Analysis of LMNB1 duplications in autosomal dominant leukodystrophy provides insights into duplication mechanisms and allele-specific expression

Abstract

Autosomal dominant leukodystrophy (ADLD) is an adult onset demyelinating disorder that is caused by duplications of the lamin B1 (LMNB1) gene. However, as only a few cases have been analyzed in detail, the mechanisms underlying LMNB1 duplications are unclear. We report the detailed molecular analysis of the largest collection of ADLD families studied, to date. We have identified the minimal duplicated region necessary for the disease, defined all the duplication junctions at the nucleotide level and identified the first inverted LMNB1 duplication. We have demonstrated that the duplications are not recurrent; patients with identical duplications share the same haplotype, likely inherited from a common founder and that the duplications originated from intrachromosomal events. The duplication junction sequences indicated that nonhomologous end joining or replication-based mechanisms such fork stalling and template switching or microhomology-mediated break induced repair are likely to be involved. LMNB1 expression was increased in patients' fibroblasts both at mRNA and protein levels and the three LMNB1 alleles in ADLD patients show equal expression, suggesting that regulatory regions are maintained within the rearranged segment. These results have allowed us to elucidate duplication mechanisms and provide insights into allele-specific LMNB1 expression levels.

Keywords: ADLD; FoSTeS; Lamin B1; MMBIR; NHEJ; duplication Alu; leukodystrophy.

Figure 1

Overview of the genomic rearrangements in the ADLD families. A: Modified output from the UCSC genome browser showing the LMNB 1 gene duplications in 20 ADLD families (16 unique duplications) and their surrounding genomic region. The duplications are marked in blue, with the exceptions of the BR1 duplication/inversion, which is in yellow and the triplicated segment, which is in green. Duplications marked with asterisks (*) have sequence insertions at their duplication junctions and show a clustering of their centromeric breakpoints within a 25 kb segment. The minimal critical region duplicated in ADLD of ∼75 kb is also shown. The location of SINE repetitive elements and microsatellite markers used in genotyping (modified UCSC genome browser tracks) are shown below. Note the enrichment of SINE elements (the majority of which are Alu repetitive elements) centromeric to the LMNB 1 gene. B: Schematic representation of the three LMNB1 duplication configurations identified. C1–T1 and C2–T2 represent the duplicated segments that are derived from the parental genomic region, C–T. Black arrows represent orientation of primers used to for PCR and sequencing across duplication and triplication junctions.

Figure 2

LMNB1 duplication junction sequences. Nucleotide positions from chromosome 5 (GCRh37/hg19) are indicated on the left of each junction. In each case, the reference sequence corresponding to the telomeric end of the duplication (red), the junction fragment present in LMNB1 duplication carriers (red and blue) and the reference sequence corresponding to the centromeric end the duplication (blue) are shown. The grey highlighted sequences represent either the presence of microhomology or nucleotide insertions at the duplication junctions. In sample A3, a single base pair deletion and an adjacent mismatch compared with the reference sequence were present. Repetitive elements present at the duplication junctions are also displayed. At the triplication junction (A6, A7, K2–3) the dotted line represents the extended part of the 146 bp segments that shows perfect homology.

Figure 3

Architecture of the BR1 inverted duplication. A: Schematic representation of the inverted duplication in BR1. The C1 and T1 junctions represent the extents of the duplication. The duplicated segment C2–T2 (brown) is inverted and embedded between junctions I1 and I2 (red vertical lines). Analysis of the junction sequences reveals that the I1–T2 junction is complex with a 78 bp J1–J2 segment (green) interspersed within it. The J1–J2 and the T2–C2 segments are in the reverse orientation. The red and green circles mark the location of the BAC probes used for FISH. Sequence alignments of the I1–J1, J2–T2, and C2–I2 junctions (center) are shown with their respective reference sequences (above and below). In the sequence alignments the regions of microhomology are marked in black. The I1–J2 sequences fall within adjacent AluY repeats which are in an opposite orientations (arrows). B: Overview of the genomic region containing breakpoints I1–I2 (red vertical lines) and J1–J2 (green vertical lines) on the reference genome. Arrows mark the orientation of the Alu elements. The array CGH plot below shows the location of a nonduplicated segment (solid red horizontal bar) surrounded by a duplicated region in the BR1 sample. The y-axis represents relative probe intensity values on a Log₂ scale. C: FISH analysis using the fluorescent labeled BAC probes RP11–692P23 (red) and RP11–772E11 (green). The red arrow points to the chromosome with the duplicated allele, whereas the white arrow shows the chromosome with the normal allele. The presence of a red–green–green–red pattern confirms the presence of the inverted duplication. The normal chromosome shows a red–green pattern. D: Model showing the replication fork switching that could give rise to the BR1 duplication. Relative locations of the duplication junctions are marked together with the Alu repetitive elements and genes involved in the rearrangement. Arrowheads show direction of DNA relative to the positive strand. Circled numbers represent FoSTeS events. Colored circles represent the duplicated genes.

Figure 4

Bioinformatics analysis of duplication breakpoints and surrounding genomic regions. A: Analysis of repetitive elements at duplication junctions. Light gray columns show all repetitive elements and dark gray columns show Alu repetitive elements only. An enrichment of Alu repetitive elements in centromeric breakpoints was present. B: Analysis of repetitive elements in 200 bp sequences surrounding duplication breakpoints in patients versus simulated breakpoints of control sequences. Only centromeric sequences show an enrichment of Alu repetitive elements. C: GC% in 4 kb sequences surrounding duplication breakpoints in patients versus simulated breakpoints in controls. All breakpoint sequences show significantly higher GC% than control sequences. D: Enrichment of CTG/CAG motifs in duplication breakpoint sequences. CTG/CAG motifs were found to be significantly enriched in telomeric breakpoint sequences. In all panels asterisks (*) represents a statistical significance of P < 0.05, and double asterisks (**) represents P < 0.001. E: Consensus sequence motif at centromeric duplication breakpoints. F: Consensus sequence motif at telomeric duplication breakpoints. In both (E) and (F), x-axis represents position of the nucleotide in the motif and the height of the nucleotide represents the probability of observing that particular nucleotide at that position. Both motifs were found to significantly overrepresented in the respective patient breakpoint sequences when compared with control sequences (P < 10⁻⁶).

Figure 5

Lamin B1 expression analysis. A: Calibration of the SNaPshot experiment using known concentrations of two plasmids containing the C and T allele of SNP #rs1051644. Percentages indicate the C:T ratio. A reproducible correlation between expected (x-axis) and measured (y-axis) values were obtained. On the right, electropherograms at different relative concentrations. B: Scheme of the wild-type heterozygous SNP rs#1051644 and of the two possible duplication configurations. The table on the bottom shows the results of the SNaPshot experiments whose graphic is in panel C (values = mean ± standard error). C: SNaPshot results of the rs#1051644 analysis on genomic DNA (gDNA) and cDNA derived from fibroblasts of controls (ctrls) and patients (ADLD), showing the C:T ratio (y-axis). Controls are shown as black-filled circles (gDNA) and empty circles (cDNA), patients are shown as black-filled squares (gDNA) and empty squares (cDNA). Heterozygous controls cluster around 50%, whereas duplication carriers cluster around 65% or 35% depending on which of the two alleles is duplicated (***P < 0.001; **P < 0.01). D: Real-time experiments measuring LMNB1 cDNA levels compared with the reference gene HMBS. Patients showed a statistically significant increase compared with controls both on mRNA derived from fibroblasts and from blood (**P < 0.01). E: Western blot analysis shows increased LMNB1 expression in patients compared with control samples (samples were normalized using the MemCode system); full Western blot images and MemCode staining are available in Supp. Fig. S1). On the right, the OD quantification of LMNB1 compared with MemCode staining. In all patients, LMNB1 protein levels were significantly increased compared with controls (**P < 0.01; *P < 0.05).