Familial thrombocytopenia due to a complex structural variant resulting in a WAC-ANKRD26 fusion transcript

Abstract

Advances in genome sequencing have resulted in the identification of the causes for numerous rare diseases. However, many cases remain unsolved with standard molecular analyses. We describe a family presenting with a phenotype resembling inherited thrombocytopenia 2 (THC2). THC2 is generally caused by single nucleotide variants that prevent silencing of ANKRD26 expression during hematopoietic differentiation. Short-read whole-exome and genome sequencing approaches were unable to identify a causal variant in this family. Using long-read whole-genome sequencing, a large complex structural variant involving a paired-duplication inversion was identified. Through functional studies, we show that this structural variant results in a pathogenic gain-of-function WAC-ANKRD26 fusion transcript. Our findings illustrate how complex structural variants that may be missed by conventional genome sequencing approaches can cause human disease.

Figure 1.

A large paired-duplication inversion SV identified as the cause of inherited thrombocytopenia. (A) Pedigree of a three-generation family affected with inherited thrombocytopenia. (B) Representative image of a peripheral blood smear from individual III-1, demonstrating severe thrombocytopenia with few platelets exhibiting normal granularity (red arrows). Scale bar: 10 µm. (C) Schematic of the genetic workup performed to identify the causal variant in this family, highlighting conventional clinical analyses (blue) versus additional approaches (red) needed to resolve this case. (D) The wild-type locus versus a local assembly of the paired-duplication inversion highlighting duplicated regions (greens), inverted segments (angular brackets), and microhomology regions (red and blue arrow heads) and with genes along each assembly (pink arrows denoting genes that are affected by the SV). ICE, Inosine Chemical Erasing Sequencing.

Figure S1.

PCR strategy to identify the causal SV. (A) PCR primers were designed across breakpoint 1 such that the only possible on-target amplification occurs in affected individuals. (B) DNA sequence used in PCR primers is shown. dup-inv-dup, duplication-inversion-duplication.

Figure 2.

The identified SV causes a partial transcript fusion of WAC and ANKRD26 genes, leading to overexpression of a functional ANKRD26 fragment. (A) Illustration of previously described ANKRD26 promoter region variants, which lead to de-repression of ANKRD26 and are either causal or associated with THC2. (B) Expression (log₂-normalized counts per million [cpm]) of the ANKRD26 and WAC genes throughout hematopoietic lineages highlights the robust ubiquitous expression of the WAC versus the regulated and lower expression of ANKRD26. (C) Sashimi plot of RNA sequencing transcriptomes of pooled controls versus affected individuals, highlighting the existence of a WAC-ANKRD26 fusion transcript that is only present in affected individuals (top; red arc). (D) Volcano plot showing that ANKRD26 is one of the most overexpressed genes in affected individuals versus controls, whereas other genes in this locus are unchanged (bottom left; red). (E) Bar plot of ANKRD26 and WAC expression, highlighting the exons involved and not involved in the SV, shows that overexpression is due to the exons that are part of the WAC-ANKRD26 fusion. adj. p-val, adjusted P value; aff., affected; FC, fold change; unaff., unaffected.

Figure S2.

ANKRD26 duplication-inversion-duplication, not promoter variants, cause ANKRD26 overexpression, leading to a suppression of platelet signatures and activation of signatures indicative of monocytes. (A) Integrated Genomics Viewer was used to show that affected individuals harbor no variants in the ANKRD26 promotor, which might explain the thrombocytopenia phenotype. (B) RidgePlot of gene expression profiles of the log₂ fold change (FC) distributions of core genes in the top 20 most significantly enriched gene sets for affected individuals versus unaffected controls. Gene sets were limited to the biological process ontology. Distributions are colored by gene set enrichment Benjamini-Hochberg­–adjusted P values (p.adjust). (C) Core genes for three platelet-associated biological processes show consistent downregulation in affected individuals versus unaffected controls. (D) Core genes for the myeloid leukocyte biological processes show consistent upregulation in affected individuals versus unaffected controls.

Figure 3.

A truncated region of ANKRD26 retains ANKRD26 functions. (A) Gene models showing full-length ANKRD26, the in-frame WAC-ANKRD26 fusion, and the 5′-truncated ANKRD26 starting from Met453. (B) Protein models of the stably expressed full-length and N-terminally truncated ANKRD26. (C) Exogenous expression of full-length ANKRD26, exon 11+ cDNA, but not WAC-ANKRD26 fusion transcript can be detected, suggesting that the fusion transcript does not result in a stable protein. Protein expression levels of ANKRD26 by Western blot (WB) from whole-cell lysates of HEK 293T cells shown by probing for N- and C-terminal ANKRD26 epitopes. (D) RT-PCR analysis of ANKRD26 mRNA from human HSPC sorted for GFP at 96 h after transduction with full-length ANKRD26 and exon 11+ cDNA confirms overexpression of ANKRD26 by 14.8 ± 0.88 and 14.79 ± 1.77, respectively. Data were normalized to the ACTB transcript level and represent mean ± SD of triplicates. (E) A representative replicate showing ERK phosphorylation assessed by flow cytometry in primary human HSPCs 96 h after lentiviral transduction with the empty control vector, full-length ANKRD26, or ANKRD26 exon 11+. (F) A bar plot showing quantification of three biological replicates of the percentage of pERK⁺ of GFP⁺ (successfully infected) cells (mean ± SD of triplicates shown). (G) A bar plot showing quantification of three biological replicates of the pERK mean fluorescence intensity (MFI) among GFP⁺ cells (mean ± SD of triplicates shown). (H) Representative histograms illustrating MFI of pERK within the GFP⁺ gate are shown for serum-starved HSPCs that were stimulated with TPO. A significant increase was noted for both full-length ANKRD26 and ANKRD26 exon 11+. P values were calculated by one-way ANOVA test followed by Dunnet's test (***, P < 0.001).