Introduction
Neutral lipid storage disease (NLSD) is a rare inherited disorder of lipid metabolism characterized by a defect in the catabolic pathway of triacylglycerols resulting in systemic accumulation of triglycerides in cytoplasmic droplets, notably in the leukocytes (
Jordan’s anomaly). Two different recessive forms have been described: (1) the NLSD with myopathy (NLSD-M) caused by molecular defects in the adipose triglyceride lipase gene (
ATGL, also called patatin-like phospholipase domain-containing 2,
PNPLA2) coding for a rate-limiting enzyme catalyzing the first step of hydrolysis of triglycerides [
1], and (2) the NLSD with ichthyosis (NLSD-I or
Chanarin-
Dorfman Disease) due to mutations in the
ABHD5 gene (also known as Comparative Gene Identification-58,
CGI-
58) coding an homonymous activator protein of ATGL [
2].
Typical NLSD-I presentation include an early onset ichthyosis (nonbullous congenital ichthyosiform erythroderma, NCIE) associated with liver and mild skeletal muscle involvement [
3]. Clinically NLSD-M is mainly characterized by adult-onset progressive myopathy with variable association of cardiomyopathy, hepatic steatosis and short stature. Muscle weakness is diffuse, but frequently predominant in proximal upper limb and distal lower limb muscles, leading to a “man in the barrel” phenotype frequently associated with neck extensor weakness [
4].
In the last years, evidence is accumulating on the usefulness of muscle imaging in defining specific patterns of muscle involvement in inherited muscle diseases to help clinicians in the diagnostic workup [
5‐
7]. Globally, metabolic myopathies have not yet been widely investigated and most of the studies concern Pompe Disease, which present a characteristic pattern of muscle involvement [
8‐
10].
Muscle imaging in NLSD have not been systematically investigated. Only few studies reported muscle MRI findings in small series of patients with NLSD-M [
4,
11,
12]. They showed a heterogeneous involvement mainly affecting the posterior compartment of the thighs, anterior and posterior compartment of the legs, deltoid, trapezius, infra- and supraspinatus. By contrast, typical NLSD-I shows milder involvement although cases presenting with severe and diffuse fatty replacement of muscles have been described [
13].
The aim of this study was to assess the skeletal muscle involvement by muscle imaging (MRI or/and CT) in a cohort of NLSD patients from the Italian network for NLSD, and in particular to establish whether there is a consistent pattern of muscle involvement.
Discussion
We present a systematic study of muscle imaging in a large cohort of NLSD patients. Imaging data have been collected from patients of the Italian Network of NLSD, harboring different genetic mutations and different clinical disease severity. Previous data from single cases and smaller cohorts show results that are globally in agreement with the pattern of muscle involvement we have recognized in this work [
4,
11,
12].
Even if the quality of imaging resolution by CT scan is lesser informative than MRI, the different degrees of muscle involvement was largely comparable between CT patients and MRI patients.
Despite clinical manifestations suggest a major clinical impairment of upper limbs in NLSD-M, muscle imaging demonstrated a more severe involvement of the lower limbs along the entire disease course (Figs.
1,
4). In the milder patient (P10), even if the upper limbs show no fatty replacement, an initial fatty replacement in lower limbs could be detected.
Medial gastrocnemius, soleus, gluteus minimus and semimembranosus are the most severely affected muscles in all patients. Moreover, leg (medial gastrocnemius and soleus) and pelvis (gluteus minimus) muscles were constantly more affected than thigh muscles (semimembranosus) suggesting that fatty replacement starts both at the pelvis and legs. Gluteus medius, biceps femoris (long head), adductor magnus and longus, gastrocnemius medialis and tibialis posterior appeared to be less severely affected. Notably, tibialis posterior is affected early in the disease course, contrary to the majority of myopathies, in particular LDMG and distal myopathies, in which tibialis posterior is frequently spared even in the late-end stages of the disease course [
17,
25‐
29]. Muscles of anterior compartment of the leg are variably affected during the disease course while quadriceps and psoas become affected in the late-end stages of disease. Psoas sparing can help to recognize NLSD from other myopathies in the late-end course of the disease when specificity of the MRI pattern involvement disappears or the muscle biopsy may be not informative [
17,
25‐
29]. Sparing of sartorius, gracilis and pectineus is constantly observed in all cases.
Once the upper limbs are involved, the first and constantly affected muscle is the infraspinatus. Supraspinatus, trapezius, deltoid, serratus and paraspinal muscles are also frequently affected, but a lesser degree. In the arms, the anterior compartment is markedly involved. By contrast sternocleidomastoid, pectoralis minor and major and triceps brachii are constantly spared in all cases and sometimes appear to be hypertrophied, differently from other reported myopathies [
30].
Taken together, this combination of muscle involvement in lower and upper limbs composes a constant pattern and represents a signature of NLSD-M in all stages of the disease. In the clinical context, muscle MRI may help to recognize NLSD-M among different conditions affecting predominantly proximal upper limbs with neck extensor weakness (“man in the barrel” or “dropped head” syndromes) [
31‐
34] or among different metabolic myopathies, notably those associated with lipidosis on the muscle biopsy [
35]. Nevertheless, more muscle imaging data are needed from these other conditions to establish if the “MRI signature” in NLSD-M is specific when compared to other muscle lipidosis.
In comparison to other muscular dystrophies and in particular with LGMD, it is notable that in the late-end stages of disease, gracilis, sartorius and biceps femoris (short head) are frequently spared, but differently from NLSD-M, calf involvement appears constantly later than thigh involvement (BMD, sarcoglycanopathies and LGMD2I) [
17,
29]. In dysferlinopathies (LGMD2B), even if posterior compartments of the thigh and calf are frequently involved, the quadriceps involvement, and particularly the vastus lateralis, appears early in the disease course, whereas in NLSD-M is constantly involved lesser than posterior compartment.
In distal myopathies, muscle replacement starts in the legs as in NLSD-M. Nevertheless, in Desminopathies semitendinosus is frequently early replaced and semimembranosus is spared until the end stage of the disease [
25]. By contrast myotilinopathy and ZASP myopathy could show similar muscle involvement in the leg and thigh to NLSD-M even if lateral gastrocnemius is frequently spared [
16]. Another myopathy with a similar pattern of myotilinopathy is LGMD1D, but peroneal compartment if frequently spared in this condition [
36].
Another important finding highlighted by our work is the characteristic aspect of fatty replacement constantly observed in all NLSD-M patients, that had never been reported in previous studies concerning NLSD. Indeed, in several muscles, fatty replacement was not homogenous along the whole length of the muscles and in early stage of the disease “patchy” areas of total fatty replacement were close to areas of muscle sparing. Moreover, in the most affected patients, “islands” of muscle sparing were present in the middle of large areas of fatty replacement. This unusual pattern of fatty replacement could represent an additional “disease signature” of NLSD-M, and could reflect different pathophysiological mechanisms of disease compared to muscular dystrophies. Nevertheless more physiological studies are necessary to support this hypothesis.
In our NLSD-I patients, muscle involvement was very mild and only showed hyperintense STIR images in the calf muscles. Nevertheless, these data should be further assessed and confirmed in a larger cohort of patients, because the overall number of muscle MRI observations of NLSD-I patients is scanty [
13].
In conclusion we describe the muscle imaging findings in a large cohort of NLSD patients. Our data provides evidence that muscle imaging can identify characteristic alterations for NLSD-M, such as a consistent pattern of muscle involvement associated to the presence of “spotted” areas of fatty replacement.
Larger cohorts are needed to assess if a distinct pattern of muscle involvement exists also for NLSD-I.