In this paper we have compared the expression profiles of FSHD-1 and FSHD-2 precursor cells in regard to healthy controls before and after myogenic differentiation. In our knowledge, this is the first report that uses human 4q-linked and non 4q-linked (or phenotypic) FSHD primary myoblasts and their in vitro differentiation to investigate global gene deregulation characterizing cells deriving from FSHD patients with a different genetic defect, but with a very similar phenotypic manifestation of the disease 
. Although the in vitro differentiation of myoblasts does not involve many of the complex series of events known to be important in vivo, such as activation of quiescent satellite cells (stem cell), maturation of the myotubes into muscle fibers and the innervations of the fibers, the cell system we used could be useful in the attempt to derive global gene expression deregulation characterizing the early stages of myogenic differentiation without interferences represented by cell contamination, inflammation or muscle regeneration as found in studies of biopsies.
It is noteworthy that only two FSHD-2 cell lines were available for the chip analysis; although this represents a small sample size we decided anyhow to include them in our analysis since this type of FSHD cell has never been analyzed; thus to render the data more significant we decided to use a lower p value (>0.001) than that used for FSHD-1 (>0.01).
A gradient of altered gene expression throughout the 4q35 chromosome linked to D4Z4 contraction has been proposed as a model for the molecular pathogenesis of FSHD-1 
. Our results did not evidence such a correlation in both FSHD-1 and FSHD-2 cells. Only three genes (SNX25, ANKRD37 and SORBS2) located approximately from 4 to 5 Mb proximal to the D4Z4 array showed in FSHD-1 myotubes a significant down-regulation. Interestingly, one of these genes (ANKRD37) was also found deregulated in muscle biopsies from FSHD-1 patients 
. Absence of significant gene expression alteration throughout the 4q35 region agrees with the data previously reported by Winokour et al. (2003) 
and Osborne et al. (2007) 
on muscle biopsies, thus excluding a position effect model for FSHD. However, we can not exclude the possibility that some of the 4q35 genes (i.e. FRG1) might be transiently deregulated during intermediate steps of the differentiation process 
However, significant results concerning the altered biological processes of the pathological cells were obtained, by deriving in regard to controls the global deregulation of gene expression in FSHD-1 and FSHD-2 myoblasts and myotubes and by comparing the two pathological differentiation processes to the normal one. By combining the two approaches, we derived that gene deregulation was essentially a feature of FSHD-1 proliferating cells and of FSHD-2 differentiated cells. FSHD-1 myoblasts showed a highly significant gene deregulation linked to cell cycle control essentially affecting G1/S and G2/M transitions. These results are in agreement with previous data derived by the analysis of FSHD-1 cells, and showing the up-regulation of p21, known to arrest progression at G1/S interface, and of WEE1, a negative regulator of entry into mitosis (G2/M transition) 
Furthermore, FSHD-1 myoblasts showed the up-regulation of PAX3, a key upstream regulator of the myogenic program: PAX3 up-regulates the myogenic determination gene MYOD1 that, in turn regulates MYOG expression 
. However, while in embryonic tissues the ability of PAX3 to activate the myogenic program is well documented 
, in adult-derived cells this effect is still under discussion.
In our system, the found premature up-regulation of MYOD1 mRNA could be ascribed to PAX3 mRNA up-regulation; furthermore, as previously reported 
, also in our cellular system MYOD-mediated induction of myogenesis is accompanied by the down-regulation of cyclins. Thus, PAX3 up-regulation might contribute to the early cell cycle arrest shown by FSHD-1 myoblasts. In spite of the up-regulation of MYOD1 mRNA, FSHD-1 proliferating cells did not show the occurrence of later marker of myogenic differentiation, such as myogenin and sarcomeric myosin. This could probably due to the absence of other required transcription factors such as myogenic enhancer factors (MEFs), essential for muscle differentiation. Thus, FSHD-1 cells seem to be characterized by a premature and partial activation of the myogenic program that could be related to the observed defect in cell cycle progression.
Remarkably, other two genes SUV39H1 and HMGB2 both involved in chromatin remodeling were down-regulated in FSHD-1 myoblasts. SUV39H1 is a histone methyl-transferase involved in D4Z4 H3K9me3 
, whereas HMGB2 is part of a multi-protein complex shown to bind a 27bp binding element (DBE) within D4Z4 units 
. In normal cells both gene activities, in association with other factors, may play an important role in the establishment and maintenance of the higher order chromatin structure of the D4Z4 array (facultative heterochromatin). In FSHD-1 cells, the down-regulation of SUV39H1 and HMGB2 genes could correlate with the hypothesized more open chromatin conformation of the contracted 4q alleles 
Thus, in addition to the early partial activation of the myogenic program, proteins involved in chromatin organization are also modulated in FSHD-1 samples, suggesting that their absence might contribute to the epigenetic defect of the D4Z4 array.
Conversely, FSHD-2 cells were characterized by a significant alteration of gene expression only after the in vitro
transition from myoblasts to myotubes. Effectively, FSHD-2 myotubes showed the deregulation of genes essentially involved in non-coding RNA metabolism and in nucleolus organization, implied in protein synthesis. FSHD-2 myotubes also showed mitochondrial abnormalities, including energy production, response to oxidative stress and mitochondrial dynamics. Mitochondrial abnormalities and dysfunction in protein synthesis have been also reported for other muscular dystrophies 
FSHD-2 differentiation analysis also evidenced the deregulation of the cell cycle and of proteasomal ubiquitin-dependent process. Importantly, ubiquitin-dependent proteolysis has been suggested to govern terminal muscle differentiation by coordinating cellular division and differentiation 
Interestingly, both FSHD-1 and FSHD-2 cells were affected in sterol biosynthetic process, showing the deregulation, although in the opposite direction, of the same genes. The alteration of cholesterol homeostasis could primarily cause cell damage in membranes lipid rafts, where different proteins are incorporate (e.g. GPI-anchored and cholesterol-linked proteins), and in caveolae a subclass of rafts 
. It was previously reported that caveolae structure alteration could affect myotube formation 
, and that FSHD-1 biopsies are characterized by the impairment of biological processes involved in the synthesis of GPI anchored proteins 
In normal cells reactive oxygen species (ROS) generation is counterbalanced by the action of antioxidant enzymes, such as mitochondrial superoxide dismutase (SOD2) and of those involved in glutathione metabolism. The found deregulation of SOD2 in FSHD-1 myoblasts and of glutathione reductase (GSR) and peroxidase (GPX4) in FSHD-2 myotubes could suggest for both FSHD manifestations the occurrence of a similar increased susceptibility to oxidative stress. The deregulation of enzymes involved in oxidative stress resistance and the consequent increased susceptibility to oxidative stress have been already reported for FSHD-1 myoblasts and biopsies 
Finally, both FSHD-1 and FSHD-2 cells showed the involvement in the gene deregulation network of some microRNAs (miRNA), a class of molecules previously shown to play an important role in the regulation of muscle development 
. Among a total of five miRNAs found deregulated in the present work, two (mir186 and mir15a) were previously reported to be commonly deregulated in more than three (including FSHD) types of muscular disorders 
. The remaining three miRNAs were detected in FSHD-1 myoblasts (mir-23b) and in FSHD-2 myotubes (mir-149, mir-26a2 and mir-23b). Interestingly, one predicted target of the supposedly FSHD-specific miRNA 23b is a gene involved in the chromatin conformation of the 4q D4Z4 array (HMGB2 down-regulated in FSHD-1 myoblasts) 
Although future work is certainly needed to confirm the herein derived observations, taken together our results seem to recapitulate previously reported defects of FSHD-1, and to add new insights into the gene deregulation characterizing both FSHD-1 and FSHD-2. In general, FSHD-1 cells showed an alteration of cell cycle control, a defect in cholesterol homeostasis and presumably in the mitochondrial capacity to buffer oxidative stress. With the exception of cholesterol homeostasis, FSHD-2 cells shared all these features by deregulating different genes. FSHD-2 cells also showed a general deregulation of protein synthesis and degradation. In this regard, proteasome ubiquitin-dependent protein degradation could be viewed as an impairment in exit from the cell cycle. Thus both FSHD manifestations presented cellular deficiencies that do not arise from a 4q position effect mechanism, but rather from a general alteration of gene expression in which miRNA deregulation may play a role.