI results for two 8-year-olds with 18q-. Both children had many characteristics of 18q- syndrome, including dysmorphic featuresand mental retardation. However, patient 18 had only one copy of the gene for MBP (a), while patient 8 had two copies of this gene (b). T2-weighted MRIimages demonstrate abnormally increased signal intensity of the deep cerebral white matter in patient 18 (c), while the MRI of patient 8 is normal (d).Arrows in (a) and (b) point to FISH probe hybridizing to the MBP gene.

426 Gay et al.

Patient 8 was the only 18q- patient with two copies ofthe gene for MBP. T1 and T2 relaxation times were inthe normal range in all four white matter ROIs in thisindividual. Plots of T1 and T2 vs. age demonstratedthat mean T1 and T2 values in all regions were age-appropriate (Fig. 5).

DISCUSSION

Conventional spin-echo MRI scans and quantitativeMR relaxometry techniques demonstrated abnormalcerebral and cerebellar white matter–best described asincomplete myelination–in all of our patients with onecopy of the MBP gene. MBP is crucial for normal my-elinogenesis, a major developmental process occurringprimarily in the first 18 months of life [De Vellis, 1990;Wrabetz et al., 1990; Kamholz et al., 1988]. Severalcase reports suggested an association between abnor-mal myelination and MBP hemizygosity in patientswith 18q- [Miller et al., 1990; Weiss et al., 1991; Vogelet al., 1990; Ono et al., 1994]. Miller et al. [1990] pos-tulated that, in patients with 18q- syndrome, ‘‘im-paired myelination of the central white matter tract-s...(is) most likely due to failure of expression of theMBP gene.’’ Our findings agree with this assertion.Furthermore, our only patient who had normal-

appearing white matter and normal white matter T1and T2 indices also had two copies of the MBP gene. Incontrast to Miller et al. [1990], we found evidence forage-related maturation in white matter, as well as evi-dence for incomplete myelination in the corpus callo-sum. This finding is most likely because our analysisincluded younger patients than did Miller’s group,since we demonstrated the most pronounced age-related changes between five months and four years ofage.

Reduced MBP gene dosage separated the MRI scansof 18q- patients from the scan of patient 8 and from thescans of normal controls. All patients with only onecopy of the MBP gene had abnormal white matter in-dices. Furthermore, T1 and T2 relaxation curves pla-teaued at significantly higher levels in these patientsthan in controls, suggesting that white matter T1 andT2 indices in MBP-deficient patients do not normalize.Many authors refer to this type of white matter abnor-mality as ‘‘delayed myelination’’ [Ono et al., 1994;Miller et al., 1990; Barkovich, 1990; Bird et al., 1989].Delayed myelination implies that myelination beginslater in affected patients but eventually becomes nor-mal or complete. Our MRI relaxometry results demon-strate that the phrase ‘‘incomplete myelination’’ is amore accurate descriptor for the white matter abnor-mality in this disorder. That is, T1 and T2 did not ap-proach normal values, suggesting that the total myelincontent in brain white matter was reduced at all agesin 18q- patients, a finding in agreement with previousneuropathological studies [Felding et al., 1987; Vogel etal., 1990].

Although white matter signal intensity was higherthan normal at all ages, T2-weighted MRI scans dem-onstrated age-related white matter maturationalchanges in 18q- patients. Similarly, white matter T1and T2 values decreased steadily until about four yearsof age, but these values did not approach control T1and T2 values. The trend for decreasing T1 and T2 withincreasing age in 18q- patients suggests that someevents contributing to myelinogenesis are occurring.These events probably include expression of myelincomponents regulated by genes at loci not disrupted by18q deletions [De Vellis, 1990].

Haploinsufficiency, resulting from reduced gene dos-age, expression, or protein activity, is a commonmechanism of disease in chromosomal deletions[Wilkie, 1994]. When a copy of the MBP gene is lost, adominant phenotype occurs, consistent with a loss-of-function mutation. Therefore, the MBP gene appears toprovide another example of a haploinsufficient locus.Reduced copy number of the MBP gene may lead todecreased production of MBP, a protein produced inlarge quantities by oligodendrocytes during infancyand early childhood. Since MBP is an essential compo-nent of normal myelin, decreased MBP availabilitymight cause either reduced quantities of myelin, func-tionally abnormal myelin, or both.

A specific deleted critical region–not deletion size–leads to abnormal myelination in this disorder. Patient8 had a large interstitial deletion sparing the gene forMBP. Despite the size of the deletion, this patient’swhite matter appearance and T1 and T2 indices were

Fig. 3. Comparison of chromosome 18 deletion size in three patients.Patient 33, a 6-year-old, had the smallest terminal deletion; patient 13, a13-year-old, had a large interstitial deletion. Both deletions included thecritical region encoding the gene for MBP. MRI scans and T1 and T2 in-dices demonstrated abnormal white matter in both patients. Patient 8 alsohad a large interstitial deletion, but this deletion spared the critical regionencoding the MBP gene. Open circles represent markers absent. Closedcircles represent markers present. The MBP gene is tightly linked tomarker D18S554.

MRI and Myelin Basic Protein in 18q- 427

Fig. 4. Age-matched T2-weighted MRI scans, patients with the 18q- syndrome (a,c,e,g) and controls (b,d,f,h). A 5 month old patient (a) is comparedto a 3 month old control (b). Scans c and d are from an 18 month old patient and control; e and f, 5 year olds; g and h, 10 year olds. Scans b,d,f and h depictthe normal changes in gray-white contrast as myelination proceeds, with white matter achieving an essentially adult appearance by about 18 months ofage. In patients with 18q- having only one copy of the gene for MBP, white matter signal intensity decreases with increasing age, but never normalizes,suggesting incomplete myelination. (Qualitative MRI scores for patients are: a 4 4; c 4 4; e 4 3; g 4 2. Scores for controls are: b 4 4 [normal infantpattern]; d, f, and h 4 1.)

428 Gay et al.

Fig. 4. (Continued).

MRI and Myelin Basic Protein in 18q- 429

normal even though the deletion included chromosom-al material also deleted in other 18q- patients. Molecu-lar analysis revealed that the deleted chromosome 18qin patient 8 conserved a unique region of approxi-mately 2 Mb not found in any of our other patients–thecritical region for normal myelination in 18q-.

Kline et al. [1993] suggested that other genes on 18qnonadjacent to the gene for MBP may influence thedegree of abnormal myelination. They reported normalMRI scans in two of five patients with distal 18q dele-tions that appeared to include the region encoding theMBP gene. However, they did not use quantitative MRItechniques, nor did they perform a specific assay forthe MBP gene. Using visual MRI inspection alone, theycould have missed more subtle white matter abnor-malities in these two patients.

In addition to better characterizing incomplete my-elination, MRI relaxometry detects white matter ab-normalities that might not be detected visually usingconventional MRI. Several authors reported that thecorpus callosum appeared normal in 18q- patients evenif adjacent central white matter exhibited abnormallyincreased signal intensity using T2-weighted images[Miller et al., 1990; Kline et al., 1993; Ono et al., 1994].We demonstrated that T1 and T2 relaxation times inthe genu of the corpus callosum were significantlylonger in our patients than in age-matched controlseven when this structure appeared normal.

Thus, MRI and MRI relaxometry data support thehypothesis that loss of a copy of the MBP gene isstrongly associated with incomplete myelination in18q- syndrome. Ongoing MRI studies include analyz-ing a group of very young patients and controls to char-acterize the nature of white matter maturation in 18q-in the first 12 months of life; performing serial follow-

up MRI on 18q- patients to analyze longitudinally theeffects of age on myelination patterns; and furthercharacterizing white matter abnormalities in 18q- us-ing magnetization transfer [Dousset et al., 1992;Hiehle et al., 1994] and diffusion [Nomura et al., 1994;LeBihan et al., 1991] techniques. Continuing molecularcharacterization of the terminal region of 18q will helpdetermine if other tightly linked deleted genes mightcontribute to incomplete myelination in this disorder.

ACKNOWLEDGMENTS

We thank Dr. Charles Dreyer for assistance in vali-dating our data analysis methods and Dr. Daniel Halefor critically reviewing the manuscript. We also thankIvelisse Galarza-Torres and Dr. Jinqui Li for assis-tance in acquiring the MRI images, and Cono Fariasand Shawn Mikiten for assistance with producing thefigures.This work was supported by funding from theMorrison Trust and the MacDonald family with match-ing funds from Microsoft, Inc.

REFERENCES

Armspach JP, Gounot D, Namer IJ, Ohlenbusch HH, Rumbach L, Cham-bron J (1993): Quantitative cerebral magnetic resonance imaging dur-ing ACTH treatment of multiple sclerosis. Magn Reson Imaging 11:1147–1153.

Barkovich AJ (1990): Normal development of the neonatal and infantbrain. In Barkovich AJ (ed): ‘‘Pediatric Neuroimaging.’’ New York:Raven Press, pp. 5–34.

Barkovich AJ, Kjos BO, Jackson DE, Norman D (1988): Normal matura-tion of the neonatal and infant brain: MR imaging at 1.5 T. Neurora-diology 166:173–180.

Bird RC, Hedberg M, Drayer BP, Keller PJ, Flom RA, Hodak JA (1989):MR assessment of myelination in infants and children: Usefulness ofmarker sites. AJNR 10:731–740.

Fig. 5. Curve-fitting for frontal white matter T1 (a) and T2 (b) vs. age. Scattergrams were best described by exponential decay functions for all whitematter ROIs [ti 4 (a) z (age−b), where ti represents T1 or T2, ‘‘a’’ estimates the intercept, and ‘‘b’’ defines the rate of decay of ti]. For log-transformed data,regression coefficients (−0.25 # b # −0.055, P # 0.004) and Pearson’s correlation coefficients (−0.95 # r # −0.76, P # 0.004) were significant in all fourregions studied. Except for patient 8, values for affected individuals fit on the same curves even though deletion sizes varied. (Squares represent patientvalues, circles represent control values, and large triangle represents patient 8 values.)

430 Gay et al.

Bottomley PA, Foster TH, Argersinger RE, Pfeifer LM (1984): A review ofnormal tissue hydrogen NMR relaxation times and relaxation mecha-nisms from 1–100 MHz: Dependence on tissue type, NMR frequency,temperature, species, excision, and age. Med Phys 11:425–448.

Brody BA, Kinney HC, Kloman AS, Gilles FH (1987): Sequence of centralnervous system myelination in human infancy. I. An autopsy study ofmyelination. J Neuropath Exp Neurol 46:283–301.

Cody JD, Brkanac Z, Pierce JF, Plaetke R, Ghidoni PD, Kaye CI, Leach RJ(1997): Preferential loss of the paternal alleles in the 18q- syndrome.Am J Med Genet (in press).

De Vellis J (1990): Developmental and hormonal regulation of gene expres-sion in oligodendrocytes. Ann NY Acad Sci 605:81–89.

Dousset V, Grossman RI, Ramer KN, Schnall MD, Young LH, Gonzalez-Scarano F, Lavi E, Cohen JA (1992): Experimental allergic encephalo-myelitis and multiple sclerosis: Lesion characterization with magneti-zation transfer imaging. Radiology 182:483–491.

Downs JH III, Alyassin AM, Herndon RC, Lancaster JL (1992): High speedcalculation of T1, T2, and proton density parametric images of the headfrom clinical MRI data. Med Phys 19:CC114–120.

DuPont BR, Gay CT, Cody JD, Plaetke R, Hardies LJ, Rauch RA, GhidoniPD, Fox PT, Kaye CI, Leach RJ (1995): Genotype at the myelin basicprotein locus with correlation to MRI in the 18q- syndrome. Am J HumGenet 57:A238.

Felding I, Kristoffersson U, Sjostrom H, Noren O (1987): Contribution tothe 18q- syndrome. A patient with del (18) (q22.3qter). Clin Genet31:206–210.

Fleiss JL (1981): ‘‘Statistical Methods for Rates and Proportions.’’ 2nd ed.New York: Wiley & Sons, p 218.

Fleiss JL (1986): ‘‘The Design and Analysis of Clinical Experiments.’’ NewYork: Wiley & Sons.

Gay CT, Lancaster JL, Floyd JL, Marcotte H, Kaye CI, Fox PT (1994):Quantitative magnetic resonance imaging of dysmyelination in the18q- syndrome. Ann Neurol 36:547–548.

Ghidoni PD, Cody J, Danney M, Rauch R, Gay C, Leach RJ, Kaye CI(1994): Growth hormone deficiency in 18q deletion syndrome. Am JHum Genet 56:A105.

Ghidoni PD, Hale DE, Cody JD, Thompson NM, McClure EB, Gay CT,Danney M, Leach RJ, Kaye CI (1997): Growth hormone deficiency as-sociated in the 18q deletion syndrome. Am J Med Genet 69:7–12.

Gyapay G, Morissette J, Vignal A, Dib C, Fizames C, Millasseau P, Marc S,Bernardi G, Lathrop M, Weissenbach J (1994): The 1993–94 Genethonhuman genetic linkage map. Nature Genet 7:246–249.

Hale DE, Ghidoni PD, Kaye CI, Cody JD, Galbreath DS, Leach RJ, Gay CT,Danney MM (1995): Growth failure and growth hormone deficiency in18q- syndrome. Pediatr Res 37:90A.

Herndon RC (1995): ‘‘Quantification of White Matter and Gray MatterFrom Magnetic Resonance Images Using Fuzzy Classifiers.’’ Disserta-tion. University of Texas Health Science Center at San Antonio.

Hiehle JF, Lenkinski RE, Grossman RI, Dousset V, Ramer KN, SchnallMD, Cohen JA, Gonzalez-Scarano F (1994): Correlation of spectroscopyand magnetization transfer imaging in the evaluation of demyelinatinglesions and normal appearing white matter in multiple sclerosis. MagnReson Med 32:285–293.

Holland BA, Haas DK, Norman D, Brant-Zawadzki M, Newton TH (1986):MRI of normal brain maturation. AJNR 7:201–208.

Ikeuchi TP (1984): Inhibiting effect of ethidium bromide on mitotic chro-mosome condensation and its application to high resolution chromo-some banding. Cytogenet Cell Genet 38:56–61.

Kamholz J, Spielman R, Gogolin K, Modi W, O’Brien S, Lazzarini R (1987):The human myelin-basic-protein gene: Chromosomal localization andRFLP analysis. Am J Hum Genet 40:365–373.

Kamholz J, Toffenetti J, Lazzarini RA (1988): Organization and expressionof the human myelin basic protein gene. J Neurosci Res 21:62–70.

Kinney HC, Brody BA, Kloman AS, Gilles FH (1988): Sequence of central

nervous system myelination in human infancy. II. Patterns of myelina-tion in autopsied infants. J Neuropath Exp Neurol 47:217–234.

Kline AD, White ME, Wapner R, Rojas K, Biesecker LG, Kamholz J, ZackaiEH, Muenke M, Scott CI, Overhauser J (1993): Molecular analysis ofthe 18q- syndrome–and correlation with phenotype. Am J Hum Genet52:895–906.

Koenig SH, Brown RD, Spiller M, Lundbom N (1990): Relaxometry ofbrain: Why white matter appears bright in MRI. Magn Reson Med14:482–495.

Le Bihan D, Turner R, Moonen CTW, Pekar J (1991): Imaging of diffusionand microcirculation with gradient sensitization: Design, strategy, andsignificance. J Magn Reson Imaging 1:7–28.

Lebron D, Perle MA, Ostrer H, Reich E, Hilz M, Au L, Kolodny EH (1994):Cerebral myelin deficiency in teenage girls with the 18q- syndrome.Pediatr Neurol 11:145.

Miller G, Mowrey PN, Hopper KD, Frankel CA, Ladda RL (1990): Neuro-logic manifestations in 18q- syndrome. Am J Med Genet 37:128–132.

Moorehead PS, Nowell PC, Mellman WJ, Battips DM, Hungerford DA(1960): Chromosome preparations of leukocytes cultured from humanperipheral blood. Exp Cell Res 20:613–616.

Nomura Y, Sakuma H, Takeda K, Tagami T, Okuda Y, Nakagawa T (1994):Diffusional anisotropy of the human brain assessed with diffusion-weighted MR: Relation with normal brain development and aging.AJNR 15:231–238.

Ono J, Harada K, Yamamoto T, Onoe S, Okada S (1994): Delayed myelina-tion in a patient with 18q- syndrome. Pediatr Neurol 11:64–67.

Pinkel D, Landegent J, Collins C, Fuscoe J, Segraves R, Lucas J, Gray J(1988): Fluorescence in situ hybridization with human chromosome-specific libraries: Detection of trisomy 21 and translocations of chro-mosome 4. Proc Natl Acad Sci USA 85:9138–9142.

Rodichok L, Miller G (1992): A study of evoked potentials in the 18q-syndrome which includes the absence of the gene locus for myelin basicprotein. Neuropediatrics 23:218–220.

Shepherd NS, Pfrogner BD, Coulby JN, Ackerman SL, Vaidyanathan G,Sauer R, Balkenhol TC, Sternberg N (1994): Preparation and screeningof an arrayed human genomic library generated with the P1 cloningsystem. Proc Natl Acad Sci USA 91:2629–2633.

SPSS (1993): ‘‘SPSS.’’ Chicago: SPSS.

Thompson NM, McClure EB, Ghidoni PD, Galbreath DS, Cody JD, Gay CT,Sloan TB, Leach RJ, Kaye CI (1995): Cognitive/behavioral phenotype in18q- syndrome. Pediatr Res 37:153A.

Van der Knaap MS, Valk J (1990): MR imaging of the various stages ofnormal myelination during the first year of life. Neuroradiology 31:459–470.

Vogel H, Urich H, Horoupian DS, Wertelecki W (1990): The brain in the18q- syndrome. Dev Med Child Neurol 32:732–737.

Weiss BJ, Kamholz J, Ritter A, Zackai EH, McDonald-McGinn DM,Emanuel B, Fishbeck KH (1991): Segmental spinal muscular atrophyand dermatological findings in a patient with chromosome 18q dele-tion. Ann Neurol 30:419–423.

Wertelecki W, Gerald PS (1971): Clinical and chromosomal studies of the18q- syndrome. J Pediatr 78:44–52.

Wilkie AOM (1994): The molecular basis of genetic dominance. J MedGenet 31:89–98.

Wilson MG, Towner JW, Forsman I, Siris E (1979): Syndromes associatedwith deletion of the long arm of chromosome 18[del(18q)]. Am J MedGenet 3:155–174.

Wrabetz LG, Kia-Noury D, Shumas S, Grinspan J, Pleasure D, Kamholz J(1990): Transcriptional regulation of the human myelin basic proteingene. Ann NY Acad Sci 605:354–357.

Yakovlev PI, Lecours AR (1967): The myelogenetic cycles of regional matu-ration in the brain. In Minkowski A (ed): ‘‘Regional Development of theBrain in Early Life.’’ Oxford: Blackwell, pp 3–69.

MRI and Myelin Basic Protein in 18q- 431