An update on MSA: premotor and non-motor features open a window of opportunities for early diagnosis and intervention

Multiple system atrophy (MSA) is a sporadic, adult-onset, progressive, rare neurodegenerative disorder of uncertain etiology. It manifests with diverse clinical features, characterized mainly by a combination of parkinsonism with a poor response to levodopa treatment, cerebellar features, and autonomic failure. The definite diagnosis of MSA can still only be established pathologically with the presence of glial cytoplasmic inclusions (GCI) at postmortem [1‐3].

In the past, a variety of terms were used to describe MSA. In 1900, Joseph Jules Dejerine and André Thomas first reported two cases presenting with cerebellar ataxia and pathological lesions in the cerebellum and brainstem, described as olivopontocerebellar atrophy. Sixty years later, Shy and Drager described two patients with parkinsonism and autonomic symptoms including orthostatic hypotension [4] and Van der Eecken et al. reported cases with parkinsonism and striatonigral degeneration at autopsy in the same year. The term ‘MSA’ was first introduced in 1969 by Graham and Oppenheimer to represent all three neurological entities: olivopontocerebellar atrophy (OPCA), the Shy–Drager syndrome, and striatonigral degeneration (SND) [5]. Papp et al. first demonstrated the presence of argyrophilic glial cytoplasmic inclusions in the central nervous system of patients diagnosed with MSA in 1989 [6]. These inclusions were later reported to be positive for α-synuclein, and nowadays, alpha-synuclein inclusions in oligodendroglia are the recognized neuropathologic hallmarks of MSA and may even represent a primary pathologic event [7]. Together, Parkinson’s disease, dementia with Lewy bodies, and MSA are called α-synucleinopathies [8]. The mechanisms by which α-synuclein, a neuronal protein, ends up in the oligodendroglia causing pathology in MSA remain unclear. There is growing evidence that oligodendroglial pathology and “prion-like” spreading of misfolded α-synuclein may be the primary events in MSA leading to neurodegeneration [7‐9]. In addition, proteasomal and mitochondrial dysfunction [10], dysregulation of myelin lipids [11, 12], genetic factors [13], microglial activation [14], neuroinflammation [15], proteolytic disturbance, autophagy [16], and other factors contributing to oxidative stress [17] have been reported as part of the pathogenesis of MSA.

MSA is an orphan disease [18] with an estimated incidence of 0.6–3 per 100,000 people [19, 20]. The incidence increases with age reaching 12/100,000 in those aged over 70 [21], with no difference between men and women [22, 23]. Prevalence estimates range from 1.9 to 4.9 [24] and may reach up to 7.8 per 100,000 after the age of 40 [20, 22, 23].

Given its varied clinical manifestation, MSA is frequently misdiagnosed, especially at disease onset. An autonomic presentation of MSA can be indistinguishable from pure autonomic failure (PAF). PAF is currently considered an idiopathic, sporadic, rare neurodegenerative disorder characterized by autonomic failure without other neurological symptoms or signs [25]. Recently, PAF has received more attention from researchers, partly due to shared common features with other forms of synucleinopathy. Furthermore, recent studies show that a significant proportion of patients with PAF eventually develop DLB, PD, or MSA over time. There are certain features that could predict these conversions [26, 27] raising the question of whether PAF is indeed a distinct disease entity or simply a prodromal disease stage of alpha-synucleinopathies.

A patient presenting with parkinsonism with autonomic involvement may be misdiagnosed as suffering from Parkinson’s disease. Ataxias resulting from diverse etiologies such as toxins, immune-mediated, or genetic forms—for example fragile X–associated tremor ataxia syndrome, spinocerebellar ataxia (especially type 6), or late-onset Friedreich’s ataxia [2]—can all mimic the cerebellar MSA phenotype. False-positive MSA diagnoses are also frequent. About 20% of cases diagnosed in life as MSA subsequently have a pathology confirming PD or DLB [28].

MSA is characterized by worsening of motor and non-motor features over an average of 10 years, with a fast progression from disease onset, helping to distinguish it from similar degenerative conditions, described above [29]. Approximately 50% of patients require walking aids within 3 years from the onset of motor symptoms [30]; 60% require a wheelchair after 5 years [31] with a median time to becoming bedridden of 6–8 years [30]. However, a more benign MSA variant with longer survival of over 15 years has been reported in pathology-confirmed cases [[2, 32], as well as an aggressive MSA phenotype, with a very short disease duration of less than 3 years [33]. Older age at onset [30, 31, 34‐36], a parkinsonian phenotype [29, 35], early development of severe autonomic failure [30, 35, 37, 38], and severe urinary retention and nocturnal stridor [39, 40] are all negative prognostic factors, whereas a cerebellar phenotype [34] and later onset of autonomic failure [32] predict slower disease progression. Bronchopneumonia, urosepsis, and sudden death are the main causes of death in MSA. Sudden death, which often occurs at night, is thought to result from either acute bilateral vocal-cord paralysis or acute disruption of the brainstem cardiorespiratory drive.

Striatonigral degeneration (SND) and olivopontocerebellar atrophy (OPCA) are pathological subtypes of MSA. Four levels of pathological severity have been described on autopsy of MSA brains. The typical α-synuclein immunoreactive inclusion pathology (GCIs) within oligodendrocytes [41] are at the most severe end and are required for a definite postmortem diagnosis of MSA. Less frequent are neuronal cytoplasmic (NCI) and neuronal nuclear inclusions (NNI), glial nuclear (GNI), astroglial cytoplasmic inclusions, and neuronal threads, also composed of α-synuclein. The next level of severity is characterized by selective neuronal loss and axonal degeneration involving multiple regions of the nervous system with predominant involvement of the striatonigral and OPC systems, followed by myelin degeneration with pallor and reduction in myelin basic protein (MBP), with accompanying astrogliosis. The final and most mild form of pathology is microglial activation alone [17].

Glial cytoplasmic inclusions (GCIs) and the resulting neurodegeneration commonly occur across multiple CNS sites. In addition to the striatonigral and OPC involvement, CGIs are found in the autonomic nuclei of the brainstem [locus coeruleus (LC), nucleus raphe, dorsal vagal nuclei, etc.], spinal cord, sacral visceral pathways [42], and the peripheral nervous system [43], defining MSA as a truly multi-system disease [7].

Currently, MSA is considered an oligodendrogliopathy with a secondary neuronal involvement. Both a reactivation of α-syn gene (SNCA) [44] hypothesis and an uptake from neurons or from the extracellular environment [45, 46] have been proposed. A prion-like mechanism of propagation in MSA [47, 48] predicts the existence of “strains” with defined incubation times and patterns of neuropathology. Strains of alpha-synuclein (fibrils and ribbons) represent different conformational polymorphs of the protein. The existence of different alpha-synuclein strains may be related to different phenotypes. For example, fibrillar α-syn polymorphisms underlies different propagation patterns of α-syn pathology [49], while α-synuclein in glial cytoplasmic inclusions (GCI-α-Syn) forms more compact and about 1,000-fold more potent structures than Lewy bodies-α-Syn in seeding α-Syn aggregation, consistent with the highly aggressive MSA [50]. Also, it has been shown that oligodendrocytes, and not neurons, transform misfolded α-Syn into a GCI-like strain, emphasizing the fact that different intracellular milieus generate distinct α-Syn strains and the GCI-α-Syn maintains its high seeding activity when propagated in neurons.

The diagnostic criteria for MSA define three levels of certainty: possible, probable, and definite MSA [3] (Table 1). However, a widening range of pathological and clinical presentations have been described in MSA, including several subtypes of MSA that do not fit into the current classification (Table 2) [51].

Based on the predominant clinical phenotype, MSA is categorized into MSA-P when parkinsonism is predominant and is associated with SND, and MSA-C, with OPCA, when associated with dominant cerebellar features. The combination of both forms, “mixed” MSA, also exists [3]. In Asian populations, the majority of MSA cases—about 70–80% [52]—are of MSA-C type, whereas in the Caucasian population, MSA-P type predominates (about 67–84%) [29, 53]. MSA-P is characterized by parkinsonism with rigidity, postural instability with a tendency to fall, and poor response to levodopa [54]. The motor symptoms are usually symmetrical [55]. Rest tremor is rare, whereas irregular postural and action tremor may occur [57]. MSA-C is associated with cerebellar ataxia affecting arms, legs, action tremor, downbeat nystagmus, and hypometric saccades [56]. Hyperreflexia and Babinski sign may occur in 30–50% of patients, while abnormal postures, such as bent spine, antecollis, and hand or foot dystonia, are rare [56]. Dysphonia, repeated falls, drooling, dysphagia, dystonia, and pain occur in advanced stages of the disease [57]. Spinal myoclonus in MSA-C caused by α-synuclein deposition in the spinal cord has also been reported [58].

The onset of motor symptoms is on average around 56 years of age (± 9 years), with both sexes equally affected [59]. However, up to 75% of MSA cases have a prodromal phase with non-motor symptoms, such as cardiovascular autonomic failure, orthostatic hypotension, urogenital and sexual dysfunction, REM-sleep behavior disorder, and respiratory disorders. These may precede the motor presentation by months to years [60]. Furthermore, up to 95% of MSA patients experience non-motor symptoms at some point during the course of the disease with autonomic failure, in particular urogenital (urinary incontinence and impaired detrusor muscle contractibility) and cardiovascular symptoms being frequent and early features of MSA [61].

In this review, we describe this wide clinical spectrum of features now seen in MSA with a focus on the non-motor symptoms and the premotor phase, and provide an up-to-date overview of the current understanding in this fast-growing field.

The distribution of neurodegenerative changes in patients with MSA is broadly reflected by α-synuclein-positive oligodendroglial cytoplasmic inclusions [2]. During the course of the disease, most patients present with a combination of gastrointestinal, cardiovascular, urogenital, or thermoregulatory abnormalities of different severities [2, 36]. Several brain regions of MSA patients are severely depleted of dopamine and norepinephrine including the corpus striatum, nucleus accumbens, substantia nigra, locus coeruleus, hypothalamus, and septal nuclei [62]. A severe loss of A5 noradrenergic neurons in the pontine tegmentum has been shown in MSA, leading to respiratory and cardiovascular manifestations, along with similar severe losses of noradrenergic neurons in the locus ceruleus [63, 64]. The loss of C1-group neurons (neurons that synthesize epinephrine) contributes to orthostatic hypotension in MSA, with the severity of neuronal loss in MSA being greater than in PD and DLB [65, 66]. Impaired reflex release of vasopressin in response to hypotension or hypovolemia in MSA may be a consequence of degeneration of noradrenergic projections to the magnocellular vasopressin neurons from A1 neurons, localized in the caudal ventrolateral medulla [65, 67]. Reduced orexin immunoreactivity, likely associated with sleep apnea syndrome, has been observed in the nucleus basalis of Meynert in MSA patients [68]. Tyrosine hydroxylase neuronal loss in the periaqueductal gray of MSA, similar to Lewy body disease, has also been reported [69].

Involvement of the insular cortex has been related to non-motor symptoms in MSA, including autonomic dysfunction and arousal [70]. Activation of the subgenual and pregenual portions has been associated with cardiovagal responses in normal people; activation of the dorsal anterior cingulate cortex (the midcingulate cortex), part of the salience network, has been shown to be associated with sympathetic activation related to behavioral arousal [71].

The brainstem regions, in particular those where the respiratory centers are located, are frequently affected in MSA. The ventral medullary arcuate nucleus, likely responsible for respiratory chemosensitivity, degenerates in MSA [72]. In addition, the pre-Botzinger complex of the medulla and the neurokinin-1 immunoreactive neurons, responsible for respiratory rhythmogenesis, are similarly markedly affected [73, 74]. The neuronal loss in MSA, from the dorsal motor nucleus of the vagus (DMV), located in the dorsomedial medulla [75], may contribute to baroreflex-triggered cardioinhibition and a respiratory sinus arrhythmia [76].

REM-sleep behavior disorder (RBD) is commonly seen in this disease and may relate to pathology affecting the pontomedullary brainstem nuclei in MSA [77, 78]. During REM sleep, atonia is controlled by the pontomedullary structures [the magnocellular reticular formation (MCRF), sublaterodorsal (SLD), and pedunculopontine and laterodorsal (PPN/LDT) nuclei] [77]. Elimination of atonia during REM sleep, caused by lesions in SLD/SC and MCRF (even unilaterally), likely leads to dream enacting behavior [79]. Depletion of cholinergic neurons in the PPN/LDN complex has been shown in autopsy studies of patients with MSA [80]. PPN may have a modulatory role in REM-related phenomena rather than a primary role for the atonia during REM sleep [75].

Preganglionic and postganglionic sudomotor denervation probably underlies the anhidrosis seen in MSA, with more postganglionic dysfunction seen later in the disease [76]. Postganglionic cardiac sympathetic denervation also has been shown to occur in a minority of cases of MSA[81}. The preganglionic neurons innervating the detrusor muscle [82] and loss of neurons of the Onuf nucleus innervating the sphincter are related to early and severe bladder dysfunction.

The prodromal non-motor phase of MSA often precedes the motor symptoms by several months to years. Reports of pathologically proven MSA cases in which no classic motor symptoms (cerebellar, parkinsonian, or pyramidal tract) developed during the disease course [195] emphasize the need to consider premotor symptoms in the early evolution of the disorder. Early symptoms in MSA are frequently autonomic in nature (in 73%) [83]. In men, erectile failure typically occurs 4.2 (± 2.6) years prior to diagnosis. After erectile failure, orthostatic hypotension—a consistent hallmark of MSA [191]—fatigue following exercise, urinary urgency or hesitancy, and violent dream enactment behavior consistent with REM behavioral sleep disorder, are the most frequent initial symptoms. For most patients, the onset of motor symptoms is 1 year after the onset of autonomic symptoms. Although challenging to interpret, these symptoms can provide a window of opportunity for disease-modifying interventions if recognized in time, therefore, raising a suspicion of MSA diagnosis.

The presence of autonomic failure, erectile dysfunction, urinary urgency, nocturia, sleep problems (RBD), postural dizziness, and respiratory disturbances at presentation, in the absence of ataxia and parkinsonism, should be regarded as suggestive of premotor MSA [196]. Up to half of patients later diagnosed with MSA initially were classified as having pure autonomic failure, as they did not exhibit motor symptoms at the first presentation to a doctor [138, 179].

A pattern is emerging in MSA; with non-motor symptoms preceding the classic motor features showing a distinctive pattern of early genitourinary dysfunction followed by orthostatic hypotension in association with sleep disorders such as RBD, sleep apnoea, excessive snoring, and stridor [196]. Above all, respiratory dysfunction is the most suggestive symptom of MSA in the presence of other premotor symptoms. At the same time, in the premotor phase, olfaction and cognition [197] remain relatively intact compared with other parkinsonian disorders (Fig. 1).

The diagnostic challenge in the premotor phase of MSA is the substantial overlap with the non-motor symptoms of other neurodegenerative disorders, in particular Parkinson’s disease and pure autonomic failure. However, at the time of diagnosis, autonomic dysfunction in MSA is typically more severe compared to that in PD and autonomic dysfunction preceding the development of parkinsonism supports MSA as a more likely diagnosis [138]. Conversely, a case with parkinsonism or cerebellar ataxia in the absence of erectile dysfunction or autonomic failure would make an MSA diagnosis less likely [83].

The spectrum of premotor symptoms of MSA is broad (Fig. 2) and should, therefore, be investigated more vigorously to achieve an early diagnosis of this serious illness and to avoid unnecessary therapeutic interventions [196].

Symptomatic treatment remains the current standard of care in MSA. Management strategies are focused on symptom control with medication for parkinsonism, autonomic lability, bladder and bowel dysfunction, and mood problems [198] (Table 3). Deep brain stimulation is not recommended for MSA [199].

New strategies targeting α-synuclein are in progress [54, 200]. Unfortunately, most clinical trials have failed to show positive results due in part to the small number of enrolled patients, the inevitable accidental recruitment of non-MSA patients, and the relatively late intervention for disease-modifying treatments [201]. The only clinical trial showing positive results was intra-arterial and intravenous injection of autologous mesenchymal stem cells (MSCs), which delayed disease progression as measured by the unified MSA rating scale in patients with MSA-C [202]. Targeting the ‘prion-like’ cell-to-cell propagation of α-synuclein using an α-synuclein immunotherapy is an appealing approach and has been studied in animal models of MSA [203]. A combination of a single-chain antibody (CD5-D5) and anti-inflammatory treatment (lenalidomide) ameliorated gliosis, α-synuclein accumulation, and behavioral deficits in MBP-α-synuclein transgenic mice [204].

The clinical spectrum of MSA and, especially, the non-motor features at disease onset represent a growing challenge for clinicians, researchers, and drug trials. Besides the current diagnostic criteria, several MSA subtypes, not fully fitting the currently used diagnostic criteria, are emerging, highlighting the complexity of the disease as well as the need for revised diagnostic tools. The ability to make an earlier diagnosis and thus intervention in MSA may lead to an increased quality of life, a better prognosis, as well as a greater chance of finding disease-modifying drugs.