Integrated proteogenomic analysis for inherited bone marrow failure syndrome

Abstract

Recent advances in in-depth data-independent acquisition proteomic analysis have enabled comprehensive quantitative analysis of >10,000 proteins. Herein, an integrated proteogenomic analysis for inherited bone marrow failure syndrome (IBMFS) was performed to reveal their biological features and to develop a proteomic-based diagnostic assay in the discovery cohort; dyskeratosis congenita (n = 12), Fanconi anemia (n = 11), Diamond-Blackfan anemia (DBA, n = 9), Shwachman-Diamond syndrome (SDS, n = 6), ADH5/ALDH2 deficiency (n = 4), and other IBMFS (n = 18). Unsupervised proteomic clustering identified eight independent clusters (C1-C8), with the ribosomal pathway specifically downregulated in C1 and C2, enriched for DBA and SDS, respectively. Six patients with SDS had significantly decreased SBDS protein expression, with two of these not diagnosed by DNA sequencing alone. Four patients with ADH5/ALDH2 deficiency showed significantly reduced ADH5 protein expression. To perform a large-scale rapid IBMFS screening, targeted proteomic analysis was performed on 417 samples from patients with IBMFS-related hematological disorders (n = 390) and healthy controls (n = 27). SBDS and ADH5 protein expressions were significantly reduced in SDS and ADH5/ALDH2 deficiency, respectively. The clinical application of this first integrated proteogenomic analysis would be useful for the diagnosis and screening of IBMFS, where appropriate clinical screening tests are lacking.

Fig. 1. Diagnostic flowchart based on the proteogenomic analysis.

Of the 10 patients with monoallelic SBDS variant detected by DNA sequencing (DNA-seq), four had diagnostic changes after a non-targeted proteomic analysis. Two patients with a significantly reduced SBDS protein expression were finally diagnosed with SDS, one with DC, and one with FA. ^*The two patients with SDS only carried the monoallelic SBDS variant detected by DNA-seq. However, proteomic analysis revealed a reduced SBDS protein expression, and subsequent comprehensive RNA-seq read alignment evaluation confirmed the presence of biallelic SBDS variants. ADH5 alcohol dehydrogenase 5, ALDH2 aldehyde dehydrogenase 2, DBA Diamond–Blackfan anemia, DC dyskeratosis congenita, FA Fanconi anemia, IBMFS inherited bone marrow failure syndrome, NOS not otherwise specified, SDS Shwachman–Diamond syndrome.

Fig. 2. Proteomic and genomic profiling of the discovery cohort.

Patient characteristics of the IBMFS discovery cohort. An omics analysis of target-captured or whole-exome DNA, in-depth proteomic, and transcriptomic analyses was conducted in patients with IBMFS (N = 60) as the discovery cohort. Proteomic analysis identified six patients with significantly reduced SBDS protein expression; however, two of them were undiagnosed with DNA-seq analysis only. The ADH5 protein expression was significantly reduced in ADH5/ALDH2 deficiency. ADH5 alcohol dehydrogenase 5, ALDH2 aldehyde dehydrogenase 2, AML acute myeloid leukemia, ANC absolute neutrophil counts, DBA Diamond–Blackfan anemia, DC dyskeratosis congenita, FA Fanconi anemia, Hb hemoglobin, IBMFS inherited bone marrow failure syndrome, MDS myelodysplastic syndrome, NOS not otherwise specified, PBMC peripheral blood mononuclear cells, Plt platelet count, Ret reticulocyte count, SD standard deviation, SDS Shwachman–Diamond syndrome, WBC white blood cell.

Fig. 3. Unsupervised proteomic-based cluster for IBMFS subtypes.

Proteomic cluster and differentially expressed proteins. Each row represents a proteomic cluster, and each column represents a protein. Red/blue indicates up/down expression patterns of different proteins within clusters. The pie charts indicate the percentage of the pathological progression to MDS and IBMFS subtypes in each of eight proteomic-based clusters (C1–C8). The gene members of the major pathways enriched in the gene cluster (heatmap below) are shown. For each pathway, the mean sample-specific gene set enrichment analysis (ssGSEA) scores in the respective IBMFS, based on proteomic and transcriptomic data, are shown in the right. Correlations between protein and mRNA expression levels are indicated for each proteomic cluster. ADH5 alcohol dehydrogenase 5, ALDH2 aldehyde dehydrogenase 2, DBA Diamond–Blackfan anemia, DC dyskeratosis congenita, FA Fanconi anemia, GSEA gene set enrichment analysis, IBMFS inherited bone marrow failure syndrome, MDS myelodysplastic syndrome, SDS Shwachman–Diamond syndrome.

Fig. 4. Proteogenomic-based diagnostic testing for SDS and ADH5/ALDH2 deficiency.

A SBDS protein expression in each IBMFS group. Four patients carrying biallelic SBDS pathogenic variants had decreased SBDS protein expression. In patients with IBMFS with monoallelic SBDS pathogenic variants (n = 10), two (UPN213 and UPN751) had significantly decreased the SBDS protein expression. No patients in other IBMFS groups had decreased SBDS protein expression. B Differences in the SBDS protein and SBDS mRNA expression based on individual SBDS variants. C ADH5 protein expression in each IBMFS group. D Differences in the ADH5 protein and ADH5 mRNA expression based on individual ADH5 variants. Starburst plot integrating proteomic and mRNA expression analyses between patients with and without SDS (E), and between those with and without ADH5/ALDH2 deficiency (F). ACTB actin beta, ADH5 alcohol dehydrogenase 5, ALDH2 aldehyde dehydrogenase 2, DBA Diamond–Blackfan anemia, DC dyskeratosis congenita, FA Fanconi anemia, IBMFS inherited bone marrow failure syndrome, LCL lymphoblastoid cell line, NOS not otherwise specified, UPN unique patient number, RPM reads per million, SDS Shwachman–Diamond syndrome.

Fig. 5. Potential utility of targeted proteomic analysis using IBMFS-related small panels.

A Summary of the discovery and extension cohort. Non-targeted proteomic analysis was performed on the discovery cohort (n = 60) and healthy controls (HC, n = 14). Targeted proteomic analysis was performed on 417 samples from the discovery cohort (n = 60), the extension cohort (n = 330), and HC (n = 27). Comparison of SBDS (B) and ADH5 (C) protein expression levels measured by non-targeted and targeted proteomic analyses. D, E SBDS and ADH5 protein expression levels measured by targeted proteomic analysis in 417 samples from patients with IBMFS-related hematological disorders (n = 390) and HC (n = 27). In both measurement assays, SBDS (D) and ADH5 (E) protein expressions were significantly decreased in patients with SDS and ADH5/ALDH2 deficiency, respectively (P < 0.001 each). AA aplastic anemia, ADH5 alcohol dehydrogenase 5, ALDH2 aldehyde dehydrogenase 2, BMNC bone marrow mononuclear cell, DBA Diamond–Blackfan anemia, DC dyskeratosis congenita, FA Fanconi anemia, HC Healthy controls, IBMFS inherited bone marrow failure syndrome, MDS/AML myelodysplastic syndrome/acute myeloid leukemia, NOS not otherwise specified, PBMC peripheral blood mononuclear cells, SDS Shwachman–Diamond syndrome, UPN unique patient number, WAS Wiskott–Aldrich syndrome.