Figure 1. Restriction map of the 4q35 region: relative positions of FRG1, TUB4Q, D4F104S1 (p13E-11) and D4Z4. (a) EcoRI fragment detected by probe p13E-11 (b) Predominantly comprises an array of 3.3 kb tandem repeats that have a copy number of 12–100 in normal controls and usually 10 or less in FSHD patients. CEN: centromere; TEL: telomere. (c) Each 3.3 kb repeat comprises two homeodomains (shaded boxes) encompassing an ORF with an in-frame start codon (ATG) and a stop codon. It encodes the DUX4 gene. The position of the GC and TACAA boxes in the promoter-like sequence of DUX4 gene is indicated.

Figure 2. Location of D4Z4 repeats at 4q35 and 10q26. The 10q-derived D4Z4 repeats differ from the 4q-derived repeats in having an internal restriction site for BlnI. Deletion of D4Z4 repeats at 4q35 is associated with FSHD, whereas reduction in the copy number of 10q-derived D4Z4 repeats does not result in any specific phenotype.

Figure 3. Length of the smallest EcoRI fragment associated with the disease in unrelated FSHD individuals. 10 mg DNA was digested with EcoRI, fractionated on a 0.5% agarose gel for 48 h at 0.5 V cm^–1 and Southern blotted on to Hybond N (Amersham). The DNA on the membrane was hybridized with radio-labelled DNA probe p13E-11. Lane 1 contains high molecular weight markers. Lanes 2–12 contain DNA samples from unrelated FSHD patients. Lane 5 contains DNA from an FSHD patient in whom the smallest EcoRI fragment is >48 kb. The corresponding smallest EcoRI fragment size is written at the top of each lane. The 8.9 kb fragment represents a Y-specific sequence.

Figure 4. Correlation of age of onset in new mutation cases and probands of 4q35-linked FSHD families, with fragment size at locus D4F104S1.

Figure 5. Subtelomeric sequence exchange between 4q35 and 10q26 in 20% of the normal population. In the control population, 80% of individuals carry a standard configuration, with 4-type repeats on chromosome 4, and 10-type repeats on chromosome 10. In 10% of individuals, 4-derived repeats are also present on one of their chromosomes 10. Likewise, 10% of the control population carry 10-derived repeats on one of their chromosomes 4. (a) 5 g high molecular weight DNA digested sequentially with EcoRI and EcoRI/BlnI, and hybridized with probe p13E-11. In an informative situation, four different-sized EcoRI fragments are produced following a single digest, two derived from chromosome 4 and two from 10q. Digestion with enzyme BlnI will cleave two chromosome 10-specific fragments. Chromosome 4-specific fragments will be reduced by 3 kb owing to the presence of a BlnI site proximal to the first repeat but distal to the EcoRI site (d). However, in 10% of individuals (b), 4-type repeats (BlnI resistant) are translocated to chromosome 10 and therefore with EcoRI/BlnI digestion, three alleles are seen instead of the expected two fragments. Similarly, 10% of control individuals (c) carry BlnI-sensitive repeats on one of their chromosomes 4; therefore, with EcoRI/BlnI double digest, one allele (monosomy) is observed (adapted from van Deutekom et al., 1993).

Figure 6. Mosaicism in FSHD patients. Subtelomeric exchange of the repeat sequences on chromosomes 4 and 10 in the control and FSHD population. Chromosomes 4 are light grey whereas chromosomes 10 are dark grey. Top: In the control population, 80% individuals have a 4-type repeat on chromosome 4 and a 10-type repeat on chromosome 10 (adapted from van der Maarel et al., 2000). Bottom: The repeat-array constitutions of mosaic individuals from de novo FSHD families (adapted from van der Maarel et al., 2000). The deletion is indicated by an open bar. These individuals carry two cell populations indicated within a box. In the original population, no FSHD-associated rearrangement is present, whereas, in the other population, a deletion has occurred on chromosome 4. van der Maarel et al. (2000) identified mosaicism in de novo FSHD families by the detection of a fifth restriction fragment (either EcoRI/HindIII or EcoRI/BlnI) by pulsed-field gel electrophoresis (PFGE) followed by Southern blotting with the probe p13E11. They detected somatic mosaicism in 40% of de novo FSHD families (14% in an unaffected parent and on 26% of the de novo FSHD patients themselves). To determine whether interchromosomal exchanges (either 4-type repeats on chromosome 10 or 10-type repeats on chromosome 4) might play a role in the deletion of repeats on chromosome 4, they examined the repeat array constitutions of 13 mosaic individuals (i.e. those with five restriction fragments). Interestingly, 6/13 (46%) carried 4-type repeats on chromosome 10, whereas 10-type repeats on chromosome 4 were not identified in a single mosaic case (0/13). This frequency of exchange differs from the Dutch control population (van Deutekom et al., 1996) in which 10% of individuals carry 4-type repeats on chromosome 10 and vice versa. In mosaic individuals, 54% carry a standard allele configuration, whereas in 46% of cases, 4-derived alleles are transferred to chromosome 10.

Figure 7. Different proposal models of FSHD pathogenesis. (a) A cis-spreading model (position effect variegation): The 4q35 genomic region containing large arrays of D4Z4 repeats and the immediately adjacent region exhibit many heterochromatic features. It is proposed that large deletions of the D4Z4 repeat array observed in FSHD induce a ‘relaxation’ of heterochromatic region to transcriptional upregulation of genes located immediately proximal to the D4Z4 array. (b) The insulator model: here the D4Z4 array is proposed to act as a spacer between the distal heterochromatic region and the proximal euchromatic sequences at 4q35. Large deletions of the D4Z4 repeat may facilitate the spread of heterochromotanization to these more proximal regions, leading to the transcriptional downregulation of genes in this region. (c) The cis-looping model: in this model it is conjectural that DNA loops between D4Z4 and its target gene(s) in cis occur only when the formation of normal D4Z4 intra-array loops are impaired due to D4Z4 contraction. While in the normal individuals, a larger D4Z4 array prevents any interaction with distant genes in cis. The excessive contractions of the D4Z4 array in FSHD results in appropriate gene expression. (d) The nuclear oganization model; It is known that in the interphase nucleus that 4q telomeres seem to be exclusively located towards the periphery of the nucleus where they may interact with the nuclear membrane. This peripheral nuclear localization of 4qter is apparently lost in cells deficient of lamin A/C, a nuclear membrane protein. According to this model, interaction of 4qter with the nuclear lamina where chromatin and transcription factors are located, is disrupted in FSHD. It is suggested that this nuclear disruption may lead to misbalance of chromatin and transcription factors at 4qter (adapted from Marrel and Frants, 2005).

Figure 8. The FSHD disease enigma; many of the above disease features must be considered when attempting to deduce a pathological mechanism.