Brain MRI abnormalities in the adult form of myotonic dystrophy type 1: A longitudinal case series study

Abstract

This study aimed to verify whether brain abnormalities, previously described in patients with myotonic dystrophy type 1 (DM1) by magnetic resonance imaging (MRI), progressed over time and, if so, to characterize their progression. Thirteen DM1 patients, who had at least two MRI examinations, were retrospectively evaluated and included in the study. The mean duration (± standard deviation) of follow-up was 13.4 (±3.8) years, over a range of 7–20 years. White matter lesions (WMLs) were rated by semi-quantitative method, the signal intensity of white matter poster-superior to trigones (WMPST) by reference to standard images and brain atrophy by ventricular/brain ratio (VBR). At the end of MRI follow-up, the scores relative to lobar, temporal and periventricular WMLs, to WMPST signal intensity and to VBR were significantly increased compared to baseline, and MRI changes were more evident in some families than in others. No correlation was found between the MRI changes and age, onset, disease duration, muscular involvement, CTG repetition and follow-up duration. These results demonstrated that white matter involvement and brain atrophy were progressive in DM1 and suggested that progression rate varied from patient to patient, regardless of age, disease duration and genetic defect.

Introduction

Myotonic dystrophy type 1 (DM1) is an autosomal dominant multisystem disorder, whose genetic defect consists of an abnormal expansion of cytosine-thymine-guanine (CTG) trinucleotide within the 3′ untranslated region of myotonic dystrophy protein kinase (DMPK) gene on chromosome 19q13.3.^1–4 CTG expansion ranges from 50 to 4000 repeats, and usually correlates well with the age of onset, whereas normal individuals have fewer than 50 repeats.^5,6 However, several questions including the etiopathogenesis underlying the multisystemic nature and the variable expressivity of signs and symptoms in DM1 remain unresolved.^7–9

The DM1 phenotype is extremely variable and can range from the frequently severe, often fatal, congenital form to the asymptomatic or nearly asymptomatic carrier of genetic defect.¹⁰ The adult form is the most common form of DM1 and typically presents at 15–40 years of age with distal muscle weakness, myotonia, progressive muscular atrophy, hatchet face and ptosis, cognitive impairment and neuropsychiatric symptoms.^5,6,10

Interest in the neurological aspects of DM1 has increased in the last several years,¹¹ and several brain magnetic resonance imaging (MRI) studies have reported findings ranging from no or minimal abnormalities to marked brain atrophy and severe white matter involvement.^12–38 However, there is great disagreement about the nature (developmental or degenerative) and the clinical and genetic significance of brain MRI abnormalities.³⁹ Some authors, in fact, have found a significant association between age, disease duration, muscular impairment or CTG expansion and the severity of white matter involvement^20,32–38 or brain atrophy,^19,24,32,36 while others have found no significant relationships for white matter abnormalities^25–29,35 nor for brain atrophy.^26,28–31 Ota et al.²⁸ have reported a significant negative correlation with CTG repeats only for the motor and prefrontal cortical atrophy. Romeo et al.³⁵ have found that the severity of white matter lesions (WMLs) significantly correlated with age, but not with muscular impairment or CTG repeats. Di Costanzo et al.³⁴ have reported a familial aggregation of WMLs, while others did not find a characteristic WML pattern among members of the same families.^26,27 Longitudinal MRI studies are expected to settle these issues.^11,39

This study verified if the abnormalities described in the brain MRIs of patients with DM1 progressed over time and tried to characterize such progression correlating the MRI findings with the clinical and genetic features. To the best of our knowledge, this is the first long-term longitudinal MRI study of the DM1 brain.

Materials and methods

Patients

Thirteen patients with the adult form of DM1, who had undergone over time two or more brain MRIs, were retrospectively included in this study. The patients (eight men and five women) belonged to five families (Table 1). The diagnosis was based on clinical features, family history, electromyography and genetic analysis of leukocytes. At the first observation, age ranged between 16 and 60 years and symptoms duration between 0 (patients without complaints detected in family examinations) and 26 years.

Table 1.

Clinical features of patients with myotonic dystrophy type 1.

Family.	Patient			Age at baseline	Disease duration at baseline			Duration of follow-up	MIRS
no.	no.	Sex	Relationship	(years)	(years)	Inheritance	CTG repeats	(years)	Basal	Follow-up
1	1	M	Proband	60	17	Paternal	100	7	4	5
1	2	F	Parent	44	24	Paternal	500	20	4	5
1	3	F	Sibling	22	8	Paternal	300	19	2	4
2	4	F	Sibling	34	4	Paternal	625	15	3	3
2	5	M	Sibling	27	8	Paternal	625	16	2	3
3	6	M	Proband	24	12	Paternal	467	12	2	3
3	7	F	Sibling	24	10	Paternal	467	13	3	4
3	8	M	Parent	54	0	Unknown	66	12	1	2
4	9	M	Proband	43	10	Maternal	210	12	3	4
4	10	M	Sibling	16	2	Paternal	650	15	2	3
5	11	M	Proband	39	26	Maternal	405	9	4	5
5	12	F	Sibling	29	0	Maternal	255	16	2	4
	13	M		39	10	Maternal	467	8	3	4

MIRS: Muscular impairment rating scale.⁴⁰

The severity of muscular involvement, assessed by the Muscular Impairment Rating Scale (MIRS),⁴⁰ ranged from 2 (minimal signs) to 5 (severe proximal weakness). When it was not possible to perform a clinical, genetic and instrumental evaluation, parents were considered affected on the basis of family history and photographic evidence; for one patient, it was not possible to identify the affected parent. No patient had history of cerebrovascular disease, multiple sclerosis, meningoencephalitis, hypoxic insult, head trauma with loss of consciousness, progressive dementia or other neurological diseases that could explain the MRI findings. Furthermore, no patients had hypertension, hyperglycemia or hypercholesterolemia; two patients (no. 9 and 10; family no. 4) presented a body mass index (BMI) > 30, and two (no. 8 and 10) were smokers. Six patients (no. 2, 4, 5, 8, 10, 13) had hypertriglyceridemia, and two of them belonged to the same family (no. 2).

The study was reviewed and approved by the local ethics committee and all patients provided written informed consent.

Imaging

At baseline, brain MRI was performed on a 0.5 T scanner (Vectra, General Electric, Milwaukee, WI, USA). Routine examinations included sagittal T1-weighted images (repetition time (TR) = 500 ms, echo time (TE) = 20 ms) and axial proton density- (PD-) and T2-weighted spin-echo images (TR = 2000 ms, TE₁ = 40 ms and TE₂ = 100 ms). Slice thickness was 5 mm with 2 mm gaps; field of view/matrix was 24 cm/192 × 256 for sagittal and 20 cm/160 × 192 for axial series. At follow-up, brain MRI was performed on a 1.5 T scanner (Vectra, General Electric) and the protocol included axial PD- and T2-weighted spin-echo images (TR = 48,400 ms, TE₁ = 85 ms, TE₂ = 116 ms) and fluid-attenuation inversion recovery (FLAIR) images (TR = 8890 ms, TE = 123 ms, IR = 1240 ms). Slice thickness was 5 mm with 2 mm gaps.

WMLs were identified as areas with high signal on both PD- and T2-weighted images. They were recorded as periventricular (abutting the ventricular lining), lobar (outside the periventricular and temporal regions) or temporal WMLs and scored according to a semiquantitative method.⁴¹ The total score was the sum of products between the number of lesions and the category indicating their size (1 < 0.5 cm; 2 = 0.5–1.0 cm; 3 = 1.0–1.5 cm; 4 = 1.5–2.0 cm; 5 = 2.0–2.5 cm; 6 = 2.5–3.0 cm; 7 ≥ 3.0 cm). The periventricular WMLs were scored by measuring their greatest thickness on PD-weighted images at the frontal and occipital horns (“caps”) and at the body of the lateral ventricle (“rims” or “bands”).

The signal intensity of white matter postero-superior to trigones (WMPST) was evaluated by reference to standard images (see Figure 1(a) and (c) in Di Costanzo et al.²⁴), showing different degrees of signal increase. It was scored as follows: 1 = mild = signal intensity less than standard image no. 1; 2 = moderate = signal intensity equal to or more than standard no. 1, but less than standard no. 2; 3 = severe = equal to or more than standard no. 2.⁴²

Brain atrophy was assessed by the ventricular/brain ratio (VBR), calculated in the PD-weighted image showing the pars centralis of lateral ventricles at the point of longest visible antero-posterior septal connection. The width of the ventricular bodies halfway from the anterior to posterior margin was divided by the maximum width of the brain at that level.⁴²

Two blinded experienced examiners, R.C. and A.S., independently evaluated the images and their results were compared. When they disagreed, a consensus meeting was held to reach an unequivocal agreement. Inter- and intra-rater reliabilities, measured by intraclass correlation coefficient, varied from 0.79 to 0.84 and from 0.84 to 0.87, respectively. Considering that small repositioning errors⁴³ and image quality differences are inevitable across scans, and that intra-observer agreement may vary between 79% and 92% in scoring systems,⁴⁴ only differences more than 20% from baseline were considered significant changes. Furthermore, longitudinal changes were also estimated based on qualitative radiological impression. Qualitative analysis is less sensitive than quantitative, but reduces the probability of a false-positive report.⁴⁵

Genetic analysis

Genetic analysis was performed from leukocyte DNA to assess the length of CTG repeat sequence. Standard procedure consisted of DNA separation from leukocytes, digestion with NcoI and BamHI restriction enzymes, separation of DNA fragments by gel electrophoresis, transfer onto nylon membranes by Southern blotting, hybridization with DNA probe MDY1 and additional polymerase chain reaction using described primers.¹ The CTG repeat size was estimated at the point of the most heavily staining band or at the midpoint of the smear for a very diffuse band.

Statistical analysis

Data were analyzed using the statistical package for the social sciences (SPSS)/personal computer (PC) statistical software package. Variables were examined for outliers and extreme values by box and normal quantile-quantile plots, and for distribution by the Kolmogorov-Smirnov test. When a normal distribution could not be accepted, variable transformations (square, square root, logarithmic, reciprocal of square root or reciprocal transformations) were reviewed. Since the normality could not be reached for most examined variables, a nonparametric test was used. The differences in the scores obtained at baseline and at the end of follow-up were evaluated with Wilcoxon test with a significance level of p < 0.05. The Spearman rank correlation coefficient (r_s) measured the linear association between WMLs, WMPST hyperintensity and VBR score variations, detected at the end of MRI follow-up, and other examined variables (age, age at onset, disease duration, CTG repeats, follow-up duration, MIRS, cardiovascular risk factors). The Bonferroni’s correction was used to account for multiple comparisons.

Results

Clinical and MRI features of each DM1 patient are reported in Tables 1 and 2, respectively. During the follow-up, WMLs showed significant changes in five (no. 1, 2, 3, 4, 5) out of 13 patients (38.4%). In particular, for three (no. 1, 2, 3) of five patients (no. 1, 2, 3, 7, 11) who presented with lobar WMLs at the baseline, we observed the extension and the confluence of pre-existent WMLs and the appearance of new WMLs during follow-up (Table 2; Figure 1). In two (no. 4, 5) of eight patients (no. 4, 5, 6, 8, 9, 10, 12, 13) who did not present lobar WMLs at baseline, we detected the appearance of lobar WMLs during the follow-up (Table 2; Figure 2). Among the seven patients (no. 1, 2, 3, 4, 7, 11, 12) showing the characteristic temporopolar WMLs at baseline, three (no. 2, 3, 4) presented with an enlargement of lesions (Table 2; Figure 3). In one (no. 5) of six patients (no. 5, 6, 8, 9, 10, 13) who did not show such kind of WMLs at baseline, a temporopolar WML was detectable at the end of follow-up (Table 2; Figure 3). An extension of “bands” and/or “caps” signal was revealed in three patients (no. 1, 2, 3), and in two (no. 5, 12) there was the appearance of periventricular WMLs (Table 2; Figure 4). WMPST hyperintensity (Table 2; Figure 5) increased in six patients (no. 2, 3, 4, 5, 12, 13). VBR heightened (Table 2; Figure 6) in all, but in six (no. 1, 2, 3, 5, 8, 13) out of 13 patients (46.1%) the difference from baseline was more than 20%. The scores of lobar (Z = 2.0; p = 0.043), temporal (Z = 2.1; p = 0.034) and periventricular WMLs (Z = 2.1; p = 0.038), WHPST hyperintensity (Z = 2.4; p = 0.014) and VBR (Z = 3.1; p = 0.002) were significantly increased compared to the basal ones. No relationship was found between MRI score variations and age, age at onset, inheritance, disease duration, CTG repeats, MIRS, presence of cardiovascular risk factors and follow-up duration.

Table 2.

MRI findings at baseline and at the end of follow-up with results of statistical analysis.

		Patient
	Time	1	2	3	4	5	6	7	8	9	10	11	12	13	Median	p value
WML-L	Baseline	29	14	7	0	0	0	2	0	0	0	6	0	0	0	0.043
	Follow-up	43a	30a	36a	6a	8a	0	2	0	0	0	6	0	0	2
WML-T	Baseline	12	10	5	0	0	0	4	0	0	0	14	10	0	2	0.034
	Follow-up	14	12a	10a	4a	0	0	4	0	0	0	14	10	0	4
WML-PV	Baseline	6	6	3	0	0	0	0	0	0	0	4	0	0	0	0.038
	Follow-up	8a	8a	6a	0	3a	0	0	0	0	0	4	2a	0	0
HWMPST	Baseline	3	2	2	1	1	1	1	1	1	1	2	1	1	1	0.014
	Follow-up	3	3a	3a	2a	2a	1	1	1	1	1	2	2a	2a	2
VBR	Baseline	0.30	0.24	0.19	0.12	0.14	0.19	0.19	0.18	0.20	0.16	0.20	0.17	0.20	0.19	0.002
	Follow-up	0.37a	0.29a	0.24a	0.14	0.22a	0.21	0.21	0.22a	0.23	0.17	0.22	0.18	0.24a	0.22

^aIncreases more than 20% from baseline. MRI: magnetic resonance imaging; WML: white matter lesion; L: lobar; T: temporal; PV: periventricular; HWMPST: hyperintensity of white matter postero-superior to trigones; VBR: ventricular/brain ratio.

Figure 1.

Axial T2-weighted images showing the extension and confluence of pre-existent white matter lesions (WMLs) and the appearance of new WMLs in patient 3 (Table 2) during follow-up: (a) (baseline) and (b) (19 years later).

Figure 2.

Axial T2-weighted images showing the appearance of a lobar white matter lesion (WML) (arrow) in patient 5 during follow-up: (a) (baseline) and (b) (16 years later).

Figure 3.

Axial T2-weighted images showing the enlargement of pre-existent temporopolar white matter lesions (WMLs) (a) and (b) in patient 3 (Table 2) and the development of a temporopolar WML (c) and (d) in patient 5 (Table 2) during follow-up: (a) and (c) (baseline), (b) (19 years later) and (d) (16 years later).

Figure 4.

Axial T2-weighted images showing the appearance of frontal “caps” (arrows) in patient 12 (Table 2) during follow-up: (a) (baseline) and (b) (16 years later).

Figure 5.

Axial T2-weighted images showing the increase of white matter postero-superior to trigones (WMPST) hyperintensity in patient 3 (Table 2) during follow-up: (a) (baseline) and (b) (19 years later).

Figure 6.

Axial PD-weighted images showing the ventricular enlargement in patient 1 (Table 2) during follow-up: (a) (baseline) and (b) (seven years later).

Discussion

The present study represents, to the best of our knowledge, the first long-term follow-up study of brain MRI features in patients with the adult form of DM1. The main previously unreported findings are progression over time of WML number and extent and brain atrophy, appearance of WMLs in patients previously devoid of MRI abnormalities, progressive increase of signal intensity of WMPST, significant MRI changes in some families and not in others, and absence of correlation between MRI changes and clinical or genetic features.

The progression over time of WML severity in the present study suggests that white matter involvement in the adult form of MD1 is probably more degenerative (and therefore progressive) in nature than developmental (and therefore stable over time). Previous cross-sectional studies reported conflicting results since some authors found a significant correlation between WMLs and disease duration,^{20,24,32–34,36} while others failed to find any significant correlation.^25–27,29 The significant changes in the number and extent of WMLs occurred in the members of two (no. 1 and 2) out of five families (Tables 1 and 2), while in the remaining families the white matter changes were less evident or absent. This finding settles the issue about the familial recurrence of WMLs in DM1. Some authors, in fact, reported a familial aggregation of WMLs in DM1,³⁴ while others^26,27 did not find any characteristic pattern of brain MRI abnormalities in the members of the same families.

Another issue addressed by the present study is the progression of cerebral atrophy in the adult form of DM1. Previous cross-sectional studies reported conflicting results since some authors reported a significant correlation between brain atrophy and disease duration,^24,32,33,36 while others denied such a correlation.^26,28–31 In effect, our result of progressive brain atrophy is in strict agreement with a previous longitudinal pneumoencephalographic study⁴⁶ that demonstrated a slowly progressive ventricular enlargement in muscular dystrophic (MD) patients followed for 20–97 months. Furthermore, considering that brain atrophy rate in healthy individuals is 0.2% per year after age 35 years and more than 0.5% per year over 60 years,⁴⁷ we can consider abnormal the ventricular enlargement in about half of our patients.

The absence of correlation with age, age at onset, disease duration, CTG repeats, MIRS and follow-up duration suggests that the rate and extent of MRI abnormality progression vary from patient to patient and are independent from genetic analysis on leukocytes. The variability of rate of disease progression is a known feature of DM1. Mathieu et al.⁴⁰ demonstrated a great variability of the rate of MD process comparing the disease duration with the MIRS scores in a large population of patients. The MIRS, in fact, was developed in concordance with the usual distal to proximal progression of muscular involvement in DM1.⁴⁰ Di Costanzo et al.³³ inferred the variable rate of progression of white matter involvement in DM1 in a cross-sectional study, by dividing their patient population in four groups on the basis of disease duration and severity of brain MR findings. The variable rate of progression of white matter involvement in DM1 could be explained by the huge variability of disease expression.^7,10 The lack of correlation with CTG repeats is in agreement with some authors^{21,29–31,34,35} and in disagreement with others, who have found positive correlations between brain MR abnormalities and CTG repeats.^28,32 The somatic mosaicism might explain the lack of correlation between the number of CTG repeats in blood leukocytes and brain abnormalities.

The etiopathogenesis of brain abnormalities in DM1 has been recently reviewed and a complex interaction between RNAopathy, spliceopathy and tauopathy has been implicated.⁹ In brief, mutated DMPK transcripts accumulate in RNA nuclear inclusions called foci and alter the alternative splicing of various transcripts, including those encoding for tau. This protein is expressed in neurons and glia, and is involved in microtubules assembly and stabilization, and in intracellular transport regulation.⁴⁸ In DM1, the changes in alternative splicing might result in a dysfunction of tau protein, with a variation of microtubule organization and axonal transport.⁹ The consequence might be the breakdown of cytoskeleton, the impairment of neuronal structure and function, the formation of neurofibrillary tangles (NFTs), and hence the onset and progression of neurodegeneration.^9,48 A number of neuropathological studies have shown NFT and altered level of tau isoforms in several DM1 brain regions.^49–54 RNA foci have also been reported in oligodendrocytes of subcortical white matter and the corpus callosum, even if less intense than in cortical neurons,^54,55 and pathological tau proteins have been detected in frontal white matter.⁵¹ Considering that cytoskeletal integrity is also important for the function and survival of oligodendrocytes and myelin sheaths,⁵⁶ tau mis-splicing might be a unifying pathomechanism underlying the brain atrophy and the white matter abnormalities of DM1. However, the familial aggregation in the occurrence, severity and progression of WMLs in DM1 (Di Costanzo et al.³⁴ and the present study) suggest that other genetic or environmental factors might be implicated. Neuropathological studies have reported loss of myelin and axons, dilated perivascular spaces, fibrillary gliosis, absence of cellular infiltrates or typical features of active breakdown myelin sheets, and presence of few fatty granular cells.^{14,50,54,57,58} These neuropathological findings associated with the progressive white matter involvement and the increasing ventricular enlargement, detected in the present study, suggest that a slow demyelinating process might underlie WMLs in the adult form of DM1.

An important limitation of this study is that we did not use any brain volumetric techniques, which are currently widely used for assessing the progression of several neurological diseases.⁵⁹ Unfortunately, these techniques were not yet available at the time of our first MRI observations. However, visual rating scales correlate well with quantitative volumetric measures in the assessment of WML progression.^60,61 Another limitation is the differences in image quality and spatial resolution across scans. These differences were inevitable given that the first MRI studies were performed with a 0.5 T scanner. However, visual rating scales, being rather easily applicable and requiring limited technical efforts, are particularly useful in studies where different machinery with different image quality have been used.⁶²

Conclusion

In conclusion, the results of this longitudinal case series study suggest that white matter abnormalities described in patients with adult-type DM1 might be progressive and, if absent at baseline, might appear during the course of the disease. The rate and extent of progression vary from person to person, regardless of CTG length in peripheral blood, and tend to be more pronounced in some families than others, suggesting the involvement of other genetic or environmental factors. These findings, together with previously reported neuropathological findings and pathogenetic hypotheses, suggest that white matter abnormalities in DM1 might be an expression of a slow demyelination process.

Conflict of interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Author contributions are as follows: Guarantor of integrity of the entire study (G. Tedeschi and S. Cirillo); study concepts (A. Di Costanzo, R. Conforti, G. Tedeschi and S. Cirillo); study design (A. Di Costanzo, R. Conforti, G. Tedeschi); definition of intellectual content (A. Di Costanzo, R. Conforti, G. Tedeschi and S. Cirillo); literature research (A. Cristofano, M. de Cristofaro, R. Conforti, B. Brogna and A. Di Costanzo); clinical studies (A. Di Costanzo, M. de Cristofaro and R. Conforti); data acquisition (R. Conforti, A. Sardaro and A. Di Costanzo); data analysis (A. Di Costanzo, R. Conforti and G. Tedeschi); statistical analysis (A. Di Costanzo and A. Cristofano); manuscript preparation (A. Di Costanzo, A. Cristofano, R. Conforti and B. Brogna); manuscript editing (A. Di Costanzo); and manuscript review (R. Conforti and G. Tedeschi).