Introduction

Facioscapulohumeral muscular dystrophy (FSHD, OMIM no. 158900) is the third most common hereditary myopathy, after Duchenne dystrophy and myotonic dystrophy [1]. It has been estimated to affect 1 in 8333 people worldwide [2]. Linkage analysis allowed the identification of the FSHD genetic locus on 4q35 chromosome [3]. The classical form of the disease (referred to as FSHD1), in fact, has been associated with DNA rearrangements (contraction) in this region, consisting of an array of tandemly repeated units known as D4Z4. However, approximately 5% of patients display the typical disease phenotype without carrying the D4Z4 contraction. This alternative form is termed FSHD2 and it has been associated with pathogenetic mutations in Structural Maintenance of Chromosomes Flexible Hinge Domain-Containing 1 (SMCHD1, 18p11.32) and DNA Methyltransferase 3 Beta (DNMT3B, 20q11.21) genes [4]. Although two distinct forms of the disease (FSHD1 and FSHD2) have been described, recent literature indicated that FSHD2-associated genes can also act as disease modifiers in FSHD1, rising the need for testing FSHD2 genes even in FSHD1 patients [3, 4].

The molecular diagnosis of FSHD1 is still based on traditional methods as Southern blotting that is expensive and labor intensive and requires large amounts of DNA. Several efforts have been made to set up alternative molecular methods, such as molecular combing (MC) [5]. On the other hand, characterization of FSHD2-related genes is made either by traditional or modern technologies (direct sequencing and NGS, respectively) which allow high-throughput, less expensive, and reliable results. In this context, prominent attention should be given to the training of biologists and laboratory technicians which should properly know not only the latest but also the oldest technologies available to perform molecular diagnosis of peculiar disorders.

Clinical features

FSHD is a slowly progressive muscular dystrophy that typically involves facial and shoulder muscles. The clinical course of the disease can be highly variable, alternating stable and worse conditions. The transmission pattern generally follows an autosomal dominant pattern, although reduced penetrance and variable expressivity complicate the identification of affected subjects within families [1]. In addition, up to 30% of cases are due to de novo mutations and, approximately half of them, result from a post-zygotic mutation leading to mosaicism [4, 6]. The onset is usually between 6 and 20 years, suddenly in the adulthood. In the early stage of the disease, the muscle involvement can be asymmetrical, and the weakness is so slight and slowly progressive that many cases remain subclinical for several years. Generally, the first signs of the disease include difficulties in raising the arms over the head and scapular winging. Facial weakness has also been observed at the initial clinical stage of the disease and, in some cases, many years before the diagnosis. The involvement of facial muscles is generally underestimated because of the little impact on activities of daily living. The affected facial muscles are the orbicularis oculi, the zygomaticus, and the orbicularis oris which appear weak and thin. Patients are therefore unable to firmly close the eyes, to inflate cheeks, and to whistle. Masticatory, lingual, and extraocular muscles are not involved in FSHD [4]. The scapular girdle is the most involved region: trapezius and pectoral muscles at the initial stage; the sternocleidomastoid, serratus magnus, rhomboid, erector spinae, latissimus dorsi later on; and ultimately, the deltoid muscles [7]. Shoulder bones protrude out on the back, causing thereby scapular winging and clavicles sticking out and rounded shoulders. Biceps are usually less involved than triceps and brachioradialis muscles, although biceps can present a thinner proximal trait compared to the distal one (“Popeye” effect) [8]. Pelvic girdle is involved some years after the onset of the disease, leading to pelvic muscle weakness and, subsequently, to mild lordosis and pelvic instability. Pretibial involvement leads to foot drop and gait instability [8]. Beevor’s sign (the abnormal upward movement of the umbilicus when the patient raises the head from a supine position) is typical in the advanced stage of the disease, when up to 20% of patients become wheelchair-dependent [7]. Cardiac involvement (rhythm disorders and cardiomegaly) is rare. Although some studies have hypothesized anticipation phenomenon, no molecular and clinical findings confirmed this hypothesis and phenotype worsening in subsequent generations remains a rare event [9].

Although some severity markers have been recently described, in most of the cases, the prognosis of the disease is unpredictable [10]. Sudden and dramatic worsening of the disease can occur after long periods of stable conditions. In addition, the clinical spectrum is widely variable among unrelated patients, including subjects with slight muscle weakness, who are unaware of being affected, as well as patients who are wheelchair-dependent [8]. This variability in disease clinical expression has been exemplified by a study on monozygotic male twins, who carry the same genetic mutation but are affected by FSHD at dramatically different extents [4]. Before the advent of molecular diagnosis, penetrance of FSHD was evaluated considering the clinical examination of patients at risk and their age, resulting to be between 83 and 95% at 20 years of age [11]. However, a higher penetrance of the disease has been reported in males compared to females, who appeared to be mostly asymptomatic or minimally affected [7]. Recently, a prospective cross-sectional study reported the age of onset of FSHD as a severity marker. Patients with an early-onset FSHD (observed in 7–15% of all FSHD cases) have been found to experience a more severe and progressive disease with 57% patients being dependent on wheelchair compared to 20–30% of classical FSHD cases. In addition, age at onset has also been described as a prognostic marker in relation to the frequency of systemic complications [10].

Genetic aspects of FSHD1

FSHD1 has been associated with DNA rearrangements (contraction) in a polymorphic region known as D4Z4 (4q35) that is characterized by an array of tandemly repeated units extending for 3.3 kb. In normal conditions, the D4Z4 array varies from 10 to 100 repeated units (RU), whereas FSHD patients have less than 10 RU [12]. The number of repeats can be determined by Southern blotting and hybridization with the p13E-11 probe, which is able to recognize the D4Z4 region. However, the interpretation of the p13E-11 hybridization pattern is complicated by the fact that the probe can also recognize an additional locus on the 10q26 region that is nearly completely homologous to the 4q35 region (> 98%) but it is not related to FSHD1 [4]. Moreover, the high degree of sequence homology between 4qter and 10qter loci facilitates inter-chromosomal exchange resulting in total and partial translocations. These translocation alleles are derive from ancient founding events but, indeed, complicate even more the molecular diagnosis of the disease [6, 13].

Many studies suggested that D4Z4 contractions are not the only required conditions for FSHD manifestation [14]. The genomic architecture of the 4q35 region and its flanking regions has been therefore deeply investigated, in order to search for other variables contributing to FSHD etiopathogenesis. Two sub-telomeric variations distal to D4Z4 were identified on chromosome 4, namely 4qA and 4qB alleles. Both are equally distributed in the general population, but only 4qA has been associated with FSHD and it is referred to as “permissive allele” [15].

Proximal to D4Z4 array, a simple sequence length polymorphism (SSLP) exists, whose sequence length can range between 157 and 180 bp. The analysis of 4q sub-telomeric variants flanking the D4Z4 array revealed different haplotypes on 4qter, including 4A159, 4A161, 4B163, 4A166, and 4A168. Currently, all 4qA alleles are pathogenic except for 4A166. Case-control studies revealed a strong association between FSHD and 4qA161, which is the most common allele contributing to create a “permissive” genetic background for disease etiopathogenesis [12, 15]. On this subject, 4A159, 4A163, 4A166H, and 4A168 haplotypes have also been described as uncommon permissive alleles [7].

Moreover, each unit of the D4Z4 array contains a copy of the retrogene Double Homeobox 4 (DUX4, OMIM no. 606009), which is normally expressed during early embryonic development and in testis whereas it is silenced in somatic tissue [15]. Somatic repression of DUX4 requires a combination of epigenetic mechanisms necessary for maintaining a repressive chromatin condition. In FSHD1, instead, the contracted allele induces a chromatin relaxation which results in the hypomethylation of the D4Z4 and a variegated expression of DUX4 in myonuclei [16]. DUX4 expression has been found to be toxic in muscle tissue, hampering myogenic differentiation, increasing oxidative stress and, ultimately, causing muscle atrophy [17]. Interestingly, DUX4 transcripts are produced only in the presence of a polymorphic sequence (ATTAAA) residing in the 4qA variant that acts as a polyadenylation signal (PAS) to stabilize the transcript of DUX4 in muscle tissue of FSHD patients [18]. In contrast, the lack of PAS in 4qB or 10q alleles does not allow the aberrant expression of DUX4 in muscles, although 4qB is known to permit the physiological expression of DUX4 in early embryonic development (at 4-cell stage) likewise 4qA alleles [4, 19,20,21,22,23]. The size of the 4q array is determined by digestion with the restriction enzyme EcoRI, which generates fragments of variable size in relation to the number of D4Z4 repeats. EcoRI fragments are considered as one the key elements to determine the clinical phenotype, age of onset, disease course, and age at loss of ambulation [4, 7]. An inverse relationship exists between the D4Z4 repeat size and the severity and progression of disease. In fact, individuals with ≥ 11 RU (> 43 kb) are normally unaffected, whereas 1–3-RU (10–17 kb) fragments are associated with more severe FSHD1 [1]. Difficulties in predicting the disease severity have been encountered in patients carrying 4–7-RU (20–30 kb) fragments because of the high clinical heterogeneity, although this repeat size is frequently observed among FSHD1 patients. The availability of predicting factors for disease severity is critical for genetic counseling, treatment, or prevention of potential complications. In this context, the D4Z4 deletion size appears to be somewhat predictive of the overall rate of disease progression, although other sources of variation can affect disease severity and frequency of systemic features. Usually, patients with severe early-onset disease have 1–3-RU repeats, suggesting a much more robust correlation between disease severity and larger D4Z4 contractions [10].

Genetic aspect of FSHD2

Up to 5% of FSHD patients do not show D4Z4 contraction on 4q35. These patients show identical clinical features of FSHD1 cases and carry similar DNA hypomethylation levels and heterochromatin markers on D4Z4 repeats, but they differ for other genetic signatures [24]. This alternative form of the disease is referred to as FSHD2 (OMIM no. 158901) and has been recognized to be distinct and much less common than the classical type. In FSHD2 patients, D4Z4 hypomethylation is observed on all D4Z4 repeat arrays without D4Z4 contractions, although the repeat size is usually smaller or borderline (between 8 and 20 RU), making the region more susceptible to further hypomethylation [25, 26]. More than 80% of FSHD2 cases are caused by the inheritance of two independent genetic variations: a heterozygous loss-of-function mutation in SMCHD1 (OMIM no. 614982) gene and a 4qA permissive allele with the PAS inducing the toxic expression of DUX4 [18, 27, 28].

The encoded protein by SMCHD1 is a distant member of the highly conserved SMC protein family that is an essential component of the cohesin/condensin protein complexes [3]. In muscle cells, SMCHD1 has been reported to directly bind the D4Z4 repeat array, suggesting that it is essential for maintaining a repressive chromatin structure in somatic cells [18]. More than 52 SMCHD1 mutations associated with FSHD have been reported up to date [28]. These mutations decrease the DNA binding activity or the enzymatic activity of SMCHD1, resulting in hypomethylation and relaxation of chromatin in all D4Z4 repeat arrays on chromosome 4. In the presence of a 4qA allele, this is thought to bring enhancers and promoters (which are normally kept away by SMCHD1) closer enough to interact and contribute to create a transcriptionally permissive environment for the expression of DUX4 [18]. However, SMCHD1 mutations do not cause FSHD2 when combined with non-permissive 4qB alleles, suggesting the existence of a digenic mechanism for FSHD2 pathogenesis. Approximately 20% of FSHD2 individuals carrying hypomethylation at D4Z4, a SMCHD1 mutation, and a permissive 4q haplotype are asymptomatic, indicating thereby the existence of incomplete penetrance as for FSHD1 [28]. Some FSHD families have been identified with both FSHD1- and FSHD2-associated mutations. These patients appear to be more severely affected with respect to individuals carrying only the D4Z4 contraction. This observation showed that SMCHD1 can act as a disease modifier in FSHD1 patients [29]. In addition to SMCHD1, other environmental and/or (epi)genetic factors are likely to be involved in modifying disease severity and clinical manifestation. On this subject, exome sequencing was performed on eight FSHD families carrying D4Z4 hypomethylation without evidence of a pathogenic SMCHD1 mutation. In two families, a potentially damaging heterozygous variant in DNMT3B (OMIM no. 602900) gene was identified [30]. The gene encodes a known D4Z4-chromatin modifier, suggesting that DNMT3B mutations may act as disease modifiers in FSHD patients. As for SMCHD1, the effect of DNMT3B mutations on DUX4 expression and disease manifestation strictly depends on the presence of PAS and on the size of the D4Z4 repeat array [4].

Genetic counseling in FSHD

Clinical and molecular characteristics of FSHD have to be considered in order to provide patients and their families with an adequate genetic counseling. Considering that FSHD is an autosomal dominant disorder with variable penetrance and expressivity, efforts to find measurable factors able to discriminate FSHD patients have been so far tried with limited success. In particular, some specific genotypes (such as the shortest D4Z4 allele) have been associated with a more severe phenotype and a higher disease penetrance [31,32,33,34]. On the other hand, longer alleles seem to be less penetrant. In the clinical practice, the lack of a precise genotype-phenotype association and the high intra-familial variability make the genetic counseling very difficult [14].

The genetic counseling should be offered to each patient with a suspicion or a clinical diagnosis of FSHD. Pre-test and post-test genetic counseling should be performed in each case. Patients should be adequately informed concerning the available molecular tests, the possible expected and unexpected findings, and their implications for their own health and of their relatives, in order to enable them to make an informed choice when considering molecular testing [35, 36]. Given FSHD penetrance and expressivity, it is recommended that patients fully understand the main implications of a positive result before undertaking the test. During pre-test genetic counseling, it is essential to address the following topics: (i) the test is optional but can support or confirm the clinical diagnosis, (ii) the technologies which are applied, (iii) limitations (sensitivity and specificity) of the test, (iv) incidental findings, (v) the timing, (vi) presentation of the final report [37, 38]. During the diagnostic procedure, the molecular confirmation of FSHD can be helpful, especially to make an early diagnosis of disease. In fact, early stages of FSHD can be promptly treated with some physiotherapy interventions aimed to prevent the acquisition of harmful habits. However, patients sometimes prefer to avoid genetic testing, because of the lack of well-defined therapies. The details about technologies are always difficult to explain to patients, because of the laborious molecular approach utilized for FSHD diagnosis. In particular, the counselor should explain all of the analytical steps performed to achieve reliable results, including both traditional (Southern blotting and PFGE (pulsed-field gel electrophoresis); LGE (linear gel electrophoresis)) and next-generation sequencing (NGS) methods. Furthermore, the identification of D4Z4-deleted alleles on the repetitive region of 4q35 should be confirmed with a segregation analysis (if available), especially when the identification of a short allele does not correlate with clinical features because of incomplete penetrance. In this case, a detailed pre-test genetic counseling should be offered to each family member that undergoes genetic test, paying attention to explain the possible occurrence of this event. Moreover, segregation analysis can have other implications since familial links might not be confirmed (lack of biological paternity). If this information is not important for the explanation of the disorder within the family, counselors may decide to not communicate it, unless explicitly requested by the patient before undertaking the test. The time required for molecular characterization is hard to predict. In a family in which the disorder co-segregates with a contracted 4q35 allele, the test to be performed in family members is quite easy and exclusively based on the detection of the known molecular defect. Sporadic cases, which are negative to D4Z4 contraction, are usually screened for mutations in SMCHD1 and DNMT3B, requiring thereby longer times for analysis. Furthermore, epigenetic evaluations could be useful in selected cases, to identify the cause of FSHD. Occasionally, the first-level test (Southern blot and PFGE, LGE) and the segregation analyses could reveal a non-canonical segregation pattern due to translocation events, hampering thereby the achievement of a conclusive diagnosis, because of the lack of molecular techniques able to differentiate between different rearrangements [39, 40]. During the pre-test genetic counseling, the counselor should also provide an anticipation of the final report including the results of the genetic test, in order to get patients ready for the interpretation of the reported data. The post-test genetic counseling is focused on the explanation of the final report and the clinical significance of the results for patients and their family members. Patients should be informed about the risk of transmission of pathogenic mutations and the recurrence risk of disease in case of identification of the disease-associated mutation. Penetrance and variable expressivity should be extensively discussed during post-test genetic counseling. Although an accurate genotype-phenotype prevision is not possible, patients can make prenatal testing to detect the pathogenic mutation on the DNA of the fetus. In this case, the counselor should inform the patient about the limitations of the genetic test. In fact, a prenatal test for FSHD can only search for a disease-associated mutation, but it cannot predict the phenotype of the fetus, because of incomplete penetrance and variable expressivity of FSHD.

Pre-implantation genetic diagnosis (PGD) is theoretically possible, but the big amount of DNA required for the PFGE analysis is unsuited to the technologies utilized for PGD. This is the reason why PGD may be performed through linkage analyses, but the detection of short D4Z4 alleles at 4q35 (FSHD1) is generally not feasible due to the chromosomal position of the repetitive trait. Limitations of the linkage analysis for detecting 4q alleles should be communicated. In fact, the D4Z4 region is located at the telomeric end of the long arm of chromosome 4, making the selection of appropriate short tandem repeat (STR) markers to perform linkage analysis difficult. In addition, STRs are very rare within the region of interest, complicating the possibility of investigating other markers in case of non-informative results. Furthermore, the great genomic distance between the D4Z4 region and STRs significantly increases the recombination risk that should always be considered when PGD or prenatal diagnosis is performed. Invasive prenatal testing is therefore recommended in each PGD for FSHD, in order to verify the occurrence of recombination events [41]. Concerning the analysis of SMCHD1 for PGD, it is currently not recommended into the clinical practice. However, it may be useful in families showing a peculiar segregation pattern associated with a known mutation in SMCHD1. Recently, new molecular technologies showed promising results in the detection of FSHD1 molecular signature. MC and optical mapping approaches seem to be the most feasible alternative methods for FSHD molecular diagnosis. In particular, MC allows the direct visualization and cartography of numerous individual DNA molecules at 1-kb resolution. Although MC was originally developed to map genes for positional cloning and to study DNA replication, it could be helpful for the diagnosis of disease involving complex chromosomal rearrangements, including FSHD. In fact, MC can be highly advantageous to directly visualize the 4q35 and 10q26 loci and discriminate among 4qA and other alleles. In addition, this approach may help to provide an accurate sizing of D4Z4 array and, subsequently, correlate it with FSHD-related clinical features. However, the expensiveness, the need for further automation, and quicker analytical times hinder the replacement of Southern blotting with MC for D4Z4 sizing [5]. Concerning nanochannel-based optical mapping approach, it proved to be able to provide an accurate quantification of the number of D4Z4 repeats, discriminate the DNA fragments from 4q35 and 10q26, and quantify post-zygotic mosaicisms in FSHD patients. However, this approach needs to be further validated and refined before to be implemented into the clinical practice [42, 43].

Currently, no effective therapies are available to prevent FSHD. In this perspective, the usefulness of pre-symptomatic genetic test should be discussed during genetic counseling [44]. On this subject, it is usually recommended to perform a neurological evaluation before genetic testing. Sometimes, a psychological evaluation of patients and/or their familial members can be performed before the genetic test, in order to avoid harmful implications of the possible results. Further studies are needed to detect individual trajectories of disease and response to therapies/environmental factors as well as to identify genetic signatures which can be utilized for predictive purposes. Moreover, the lack of a genomic personalized medicine in FSHD often influences testing and medical decisions.

Conclusions

In the last 10 years, the availability of new-generation technologies has been crucial for an extensive and deep investigation of the human genome in health and disease conditions. A new era of genetics therefore originated, moving from stem cells to gene therapy, and gene editing is becoming really promising. In this context, genetics progress was very useful to understand the pathogenic pathways of disease and to develop innovative genome-based tools to provide patients with more effective and safer treatments [45].

In 1992, the cause of FSHD was shown to be the deletion of a region within chromosome 4 that consists of repeated DNA called D4Z4. At that time, many scientists assumed that FSHD originates by a mechanism observed for other genetic disorders: the mutation of a gene within D4Z4 makes it unable to produce the corresponding protein. However, subsequent research proved exactly the opposite: FSHD is not due to the loss of a protein, but to its excessive production. The subsequent step was to understand how D4Z4 was able to regulate protein production in FSHD.

Also, the model of inheritance of FSHD seems to follow the classical Mendelian patterns. Although it should be theoretically inherited by an autosomal dominant mode, it is not so frequent in the clinical practice. In our experience, molecular confirmation for FSHD1 is positive in about 60% of subjects with a clinical suspect of disease. This apparent low detection rate of the molecular analysis is explained by the clinical variability of the phenotype that involves just a hyperCKemia (high levels of blood creatine kinase, CK) with poor clinical signs, in most cases. It is important to remind that CK is a metabolic enzyme that is highly expressed in cardiac and skeletal muscles and is detected in blood in response to muscle injury or dystrophy. However, it cannot be utilized as an accurate diagnostic marker for FSHD, since it is highly variable among man and woman and appears slightly/moderately increased in FSHD [4]. Some people do not show any symptoms during their life despite carrying the genetic defect; others are only minimally affected or show non-muscular problems. What is surprising is that this variability exists also within the same family and involves not only the severity, but also the age of onset of the disease. In addition, we observed that as many as 2% of the general population present 8–10 RU on a permissive 4qA allele without being affected, meaning that there must be something more to explain the clinical picture that we observe in patients with a confirmed FSHD diagnosis. The identification of SMCHD1 and DNMT3B as FSHD2-causative genes and their role as disease modifiers in FSHD1 increased this knowledge, proving that unbalances between genomic context and epigenetic regulation can have a critical impact on disease development. Moreover, a lot of studies suggested that copy number variations (CNVs) are important for human phenotypic diversity and disease susceptibility. Repetitive DNA sequences account for 55% of the human genome and a significant fraction of CNVs [46]. In this context, FSHD is a pathology caused by CNVs of D4Z4 repeats affecting the regulation of the epigenetic signatures in affected patients.

The diagnosis of FSHD1, as well as Huntington’s disease or myotonic dystrophy, is made possible thanks to techniques essentially based on basic molecular biology, including DNA enzymatic digestion, Southern blot, hybridization with radiolabeled probes, methylation assays, sequencing analysis, and DNA cloning, which are subsequently combined with the expertise of the operators in discriminating a positive case from a false negative [39, 47,48,49]. On this subject, the introduction of NGS technology into the clinical practice improved the molecular detection of many genetic disorders, making possible to test several genes at the same time [50]. Unfortunately, NGS is still not suitable for characterizing repetitive traits of DNA. Many spinocerebellar ataxia, myotonic dystrophies, as well as FSHD require specific genetic analyses able to establish the exact size of repetitive DNA in a specific trait. This is the reason why the training of clinical experts, biologists, and laboratory technicians on traditional and outdated methodologies is mandatory even in the era of NGS and high-throughput technologies. The current university courses should therefore not only provide the students with a full knowledge of the most modern technologies, but also include specific training on traditional molecular techniques which are still necessary to provide the proper molecular diagnosis to many human diseases, including FSHD.