Revertant mosaicism in junctional epidermolysis bullosa due to multiple correcting second-site mutations in LAMB3

Abstract

Revertant mosaicism due to in vivo reversion of an inherited mutation has been described in the genetic skin disease epidermolysis bullosa (EB) for the genes KRT14 and COL17A1. Here we demonstrate the presence of multiple second-site mutations, all correcting the germline mutation LAMB3:c.628G-->A;p.E210K, in 2 unrelated non-Herlitz junctional EB patients with revertant mosaicism. Both probands had a severe reduction in laminin-332 expression in their affected skin. Remarkably, the skin on the lower leg of patient 078-01 (c.628G-->A/c.1903C-->T) became progressively clinically healthy, with normal expression of laminin-332 on previously affected skin. In the other proband, 029-01 (c.628G-->A/c.628G-->A), the revertant patches were located at his arms, shoulder, and chest. DNA analysis showed different second-site mutations in revertant keratinocytes of distinct biopsy specimens (c.565-3T-->C, c.596G-->C;p.G199A, c.619A-->C;p.K207Q, c.628+42G-->A, and c.629-1G-->A), implying that there is not a single preferred mechanism for the correction of a specific mutation. Our data offer prospects for EB treatment in particular cases, since revertant mosaicism seems to occur at a higher frequency than expected. This opens the possibility of applying revertant cell therapy in mosaic EB of the LAMB3 gene by using autologous naturally corrected keratinocytes, thereby bypassing the recombinant gene correction phase.

Figure 1. Revertant unaffected skin on the left lower leg of proband 078-01 in January 1999.

The skin of the left lower leg reverted to clinically healthy skin after an erosive period of 7 years (A, medial aspect; B, lateral aspect), while the skin of the right lower leg remained affected. The area of revertant skin is outlined in black.

Figure 2. IF microscopy reveals cellular mosaicism in the skin of proband 078-01.

(A and B) Staining with monoclonal antibody K140 to the β3 chain of LM-332 was markedly reduced in affected skin of the upper arm (A) compared with normal control skin (B). (C) To our surprise, a short stretch of approximately 25 basal cells embedded in an LM-332–reduced environment that stained brightly for LM-332 was observed in biopsy II (M). (D) Biopsy III (R) of the left lower leg showed an interrupted pattern, with normal and reduced staining along the epidermal-dermal junction, while biopsy IV (R) revealed bright staining for all basal cells (data not shown). Original magnification, ×40.

Figure 3. Normal hemidesmosomes in revertant skin.

(A) Ultrastructural examination of the affected skin of proband 078-01 showed a reduced number of abnormal hemidesmosomes. The intermediate filaments were not connected to the flattened basilar cell periphery. (B) The revertant skin of the lower leg reveals intermediate filaments that connect to normal hemidesmosomes in the basilar cell periphery. Black arrows indicate hemidesmosomes and white arrows intermediate filaments. Scale bars: 500 nm.

Figure 4. Identification of the different correcting LAMB3 mutations in patient 078-01.

(A) The G→A nucleotide change at position –1 of the 5ι splice site of intron 7 was present in keratinocytes with reduced staining of LM-332. (B) The second-site mutation c.596G→C was present in revertant keratinocytes of biopsy III (R). (C) An additional mutation in intron 7, c.628+42G→A, in revertant keratinocytes of biopsy IV (R). The cryptic splice site, CAGïΣΦGT, which is used when the c.628+42G→A substitution is present, is indicated by the dashed line. Red arrows indicate the inherited mutation, and green arrows the second-site mutations. Corresponding amino acid sequences are indicated above the nucleotide sequences.

Figure 5. Effects of second-site mutations at the mRNA level (patient 078-01).

(A) nt 580–848 of LAMB3 mRNA was analyzed by RT-PCR. Along with the expected 269-bp product in the control sample (lane 1), affected skin samples revealed 3 smaller transcripts (lanes 2 and 3). Isolation and sequencing showed the presence of normally spliced transcript (B, b), the exon 7–deleted transcript (c), a transcript missing the first 2 nt of exon 8 in addition to exon 7 (d), and a transcript with a deletion of 165 nt, comprising exon 7 and the first 101 nt of exon 8 (e), in biopsy I (M) (lane 2) and biopsy II (M) (lane 3). In the mosaic biopsy III (R) (lane 4), the normally spliced variant was more abundant, whereas the amount of aberrant mRNA transcripts was decreased. In the completely reverted biopsy IV (R), an additional transcript retaining the first 66 nt of intron 7 (B, a) was present (lane 5). (C) The LAMB3 gene, with sizes of exons (above) and introns (below). Splicing of normal full-length transcript is indicated with the solid line (C, b). Dotted lines depict the splicing of aberrant transcripts (C, a, c, d, and e). The mutations c.628G→A and c.628G+42G→A are indicated by black arrowheads. (D) p.R635X induces exon 14 skipping. Primers amplifying bp 1,641–2,229 demonstrated the 589-nt transcript in the control (lane 1). This fragment was almost absent in all patient biopsies (lanes 2–5), while a smaller amplimer of 210 bp was visualized. Lane M contains a 100-bp molecular size marker (A and D).

Figure 6. Revertant unaffected skin on upper arm and shoulder of non-Herlitz JEB patient 029-01 in January 2006.

(A) The region where biopsy IV (R) was taken is circled, whereas other skin areas without blistering tendency are indicated with crosses. The latter are clearly distinguishable from the surrounding erythematous atrophic skin. (B) Frontal aspect of the chest. The sites of biopsy II (R) and III (R) are labeled.

Figure 7. The inherited homozygous c.628G→A mutation was present in 029-01 proband’s keratinocytes with normal (A–C) as well as reduced (D) LM-332 staining.

(A) In biopsy I (R), a second-site mutation, c.619A→C;p.K207Q, was present in exon 7. Biopsy III (R) had an additional substitution, c.565-3T→C, in the 3ι splice site of intron 6 (B) and biopsy IV (R) an additional c.629-1G→A change in the 3ι splice site of intron 7 (C). (D) None of these additional substitutions were seen in LDM-isolated mutant keratinocytes. Red arrows indicate the inherited mutation, and green arrows the second-site mutations. Amino acid sequences are indicated above the nt sequences.

Figure 8. Effects of second-site mutations at the mRNA level (patient 029-01).

Oligonucleotide primers were similar to those described in Figure 5. A normal human control sample shows the expected 269-bp product (lane 6). The affected skin sample contained 3 additional smaller fragments due to the c.628G→A transition (lanes 1–4 and Figure 5B, c–e). The revertant keratinocytes of biopsy I (R) with the secondary c.619A→C mutation (lane 1) produced more full-length mRNA transcript and fewer 64 and 66 bp–deleted transcripts than mutant keratinocytes (lane 5). This is also seen in the revertant keratinocytes from biopsy II (R) (lane 2) and biopsy III (R) (lane 3) containing the additional c.565-3T→C mutation. In contrast, the revertant keratinocytes of biopsy IV (R) with the second-site c.629-1G→A mutation had a greater abundance of the transcript with the 66-bp deletion (lane 4). Lane M contains a 100-bp molecular size marker and lane 7 a negative control.

Figure 9. Schematic drawing showing the different second-site mutations that all corrected LAMB3:c.628G→A.

(A) The inherited germline mutation c.628G→A is depicted as black rectangles, while the second-site mutations are depicted as white rectangles. Cells with a mutant phenotype are white, and those with a revertant phenotype green. (B) Distribution of the mutations in the LAMB3 gene. The red arrow indicates the inherited mutation, and green arrows indicate the second-site mutations.