Human brain pathology in myotonic dystrophy type 1: A systematic review

Abstract

Brain involvement in myotonic dystrophy type 1 (DM1) is characterized by heterogeneous cognitive, behavioral, and affective symptoms and imaging alterations indicative of widespread grey and white matter involvement. The aim of the present study was to systematically review the literature on brain pathology in DM1. We conducted a structured search in EMBASE (index period 1974–2017) and MEDLINE (index period 1887–2017) on December 11, 2017, using free text and index search terms related to myotonic dystrophy type 1 and brain structures or regions. Eligible studies were full‐text studies reporting on microscopic brain pathology of DM1 patients without potentially interfering comorbidity. We discussed the findings based on the anatomical region and the nature of the anomaly. Neuropathological findings in DM1 can be classified as follows: (1) protein and nucleotide deposits; (2) changes in neurons and glial cells; and (3) white matter alterations. Most findings are unspecific to DM1 and may occur with physiological aging, albeit to a lesser degree. There are similarities and contrasts with Alzheimer's disease; both show the appearance of neurofibrillary tangles in the limbic system without plaque occurrence. Likewise, there is myelin loss and gliosis, and there are dilated perivascular spaces in the white matter resemblant of cerebral small vessel disease. However, we did not find evidence of lacunar infarction or microbleeding. The various neuropathological findings in DM1 are reflective of the heterogeneous clinical and neuroimaging features of the disease. The strength of conclusions from this study's findings is bounded by limited numbers of participants in studies, methodological constraints, and lack of assessed associations between histopathology and clinical or neuroimaging findings.

INTRODUCTION

Myotonic dystrophy type 1 (DM1) is a complex multisystem disease, caused by a cytosine–thymine–guanine (CTG) microsatellite repeat expansion in the 3′‐untranslated region of the dystrophia myotonica protein kinase (DMPK) gene, that has autosomal dominant inheritance. ¹ DM1 is a chronic progressive disease characterized by reduced survival, high burden of disease, and loss of quality of life. ² , ³ Brain involvement in DM1 is phenotypically heterogeneous. ⁴ Features include variable combination of developmental disorders, cognitive deficits in various domains, personality and behavioral disturbances, affective symptoms, and sleep disorders. ⁵ , ⁶ , ⁷ These clinical features of brain involvement in DM1 are accompanied by alterations in structural and functional brain imaging, as summarized in several recent reviews. ⁸ , ⁹ , ¹⁰

Despite progress in the understanding of brain correlates of these non‐motor manifestations of DM1, partly due to increasingly sophisticated imaging techniques, less is known about pathological alterations that underly these now well‐recognized clinical and imaging phenotypes in DM1. Various histopathological features have been observed in postmortem brain samples of patients with DM1. ⁴ In the current study, we systematically reviewed the published literature on microscopic brain alterations in the human DM1 brain. A thorough and comprehensive literature review was conducted to identify gaps in the present knowledge and to inform future studies evaluating microscopic brain alterations in DM1. These goals comport with suggestions and advice recently formulated by the scientific community in DM1. ¹¹ , ¹² Therefore, the aim of this study was to provide a systematic overview of the literature on microscopic pathology in the human DM1 brain.

METHODS

Search strategy

We conducted a systematic review of the literature and, where possible, aggregated homologous data. Prior to the commencement of data analysis, we registered the protocol of this study in PROSPERO, an international prospective register of systematic reviews, under ID 42018085626. It can be accessed at: http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42018085626.

A single author (RW) conducted a search in EMBASE (index period 1974–2017) and MEDLINE (index period 1887–2017) on December 11, 2017. AC, JR, and BK supervised the search, which was conducted in collaboration with an information specialist of the university library. We used free text and index search terms related to myotonic dystrophy type 1 and brain structures or regions (Appendix I). Additional studies were identified through cross‐referencing. Because we were unaware of the number, sizes, and types of studies to expect, we modified the eligibility criteria during the course of the study, as we felt the initial criteria to be too restrictive or too loose (see Appendix II).

Eligibility criteria and study selection

Studies were eligible if they reported on at least one microscopic pathological brain feature in ≥ 1 patient with confirmed clinical or genetic DM1 diagnosis according to the authors and no evidence of potentially relevant co‐morbidity (i.e. brain disease other than DM1) that could interfere with outcomes of interest. We excluded studies for reasons as follows: (1) unavailability of full text; (2) language of publication other than English, Dutch, French, or German; (3) double reporting (i.e. dual publication of same research results); and (4) conference abstracts. After deduplication, study selection comprised an initial round of title and abstract screening, followed by a second round of full‐text screening. Both screening rounds were independently performed by two authors (RW and KO). Differences in the selection of studies were discussed to meet consensus.

Data extraction

Data extraction was carried out independently by RW and KO. We extracted the following data from each selected study: study characteristics (first author, year of publication, brain areas, and pathological alterations under study); patient characteristics (number, sex, age at death, DM1 disease class, and clinical or genetic DM1 diagnosis). DM1 disease class was defined as either congenital, infantile, juvenile, adult, or late‐onset DM1. Where applicable, we also collected these variables for healthy or disease control subjects that were presented in the included studies.

Data synthesis

To summarize our findings, we classified the pathological alterations reported in the included studies and counted the number of studies reporting these pathological alterations. Accordingly, we subdivided our results on the basis of the most frequently reported alterations.

RESULTS

Search and selection results

Our primary search resulted in a total of 2099 records (Fig. 1). After deduplication and title and abstract screening, a total of 102 records remained for full‐text screening. We excluded 63 studies, mostly for reasons of absence of outcome of interest (n = 40), unsuitable publication type/conference abstract (n = 17), and language restriction (n = 5). We included 39 studies from our primary search and included an additional two studies after cross‐referencing, making a total of 41 studies (Fig. 2). ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³

Fig 1.

Literature search and study selection with a flow chart of literature search and study selection. *See text for reasons for exclusion. This figure has been created using Microsoft Visio 2007, saved in .jpg image format.

Fig 2.

Country of origin of studies. The world map shows country of origin of brain pathology studies in DM1. All studies originate from a total of seven countries. The map has been created with Mapchart.net (https://mapchart.net/) under a CC BY‐SA 4.0 license.

Study characteristics

A total of 130 DM1 patients were included in 41 studies (Table 1). However, as some research groups assessed different pathological aspects in studies on identical subjects (e.g. the group of Ono and colleagues), the number of unique subjects is likely considerably lower. Most studies included a limited number of participants; the largest series investigated the brain of 17 patients and the median number of patients was five. Healthy (n = 292) and disease controls (n = 598) were present in 23 and 16 studies, respectively. The most frequently included disease controls were patients with Duchenne muscular dystrophy (n = 253, two studies), Alzheimer's disease (n = 56, 11 studies), amyotrophic lateral sclerosis (n = 28, six studies) and progressive supranuclear palsy (n = 30, five studies). The most common pathologic features were: (1) protein or nucleotide deposits; (2) cellular alterations in neurons or glial cells; and (3) white matter alterations. For each feature, the relevant study and brain structures in which they were present are shown in Table 2.

Table 1.

Overview of included studies

Author, year, reference	ID	No. P (M:F)	Age at death (mean ± SD in years)	DM1 disease class ‡ (c:i:j:a:lo)	No. HC	No. DC	Brain areas	Histopathological changes evaluated
Rosman, 1967 13	1	4 (2:2)	47 ± 7	?	0	0	CC	Neuronal heterotopia; disordered lamination
Culebras, 1973 14	2	6 (4:2)	48 ± 9	‐:‐:+:+:‐	0	3	CC	Neuronal heterotopia; disordered lamination; astrocytes
							Thalamus	IIB
Wisniewski, 1975 15	3	1 (0:1)	48	‐:‐:‐:‐:+	0	0	Thalamus	IIB
Sarnat, 1976 16	4	4 (1:3)	0 †	+:‐:‐:‐:‐	0	0	CC	Neuronal size
							WM	Gliosis
Pena, 1980 17	5	2 (2:0)	0; 62	+:?:?:?:?	40	0	Thalamus	IIB
Young, 1981 18	6	4 (3:1)	0	+:‐:‐:‐:‐	0	0	WM	Gliosis
							BS: medullary pyramids	Gliosis
Ono § , 1989 19	7	4 (2:2)	62 ± 4	?	8	8	Thalamus	IIB
							SN	IIB; MB
							HPC	NFT; GVD; plaques
Yoshimura, 1990 20	8	1 (1:0)	61	‐:‐:‐:+:‐	3	0	CC; hypothalamus; NBM; BG; HPC; amygdala; BS; CB; olfactory bulb	Neuronal loss; neuronal size; neuronal degeneration; gliosis; NFT; plaques
							Thalamus	IIB; neuronal size; neuronal degeneration; gliosis
Garcia‐Alix, 1991 21	9	4 (4:0)	0 †	+:‐:‐:‐:‐	0	0	WM	Neuronal heterotopia; necrosis; hemorrhage; astrocytes
							CB	Neuronal heterotopia
Kiuchi, 1991 22	10	7 (4:3)	49 ± 8	?	18	0	CC: limbic cortex	NFT; plaques
							HPC	Neuronal loss
							BS: LC, raphe nuclei; WM	NFT
Hageman, 1993 23	11	5 (3:2)	0; 4	+:‐:‐:‐:‐	0	0	WM	Periventricular leukomalacia
							Thalamus	IIB
Moriuchi, 1993 24	12	17 (?:?)	50 ±?	?	0	312	CC	Neuronal loss
							?	NFT
Abe, 1994 25	13	1 (0:1)	54	?	0	0	CC; BS	Neuronal loss
							WM	Myelin loss
Oyanagi, 1994 26	14	5 (2:3)	53 ± 9	?	24	0	CC: temporal	NFT; GVD; Hirano; plaques
							Thalamus	IIB
							BG	IIB; NFT
							SN	IIB; MB
Lidang Jensen, 1995 27	15	2 (1:1)	0 †	+:‐:‐:‐:‐	0	0	CC	Neuronal size; neuronal heterotopia
							BS: hypoglossal nucleus	Neuronal size
Ono § , 1995 28	16	7 (4:3)	62 ± 5	?	8	0	BS: raphe nuclei	Neuronal size; gliosis
Ono § , 1995 29	17	1 (0:1)	53	‐:‐:‐:+:‐	0	0	BS: raphe nuclei, MRF	Neuronal size; gliosis
Ono § , 1996 30	18	7 (4:3)	62 ± 5	?	8	0	BS: MRF	Neuronal loss
Ono § , 1996 31	19	8 (5:3)	62 ± 4	?	0	9	thalamus	IIB
Vermersch, 1996 32	20	2 (2:0)	53; 61	‐:+:‐:+:‐	20	?	CC: entorhinal, inferior temporal	NFT; plaques
							thalamus; BG	IIB
							HPC; BS; amygdala	NFT
Ono, 1997 33	21	8 (5:3)	62 ± 4	?	8	9	SN	IIB
Watanabe, 1997 34	22	1 (1:0)	48	+:‐:‐:‐:‐	0	9	CC	Neuronal loss; neuronal size; neuronal heterotopia; gliosis; NFT; plaques
							Thalamus	IIB
							Amygdala	NFT
							HPC, NBM	Neuronal size
							BS: SN	NFT; gliosis; MB
							WM	Neuronal heterotopia
Ono § , 1998 35	23	8 (4:4)	63 ±?	?	12	0	BS: raphe nuclei	Neuronal size
Ono § , 1998 36	24	8 (4:4)	63 ±?	?	10	0	BS: MRF	Neuronal size
Mizukami, 1999 37	25	1 (1:0)	58	?	0	0	CC: entorhinal cortex	Neuronal loss; neuronal size; astrocytes; NFT
							Thalamus	IIB
							BS: SN, pons: LC, medulla	Neuronal size; Lewy bodies
							CB: ventral hemisphere	Neuronal loss; gliosis
							NBM	Neuronal loss; gliosis; NFT
							WM	Myelin loss; axonal loss; neuronal neuronal heterotopia
Spillantini, 1999 38	26	1 (0:1)	54	?	2	10	HPC	NFT; plaques
Endo, 2000 39	27	11 (7:4)		+:‐:‐:+:‐	30	0	CC	Neuronal heterotopia; disordered lamination; NFT
							Thalamus	NFT; IIB
							BG; BS	NFT; neuronal heterotopia
							CB	Neuronal heterotopia
							WM	Periventricular leukomalacia
Kawashima, 2000 40	28	6 (3:3)	59 ± 12	?	40	59	BG	SLI
Ono § , 2001 41	29	8 (4:4)	63 ± 5	?	8	10	CC; HPC; CB; BS: medullary arcuate nucleus	Neuronal size; neuronal loss
Kumada, 2002 42	30	7 (4:3)	53 ± 10	?	10	0	thalamus; CB; CC	Intranuclear inclusions
							SN	Neuronal size; gliosis; MB
Maurage, 2003 43	31	3 (3:0)	64 ± 2	?	15	32	CC	NFT
							WM	NFT
Jiang, 2004 44	32	10 (7:3)	56 ±?	‐:‐:‐:+:+	6	7	CC; CB; HPC; thalamus; WM: subcortical, corpus callosum; BS: SN, tegmentum	RNA foci
Maurage, 2005 45	33	3 (3:0)	59 ± 6	+:‐:‐:+:‐	3	2	CC: entorhinal; HPC; BS – LC, SN; CB	NFT
							SN	MB
Yamazaki, 2005 46	34	5 (3:2)	54 ± 20	?	0	101	WM: frontal	Tau‐positive fine granules
Oyamada, 2006 47	35	12 (6:6)	55 ± 8	?	4	0	CC; HPC: subiculum	NFT, plaques
							BS: SN, LC, central grey matter	NFT, plaques, Lewy bodies
Itoh, 2010 48	36	11 (5:6)	60 ± 5	?	0	0	CC; CB; BG; HPC; BS	Neuronal size
							CC: enthorhinal cortex; HPC; amygdala; BS	NFT, plaques
							Thalamus	IIB
							SN	MB
							SN, LC	Gliosis, Lewy bodies
							Deep WM	Gliosis
Dhaenens, 2011 49	37	5 (4:1)	60 ± 4	+:‐:‐:+:‐	3	0	CC: entorhinal cortex; HPC; amygdala	NFT
Yamazaki, 2011 50	38	5 (4:1)	57 ± 6	?	9	23	HPC	NFT; GVD
Nakamori, 2012 51	39	3 (2:1)	55 ± 6	?	3	4	HPC	NFT; GVD
Jinnai, 2013 52	40	9 (5:4)	60 ± 8	?	0	0	CC: entorhinal cortex; CB	Neuronal loss
							WM; cerebellar WM	Myelin loss; dilate perivascular spaces; axonal loss; gliosis; capillary hyalinization
							Entorhinal cortex; HPC	NFT; plaques
Michel, 2015 53	41	?	?	+:‐:‐:‐:‐?	0	0	?	RNA foci

^†Study included deceased newborns.

^‡Disease class: congenital‐onset DM1; infantile‐onset DM1; juvenile‐onset DM1, adult‐onset DM1 and late‐onset DM1, ⁵⁴ according to the authors; + denotes presence, − denotes absence. BG, basal ganglia; BS, brain stem; CB, cerebellum; CC, cerebral cortex; GVD, granulovacuolar degeneration; HPC, hippocampus; IIB, intracytoplasmic inclusion body; LC, locus coeruleus; MB, Marinesco body; MRF, medullary reticular formation; NBM, nucleus basalis of Meynert; NFT, neurofibrillary tangle; SLI, skein‐like inclusions; SN, substantia nigra; WM, white matter.

^§Not all patients in these studies were unique.

DC, disease controls; F, female; HC, healthy controls; M, male; No., number; P, patients.

Table 2.

Summary of histopathological changes, brain regions, and studies

Histopathological changes	Number of studies	Study IDs	Brain regions
1. Accumulations of material
Intracytoplasmic eosinophilic inclusion bodies	14	2; 3; 5; 7; 8; 11; 14; 19; 20; 21; 22; 25; 27; 36	Thalamus BG; BS: SN
Marinesco bodies	6	7; 14; 22; 30; 33; 36	SN
Neurofibrillary degeneration	19	7; 8; 10; 12; 14; 20; 22; 25; 26; 27; 31; 33; 34; 35; 36; 37; 38; 39; 40	Limbic system: amygdala; HPC; CC; limbic Cortex; olfactory bulb CB; BG; NBM; BS: raphe; hypothalamus WM
Granulovacuolar degeneration	4	7; 14; 38; 39	HPC; CC: temporal
Plaques	10	7; 8; 10; 14; 20; 22; 26; 35; 36; 40	Whole brain with focus on limbic system and BS
RNA foci	2	32; 41	CC; HPC; CB; thalamus; BS: SN WM: subcortical, callosal
Other accumulations	6	14; 25; 28; 30; 35; 36	CC: temporal; thalamus; CB; BG; BS: pons (locus coeruleus), medulla, SN
2. Cellular alterations
Neuronal alterations	21	1; 2; 4; 8; 9; 10; 12; 13; 15; 16; 17; 18; 22; 23; 24; 25; 27; 29; 30; 36; 40	Limbic system (CC: entorhinal; HPC) CC; BG; NBM; BS: SN, raphe nuclei, hypoglossal nucleus; medullary arcuate nucleus, MRF; CB WM
Glial alterations	15	2; 4; 6; 8; 9; 16; 17; 22; 25; 30; 36; 40	Limbic system (CC: entorhinal cortex; HPC; olfactory bulb; amygdala) CC; thalamus; BG; NBM; CB; BS: raphe nuclei, medullary pyramids, SN, LC; hypothalamus WM
3. White matter pathology	14	4; 6; 9; 10; 11; 13; 22; 25; 27; 31; 32; 34; 36; 40	WM: deep; subcortical; callosal

The most/best studied brain regions are indicated in bold font.

BG, basal ganglia; BS, brain stem; CB, cerebellum; CC, cerebral cortex; HPC, hippocampus; LC, locus coeruleus; MRF, medullary reticular formation; NBM, nucleus basalis of Meynert; SN, substantia nigra; WM, white matter.

PROTEIN AND NUCLEOTIDE DEPOSITS

Intracytoplasmic eosinophilic thalamic inclusion bodies

A total of 14 studies examined post‐mortem neuropathological samples of 66 patients for the presence of thalamic inclusion bodies (TI) (Table 2). These thalamic inclusion bodies are round to oval structures located in the cytoplasm of neurons that stain homogeneously eosinophilic with HE staining. Anti‐ubiquitin antibody positivity and tau negativity have been demonstrated with immunohistochemical methods. ³¹ No more than one to two TI were present per neuron. They were not found in post‐mortem brains of patients with congenital DM1 that died at a young age. ¹⁷ , ²³ They were found in a variable number (18–100%) of adult DM1 patients and in a variable percentage of thalamic cells (3.5–35%), depending on the study. TI were not specific to DM1 as has been found in unaffected controls and controls with other degenerative brain diseases, although the percentage of affected neurons in DM1 exceeded that in the control populations. ¹⁹ The latter finding was not present in a previous study on TI, in which the number of inclusions was higher in healthy controls than in one DM1 patient. ¹⁷

Marinesco bodies in the substantia nigra

A total of six studies evaluated the presence of Marinesco bodies (MB) in the substantia nigra neurons in post‐mortem brains of 31 patients. ¹⁹ , ²⁶ , ³⁴ , ⁴² , ⁴⁵ , ⁴⁸ Marinesco bodies are eosinophilic intranuclear bodies in the pigmented neurons of the substantia nigra (Fig. 3). MBs are not specific to DM1, as they have been observed in other brain diseases, such as Alzheimer's and amyotrophic lateral sclerosis, and in physiological aging. ¹⁹ , ⁴² MB may be accompanied by eosinophilic intracytoplasmic inclusion bodies that highly resemble thalamic inclusion bodies at the structural and ultrastructural (i.e. electron microscopy) level. ¹⁹ As with these intracytoplasmic antibodies, immunohistochemistry has demonstrated anti‐ubiquitin antibody positivity of MB. Whereas one study found a higher load of MB in DM1 patients compared to controls with systemic or neurodegenerative disease, ¹⁹ equal loads in patients versus controls were found in substantia nigra neurons in another study. ⁴²

Fig 3.

Hitological features of an MB in the substantia nigra. Black arrowhead depicts a Marinesco body (an eosinophilic structure) adjacent to a similar‐sized nucleolus (a basophilic structure). This image is kindly provided by Dr. Kusters.

Neurofibrillary tangles

A total of 19 studies reported on the presence or absence of tau pathology in 104 DM1 patients (Table 2). Most studies detected neurofibrillary tangles (NFT) in all or most of the patients that were evaluated. ²² , ⁴³ , ⁴⁵ , ⁴⁷ , ⁴⁹ Reportedly, the brain areas that demonstrated the highest burden of neurofibrillary tangles (NFTs) were the entorhinal cortex, (subdivisions of) the hippocampus, the parahippocampal cortex, the temporal cortex, and the amygdala. Other, often less affected areas included the olfactory bulb, ²⁰ brain stem, ⁴⁵ , ⁴⁷ basal ganglia, ²⁰ , ⁴⁷ cerebellum, ⁴⁵ and other (non‐temporal lobe) parts of the neocortex ²⁰ , ³⁴ , ⁴³ , ⁴⁵ , ⁴⁹ The load of NFTs was higher in DM1 patients than in controls in three studies. ²² , ⁴⁵ , ⁴⁷ In contrast to the aforementioned studies, Moriuchi et al. ²⁴ found NFT in only two out of 17 DM1 patients, and Ono et al. ¹⁹ found no or few NFTs in the hippocampus in DM1 patients. Studies reported no or incidental “plaques” accompanying the tau‐related neuropathological changes. ²⁰ , ²² , ³² , ³⁴ , ⁴⁵ , ⁴⁷ Studies demonstrated that the NFT in DM1 are preferentially composed of tau proteins devoid of the amino acids encoded by exons 2 and 3 of the microtubule associated protein tau (MAPT) gene. ³² , ⁴⁵ , ⁴⁹

Nuclear accumulations

Two studies demonstrated the presence of abnormal, mutant RNA accumulations in the nucleus of brain cells using fluorescence in situ hybridization techniques. ⁴⁴ , ⁵³ These accumulations have been coined “RNA foci.” Jiang et al. ⁴⁴ observed these foci diffusely throughout the brain of adult DM1 patients, in neurons of the cortex, hippocampus, thalamus, and brainstem, including the substantia nigra. Of note, the cerebellum displayed small foci in the Purkinje cells but not in neurons of the molecular or granule cell layer. Foci were sometimes multiple and were also occasionally observed in oligodendrocytes of white matter in subcortical and callosal areas. ⁴⁴ More recently, Michel et al. ⁵³ demonstrated RNA foci GVD in the temporal cortex that coincided in the human brain in the earliest (i.e. fetal and embryonal) stages of development.

Other accumulations

Four studies (total n = 17) noted the presence of granulovacuolar degeneration (GVD), involving electron‐dense granules within double membrane‐bound cytoplasmic vacuoles, in the temporal lobe/hippocampal area of DM1 patients. ¹⁹ , ²⁶ , ⁵⁰ , ⁵¹ GVD were also found in other neurodegenerative diseases. ⁵⁰ Ono et al. ¹⁹ (n = 4 patients) found few GVD in the hippocampal area, and two out of five patients from the study by Oyanagi et al. ²⁶ had “several” GVD in the temporal cortex that coincided with NFTs and Hirano bodies. In hippocampal pyramidal neurons, one group demonstrated a strong relationship between GVD and tau pathology. ⁵⁰ , ⁵¹ Skein‐like ubiquitin‐positive inclusions in the striatum were observed in aging controls and in patients with various neurodegenerative disease, including five out six patients with myotonic dystrophy type 1. ⁴⁰

CELLULAR ALTERATIONS

Neuronal loss

In four post‐mortem cases of children with congenital onset DM1, no neuronal loss was found in the brain on autopsy. ²¹ In adults, severe neuronal loss was demonstrated in a proportion of patients as follows: (1) dorsal raphe nucleus and superior central nucleus of the midbrain; (2) pontine raphe and pontine reticular formation; and (3) dorsal medullary central nucleus, ventral medullary central nucleus, medullary arcuate nucleus, and subtrigeminal medullary nucleus in the medulla. ²⁸ , ²⁹ , ³⁰ , ⁴¹ The authors noted this neuronal loss in the brainstem in association with hypersomnia and hypoventilation. ²⁸ , ³⁰ In contrast, Mizukami et al., ³⁷ Itoh et al., ⁴⁸ and Lidang Jensen et al. ²⁷ reported no loss of neurons in the (pontine and medullary) brain stem. In the substantia nigra, mild neuronal loss was observed in two case reports ³⁴ , ³⁷ ; two case series demonstrated neuronal loss in zero out of seven and two out of 11 patients in this region. ⁴² , ⁴⁸ Similarly, neuronal loss was observed in the cerebellum in the case in Mizukami et al. ³⁷ but absent in two case series on 11 and nine patient evaluating cerebellar granule and Purkinje cells, respectively. ⁴⁸ , ⁵² In the cerebral cortex, Maurage et al. ⁴⁵ (n = 3), Abe et al. ²⁵ (n = 1), Ono et al. ⁴¹ (n = 8), Watanabe et al. ³⁴ (n = 1), and Lidang Jensen et al. ²⁷ (n = 2) observed no neuronal loss. Other studies observed cortical neuronal loss in a variable proportion of patients: Mizukami et al. ³⁷ (one out of one), Itoh et al. ⁴⁸ (six out of 11), and Jinnai et al. ⁵² (seven out of nine). Cortical neuronal atrophy was observed by Yoshimura et al. ²⁰ and Watanabe et al. ³⁴ The basal ganglia did not demonstrate neuronal loss. ²⁵ , ³⁴ , ⁴⁸ We did not find studies that evaluated neuronal loss in the thalamus. In the hippocampus, mild neuronal loss was found by Watanabe et al. ³⁴ (n = 1) but not by Ono et al. ⁴¹ (n = 8) and Itoh et al. ⁴⁸ (n = 11). In the adjacent entorhinal cortex, however, neuronal loss was observed by Jinnai et al. ⁵² (9 out of 9) and Itoh et al. ⁴⁸ (6 out of 11). Neuronal loss was also found in the nucleus basalis of Meynert, ³⁴ , ³⁷ and in the locus coeruleus. ³⁷ , ⁴⁸ With two exceptions, ³⁰ , ⁴¹ most studies did not report how neuronal loss was specifically evaluated. Neuronal loss was quantified qualitatively or semi‐quantitatively. The study by Ono et al., ⁴¹ in which design‐based stereological techniques were used, is a notable exception.

Neuronal heterotopia

Heterotopic neurons were found in the white matter in three studies ¹³ , ²¹ , ³⁴ that evaluated the brains from nine patients. Four brains were from patients with congenital DM1 that died shortly after birth; the other five were from DM1 patients that deceased at ages 36 to 51. Heterotopic neurons were found both in subcortical and deep white matter.

Reactions of glial cells

Eight studies evaluated the presence of gliosis in the brain of 37 DM1 patients. In the two largest studies (n = 11 and n = 9), gliosis was observed diffusely throughout the deep white matter. ⁴⁸ , ⁵² Itoh et al. ⁴⁸ also found astrogliosis in two out of 11 brains in the locus coeruleus and substantia nigra. In contrast, Kumada et al. ⁴² observed no gliosis in the substantia nigra in any of seven brains. Ono et al. found gliosis in six out of eight patients in the raphe nuclei and superior central nucleus of the brainstem that accompanied neural loss. ²⁸ , ²⁹ Yoshimura et al. ²⁰ and Mizukami et al. ³⁷ observed cortical gliosis. In the case described by Yoshimura et al., ²⁰ cortical gliosis was accompanied by diffuse gliosis throughout the brain, including the cerebellum, the brain stem, and the limbic system. In another case report, no cortical gliosis was found. ³⁴ Studies did not typically specify the type of gliosis (e.g. reactive astrogliosis, microgliosis/activated microglia cells, or oligodendrocyte response).

WHITE MATTER PATHOLOGY

Various neuropathological findings in white matter were reported in 14 studies (Tables 1 and 2). There was loss of myelin from deep and subcortical white matter, and widened perivascular spaces (Figs 4, 5), giving the white matter an “état criblé” appearance. The largest study (n = 11) reported loss of axons accompanying the myelin loss, ⁴⁸ whereas in two case reports myelin loss coincided with “relative axonal sparing.” ²⁵ , ³⁷ Itoh et al. ⁴⁸ also found capillary hyalinization and fibrillary gliosis.

Fig 4.

Perivascular dilatation in the cerebrum. In the temporal lobe, enlarged perivascular spaces are observed on sections, stained Klüver–Barrera, from a DM1 patient (B) but not a control individual (A).

Fig 5.

Perivascular dilatation in the cerebrum. Enlarged perivascular spaces are identified by polarized light imaging in the same patient as in Figure 4.

DISCUSSION

The various and anatomically diffuse microscopic brain alterations in myotonic dystrophy type 1 are a reflection of the heterogeneous clinical features and imaging characteristics of brain involvement of the disease (Fig. 6). On the basis of our findings, we classified microscopic brain pathology in DM1 as: (1) protein and nucleotide deposits; (2) cellular anomalies; and (3) white matter pathology (which will be discussed in more detail in the next sections).

Fig 6.

Summary of brain imaging findings. †Inconsistent findings across studies. BS, brain stem; CB, cerebellum; CC, cerebral cortex; HPC, hippocampus; SN, substantia nigra; WM, white matter. Brain MRI slice images are provided by Nevit Dilmen, used under Attribution‐ShareAlike 3.0 Unported (CC BY‐SA 3.0) license [https://creativecommons.org/licenses/by‐sa/3.0/deed.en], downloaded from commons.wikimedia.org.

Protein and nucleotide deposits

The presence of various deposits has been noted in histopathologic studies in DM1. The first descriptions include the intracytoplasmic eosinophilic bodies in the thalamus and the MBs in the substantia nigra. Their clinical relevance is uncertain, as MB could be found in persons as young as 21 years of age and are not associated with a specific disease. ⁵⁵ However, their prevalence in the substantia nigra increases with age and they could be considered as a marker for neurodegeneration. ⁵⁶ Immunohistochemically, MB are generally positive for proteins of the ubiquitin–proteasome system. ⁵⁷ Their increased occurrence in the substantia nigra in DM1, if validated in controlled, larger studies, might imply a relationship between protein degradation systems, abnormal proteins resulting from aberrant splicing in DM1, and neurodegeneration. Interestingly, echography studies have demonstrated hyperechogenicity of the substantia of the substantia nigra in a proportion of DM1 patients, in comparison to controls. ⁵⁸ Furthermore, the presence of accumulations and neuronal loss in the brainstem (see below) is interesting in light of abnormal levels of dopamine and serotonin metabolites in the cortex and the brain stem, respectively, in animals studies. ⁵⁹ These findings invite the study of in vivo (i.e. CSF) and ex vivo neurotransmitter metabolism in DM1‐affected individuals.

More recently, the demonstration of RNA “foci,” toxic accumulations of mutant RNA, in the brain provide a direct link between the molecular pathophysiology of DM1 and visible histopathological alterations. ⁴⁴ , ⁵³ The CTG repeat expansion in the 3′‐untranslated region of the DMPK is transcribed into long CUG‐containing mRNAs that remain in the nucleus and may sequestrate RNA‐binding proteins and, consequently, lead to transcriptional dysregulation. ⁶⁰ The formation of these alterations is likely located upstream in the cascade of pathophysiological brain changes in DM1, and they have been shown to start very early in the disease process. Interestingly, the development of brain “foci” is well captured in animal models and, consequently, is an aspect that is open to future therapeutic study in DM1.

Tau protein is encoded by a single gene (MAPT) on chromosome 17q21 and represented by different isoforms generated by alternative splicing from the MAPT gene. An expanding spectrum of diseases now classified as “tauopathies” is characterized by NFTs that consist of aggregates of hyperphosphorylated and abnormally phosphorylated isoforms of tau proteins. ⁶¹ , ⁶² In several of the tauopathies, aberrant splicing of the MAPT RNA has been demonstrated, and aberrant alternative tau splicing could promote tau pathology. ⁶³ Tau missplicing results in isoform‐specific impairment in normal physiological function and enhanced recruitment of excessive tau isoforms into the pathological process. ⁶³ In DM1, recent work has suggested an interplay between toxic RNA, disordered splicing, and the dysregulation of microtubule‐associated protein tau. ⁶⁴ , ⁶⁵ In DM1, the “foci,” consisting of accumulations of toxic RNA, are known sequester RNA‐binding proteins, such as splicing factors, consequently leading to dysregulated alternative splicing in neurons. ⁶⁴ , ⁶⁵ Accordingly, alterations of tau alternative splicing have been reported in DM1 with missplicing of exons 2 and 6 in the brain. Thus, DM1 is characterized by the preferential accumulation of the 0N3R isoform. ³² Moreover, several studies provide evidence of the effects of missplicing on tau aggregation, protein–protein interactions, microtubule stabilization, and axonal transport, which may be relevant to myotonic dystrophy type 1. ⁶⁴ , ⁶⁵

In DM1, neurofibrillary degeneration (NFD) is preferentially seen in parts of the limbic system including (parts of) the hippocampus and adjacent cortex, but other brain regions may also show signs of tau pathology. Clinical features of DM1, such as diminished olfaction, apathy, and learning and memory disorders may reflect involvement of the limbic system; ⁵ , ⁶⁶ , ⁶⁷ however, the exact relation with NFD in this area is unknown. It should be noted that tau pathology can be present in brains of individuals that have shown healthy aging from a clinical perspective, and that tau missplicing is not restricted to tauopathies but can also be present in non‐tau proteinopathies. ⁶² , ⁶³ It remains to be seen to what extent tau pathology in DM1 is causal to the heterogeneous brain phenotypes of the disease. Besides this uncertainty about the clinicopathological correlations, the spatiotemporal progression of NFD in DM1 is undiscovered at present. PET‐Tau imaging offers a means to study such progression in vivo and could be applied to DM1 in future studies. ⁶⁸ In addition, the relationship of tau pathology with other histopathological features of DM1 should be further studied. Understanding the pathophysiology of DM may increase our knowledge of the complexity of the pathological interplay among RNAopathy, spliceopathy, and proteinopathy implicated in tauopathies and other neurodegenerative diseases. ⁶⁵

Cellular alterations

Neuronal loss in the brainstem was inconsistently present across studies. Neuronal loss in the brainstem may relate to clinical features of excessive daytime sleepiness and ventilatory dysfunction. ²⁸ , ²⁹ Brainstem neuronal loss was not captured in a DM mouse model with respiratory dysfunction. ⁶⁹ In the cerebellum, neuronal loss has been the subject of limited study but does not seem to be a generalized finding. In common with other trinucleotide repeat expansion, somatic instability (i.e. increase in size) of the CTG repeat was lower in cerebellar tissue than in other brain regions. ⁷⁰ Of note, most structural imaging studies (voxel‐based volumetry and diffusion tensor imaging) are less suitable for the study of the brain stem and the cerebellum in comparison to the hemispheral grey matter. However, alterations in brainstem raphe echogenicity in DM1, noted by Krogias and colleagues, may be a possible correlate to brainstem histopathological changes. ⁵⁸ In the cerebral cortex, histopathological results were mixed, with some studies demonstrating no neuronal loss and others only in a proportion of patients. Cortical neuronal loss is in line with magnetic resonance spectroscopy studies that found decreased concentrations of N‐acetylaspartate, mostly in cortical grey matter. Possible explanations for the divergent findings between studies may include heterogeneous patient populations, (unrecognized) co‐morbidities, and technical reasons. Importantly, with few exceptions, neuronal loss was “quantified” in a qualitative (i.e. present/absent) or semi‐quantitative manner in all studies. Although gross changes may be detected with these methods, the number of a specific cell type cannot be simply deduced from the number of cross‐sectional profiles found in thin tissue sections, as this parameter also depends on cell volume, tissue orientation, and tissue atrophy. ⁷¹ , ⁷² Consequently, pending more advanced studies, these results on neuronal loss should be interpreted with caution. In contrast to studies on neuronal loss, we found no studies that evaluated synaptic density in the brain of DM1 patients. Such studies would be of interest, as imaging studies have found grey matter volume loss throughout the cortex in human DM1 patients. ⁷³ , ⁷⁴ One correlate of grey matter volume loss based on voxel‐based‐morphometry could be neuronal loss, but given the findings in this study, this is unlikely the only explanation. Reduced synaptic density may provide a complementary explanation that could be worth exploring. ⁷⁵ In cell and animal models, there is evidence that the CTG repeat expansion of DM1 may affect synaptogenesis and synaptic function. ⁵⁹ , ⁷⁶ Finally, as a developmental alteration, heterotopic neurons may be found in the white matter of DM1 patients. Their clinical relevance in relation to central nervous system features of congenital and infantile‐onset DM1 is currently unknown.

White matter changes

Degenerative characteristics of white matter include loss of myelin in combination with a varying amount of axonal loss, dilation of perivascular spaces, gliosis, and capillary hyalinization in deep and subcortical white matter. These findings are in line with white matter alterations captured with in vivo neuroimaging: dilated perivascular spaces, T2/FLAIR hyperintensities, and alterations in microstructurem, as demonstrated with DTI imaging. ⁷⁴ , ⁷⁷ , ⁷⁸ Although the white matter histological alterations in DM1 share characteristics with cerebral small‐vessel disease, other characteristics of the latter, such as microbleeds and lacunes, have not been observed in DM1. Consequently, chronic brain ischemia seems unlikely to be the major mechanism through which white matter lesions originate. Renard et al. (2018) proposed increased burden due to microvascular changes and lack of drainage of interstitial fluid and degraded protein products as a mechanism for anterior temporal white matter lesions in DM1. ⁷⁹ This hypothesis was based on the co‐localization of dilated perivascular spaces and white matter hyperintensities in the anterior temporal pole, which DM1 shares with the monogenetic small vessel disease cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL). ⁷⁹ , ⁸⁰ Of interest, dilated perivascular spaces were recently demonstrated to be associated with the burden of neurofibrillary tangles in Alzheimer's disease. ⁸¹ At the basis of the association of dilated perivascular spaces with various tauopathies, including DM1 and Alzheimer's disease, could be a common pathophysiological mechanism (e.g. astrocytic dysfunction) worth further exploring.

CONCLUSION

Summarizing our results, clinical and neuroimaging heterogeneity of DM1 reflect the various histopathological changes in DM1. Arbitrarily, histopathological findings can be classified as: (1) protein and nucleotide deposits; (2) cellular alterations including neuronal loss and gliosis; and (3) white matter pathology. We observed both anomalies that could be considered developmental in nature (e.g. heterotopic neurons) and neurodegeneration. Indeed, many of the neuropathological findings in DM1 share similarities with other neurodegenerative diseases and can be seen in “physiological” aging (albeit to a lesser degree). The latter findings are compatible with the concept of DM1 as a progeroid/accelerated aging disease. ⁸² , ⁸³ However, in common with the highly variable and unique admixture of clinical brain features in DM1, the specific combination of histopathological findings seems to be unique. For some aspects such as neurofibrillary degeneration, great progress has been made in linking the pathophysiological concepts of RNAopathy and spliceopathy to histopathological changes. ⁶⁴ These interrelations remain elusive for other pathological findings at this time.

In a seminal editorial on “myotonic dystrophy as a brain disorder”, Ashizawa (1998, p. 291)noted that “availability of brain autopsy materials in DM is limited and the MR imaging data are difficult to correlate with the histological findings.” ⁸⁴ Unfortunately, some 20 years later, we must conclude that this has not changed. Most published studies are case reports and small case series. These stand in sharp contrast to a large and increasingly sophisticated body of literature on neuroimaging in DM1. ⁸ , ⁹ However, studies that correlate in vivo neuroimaging with post‐mortem histopathological findings are virtually non‐existent. Clearly, there is great need for larger follow‐up studies that evaluate the associations between the many neuroimaging findings and microscopic alterations in relation to clinical DM1 features. Furthermore, more detailed morphometric studies of animal model brains would inform us of which aspects of the human DM1 histopathology are captured by the (mouse) models. To facilitate such studies, international collaboration with an established set of clinical and imaging variables and a logistical pipeline that allows identical handling (e.g. design‐based stereology for morphometric aspects) of brain autopsy samples is essential and should be high on the DM1 scientific research agenda. ¹¹ , ¹²

DISCLOSURE

The authors declare no conflict of interest.

1. In the original study protocol, we only included studies with patients with genetically proven DM1. This criterion was modified to the criterion presented in Appendix III. Because a lot of our studies were published before or shortly after the discovery of the DMPK gene in 1992, we felt that this item was too restrictive as we had to exclude approximately 50% of otherwise eligible studies based on this criterion.

2. In the original study protocol, we included studies that obtained results exclusively by macroscopic and/or genetic examinations. This criterion was modified to the criterion presented in Appendix III. We felt that the inclusion of studies on each study level (i.e. macroscopic, microscopic, and genetic) would be too extensive. We decided to focus on microscopic alterations and excluded those that did not present at least one microscopic outcome.

PubMed (836 hits)

EMBASE (1263 hits)

(“Myotonic Dystrophy”[Mesh] OR Curschmann*[tiab] OR Dystrophia Myotoni*[tiab] OR Dystrophic Myoton*[tiab] OR Myotonia Atrophica*[tiab] OR Myotonia Dystrophi*[tiab] OR Myotonic Dystroph*[tiab] OR Steinert*[tiab] OR Steinert's[tiab]) AND (“Brain”[Mesh] OR “Myelin Sheath”[Mesh] OR Amygdal*[tiab] OR Brain[tiab] OR Brains[tiab] OR Brainstem*[tiab] OR Caeruleus[tiab] OR Callos*[tiab] OR Cerebr*[tiab] OR Ceruleus[tiab] OR Coeruleus[tiab] OR Corpus Callos*[tiab] OR Cortex*[tiab] OR Cortical[tiab] OR Cortices[tiab] OR Corticospinal*[tiab] OR Diencephal*[tiab] OR Encephal*[tiab] OR Epithalam*[tiab] OR Fascicul*[tiab] OR Glia[tiab] OR Glial[tiab] OR Gray Matter*[tiab] OR Grey Matter*[tiab] OR Hippocamp*[tiab] OR Hypothalam* [tiab] OR Limbic*[tiab] OR Lymbic*[tiab] OR Mesencephal*[tiab] OR Mesolimbic*[tiab] OR Metencephal*[tiab] OR Midbrain*[tiab] OR Myelencephal*[tiab] OR Olfactory Pathway*[tiab] OR Parahippocampal*[tiab] OR Perforant Nerve Tract*[tiab] OR Perforant Path*[tiab] OR Telencephal*[tiab] OR Ventricl*[tiab] OR Ventricul*[tiab] OR White Matter*[tiab])

(myotonic dystrophy/ OR (Curschmann* OR Dystrophia Myotoni* OR Myotonia Atrophica* OR Myotonia Dystrophica* OR Myotonic Atroph* OR Myotonic Dystroph* OR Steinert* OR Steinert's).ti,ab,kw.) AND (exp brain/ OR exp. brain tissue/ OR exp. brain ventricle/ OR myelin/ OR myelin sheath/ OR perforant nerve tract/ OR (Amygdal* OR Brain OR Brains OR Brainstem* OR Caeruleus OR Callos* OR Cerebr* OR Ceruleus OR Coeruleus OR Corpus callos* OR Cortex* OR Cortical OR Cortices OR Corticospinal* OR Diencephal* OR Encephal* OR Epithalam* OR Fascicul* OR Glia OR Glial OR Gray Matter* OR Grey Matter* OR Hippocamp* OR Hypothalam* OR Limbic* OR Lymbic* OR Mesencephal* OR Mesolimbic* OR Metencephal* OR Midbrain* OR Myelencephal* OR Olfactory Pathway* OR Parahippocampal* OR Perforant Nerve Tract* OR Perforant Path* OR Telencephal* OR Ventricl* OR Ventricul* OR White Matter*).ti,ab,kw.)

Number	First author	Year	Journal	Notes
Language restrictions
1	Tomi H	1986	Clinical Neurology	Article in Japanese
2	Mitake S	1989	Rinsho Shinkeigaku	Article in Japanese
3	Yoneyama S	1992	Rinsho Shinkeigaku	Article in Japanese
4	Haranaka M	2000	No To Hattatsu	Article in Japanese
5	Iwata T	2009	Rinsho Shinkeigaku	Article in Japanese
Conference abstracts
1	Dhaenens C	2009	Medizinische Genetik
2	Gourdon G	2009	Medizinische Genetik
3	Kang Y	2009	Medizinische Genetik
4	Kuru S	2010	Neuropathology
5	Saito Y	2010	Neuropathology
6	Caillet‐Boudin M	2011	Alzheimer's and Dementia
7	Ikeda K	2011	Neuropathology
8	Jinnai K	2011	Neuropathology
9	Kuru S	2011	Neuropathology
10	Matsuda M	2012	Neuropathology
11	Imanishi A	2013	Sleep
12	Komene T	2013	Neuropathology
13	Oya Y	2013	Neuropathology
14	Servais L	2015	Neuromuscular Disorders
15	Iwasaki Y	2016	Clinical Neurology
16	Sano T	2016	Clinical Neurology
17	Shioya A	2016	Clinical Neurology
Cell/tissue culture studies
1	Hernandez‐Hernandez O	2006	Journal of Neuroscience Research
2	Ghanem D	2009	FEBS Letters
3	Marteyn A	2011	Cell Stem Cell
4	Velazquez‐Bernardinho P	2012	Molecular Biology Reports
5	Reddy K	2014	Nucleic Acids Research
Animal studies
1	Balasubramanyam A	1998	Journal of Comparative Neurology
2	Seznec H	2001	Human Molecular Genetics
3	Sato S	2002	Human Molecular Genetics
4	Westerlaken J	2003	Brain Research
5	Garcia‐Lopez A	2008	PLoS One
6	Oude Ophuis R	2009	Muscle Nerve
7	Charizanis K	2012	Neuron
8	Wang E	2012	Cell
9	Hernandez‐Hernandez O	2013	Rare Diseases
10	Poulos M	2013	Human Molecular Genetics
11	Choi J	2016	Scientific Reports
12	Wang P	2017	Human Molecular Genetics
Genetic studies
1	Ishii S	1996	Human Genetics
2	Gennarelli M	1999	Neuromuscular Disorders
3	Korade‐Mirnics Z	1999	Human Molecular Genetics
4	Sergeant N	2001	Human Molecular Genetics
5	Leroy O	2006	Biochimica et Biophysica Acta
6	Leroy O	2006	Journal Neuroscience Research
7	Dhaenens C	2008	Experimental Neurology
8	Axford M	2011	Journal of Medical Genetics
9	Lopez Castel A	2011	Human Molecular Genetics
10	Suenaga K	2012	PLoS One
11	Axford M	2013	PLoS Genetics
12	Hernandez‐Hernandez O	2013	Brain
13	Goodwin M	2015	Cell Reports
CSF studies
1	Hirase T	1984	Brain and Development
2	Martinez‐Rodriguez J	2003	Sleep
3	Ciafaloni E	2008	Neurology
4	Winblad S	2008	European Journal of Neurology
Double reporting
1	Ono S	1987	Journal of the Neurological Sciences
2	Yoshimura N	1990	Clinical Neuropathology
Significant neurological comorbidities
1	Kobayashi S	2005	Journal of the Neurological Sciences
No patient tissue used
1	Gennarelli M	1995	Biochemical and Biophysical Research Communications
2	Whiting E	1995	Human Molecular Genetics
3	Wang J	2007	Journal of Neurochemistry
4	Cleary J	2010	Nature Structural & Molecular Biology