Whole exome sequencing of 20 patients with pulmonary hypertension caused by fibrosing mediastinitis

To the Editor: Fibrous mediastinitis (FM) is a rare group of diseases characterized by excessive proliferation of fibrous tissue in the mediastinum, which can cause bronchial stenosis, superior vena cava (SVC) obstruction, and pulmonary artery (PA) and vein (PV) stenosis. Moreover, long-term obstruction of PAs and PVs may lead to pulmonary hypertension (PH), right heart failure (RHF), and death.[1] Although it is generally held that FM is an aberrant immunoproliferative response of the mediastinum to a triggering factor, such as infection of Mycobacterium tuberculosis and Histoplasma capsulatum, the exact etiology and pathogenesis of FM remain unclear; in particular, the genetic susceptibility of this fatal disease is unknown. In this study, we performed whole exome sequencing (WES) on blood samples from 20 patients with PH-FM and sought to identify exome variations, particularly the single nucleotide polymorphism (SNP) and insertion-deletion (InDel) variations underpinning this rare but fatal disease and explore the possible roles of these genomic variations in disease initiation and progression. Ethical approval was obtained from the Ethics Committee of Gansu Provincial Hospital, and written informed consent was obtained from all participants prior to their enrollment in the study. A total of 20 patients with FM (9 males and 11 females, mean age 66.0 ± 7.6 years) were included in this study. The clinical, functional, and hemodynamic data of the patients are summarized in Supplementary Table 1, https://links.lww.com/CM9/C692. FM diagnosis was determined by contrast-enhanced chest computed topography (CT). Patients with peak tricuspid regurgitant velocities >2.8 m/s in echocardiography were suspected of PH.[1] The diagnosis of PH was confirmed by right heart catheterization with a mean pulmonary artery pressure (mPAP) ≥20 mmHg.[1] The demographic information of these patients was sorted and analyzed, and peripheral blood samples were collected. Genomic DNA was extracted from the blood samples, and WES was performed by Beijing Beirui Hekang Biotechnology Company (Beijing, China). The genomic DNA was randomly sheared into fragments 180–250 bp in length. High-throughput and in-depth sequencing was performed in an HiSeq 2500 System (Illumina Inc., San Diego, CA, USA) after enriching all exon regions of the human genome with the Agilent-V6 or IDT-39 M liquid chip capture system. Filtering of low-frequency deleterious variants was performed as follows: (1) Variants were filtered against the 1000 Genomes Project (1000g2015aug_all, 1000g2015aug_eas), ESP6500SI-V2 (esp6500siv2_all), and Exome Aggregation Consortium (ExAC; ExAC_ALL, ExAC_EAS) databases, and variants with a minor allele frequency (MAF) <0.05 were retained; (2) the synonymous single nucleotide variants (SNVs) that were not relevant to the amino acid alternation were discarded; (3) the variants that are unannotated with genes were discarded; (4) only SNVs occurring in exons or splice sites were further analyzed since we were interested in amino acid changes; and (5) Bioinformatics mutation prediction software programs were used, including SIFT 6.2.1 (A*STAR Bioinformatics Institute, Singapore), Align-GVGD 2.0.2 (Huntsman Cancer Institute, University of Utah, Salt Lake City, UT, USA), MutationTaster2021 (Charité–Universitätsmedizin Berlin, Berlin, Germany), and CADD v1.6 (University of Washington, Seattle, WA, USA). At least half of these four software programs show that the SNV is not benign and can be retained. The mutations were annotated in detail through ANNOVAR software (2011–2021 releases; PA, USA). A Verita Trekker mutation system developed by Beijing Beirui Hekang Biotechnology Company was used to detect the SNV and InDel sites. The data revealed that there were 8104.5 InDels on average in the 20 DNA samples. Compared with those of the normal controls, the proportions of the InDels at different lengths were approximately equal. The SNP analysis showed that the total number of SNPs was 76,257.3 on average. The SNPs can be categorized into SNPs in the coding region and those in the non-coding region. The former could be further divided into synonymous SNPs (without amino acid changes) and non-synonymous SNPs (with amino acid changes). Among the total SNPs, the synonymous and non-synonymous SNPs in the 20 samples accounted for 15.35% (11,688/76,154) and 14.43% (10,915/76,154), respectively. There were 19,400 identical benign sites, 72 identical variants of uncertain significance (VUS), and 380 identical VUS false positive (VUS_FP) sites. The functional regions of both the benign sites and the VUS included synonymous and non-synonymous SNVs, as well as other regions. Non-synonymous SNVs were identified at the highest frequency in the VUS. Moreover, the VUS functional regions harbored 31 intronic, 24 exonic, and 9 intergenic variations. Furthermore, within the 24 exonic VUS, 13 non-synonymous SNVs were distributed in the MUC5AC, HYDIN, ZNF717, MADCAM1, and POTED genes; 6 synonymous SNVs were distributed in the HYDIN, ANKRD36C, ZNF717, and KMT5A genes; and 5 non-frameshift substitutions were distributed in the MUC3A and OR4C3 genes [Supplementary Table 2, https://links.lww.com/CM9/C692]. The Dann scoring system was used to score the harmfulness of the SNVs and InDels (the higher the score, the more harmful), and 22 VUS and 75 VUS_FP were scored higher than 0.7. For all the mutations identified in this study, the gene ontology (GO) analysis showed that the biological processes (BPs) of the mutated genes were enriched in biological adhesion, cell adhesion, and so on; the molecular functions (MFs) of the mutated genes were enriched in metal ion binding, hydrolase activity, etc.; and the cellular components (CCs) of the mutated genes were enriched in inherent components of membrane, cell periphery, plasma membrane, etc. [Supplementary Figure 1A, https://links.lww.com/CM9/C692]. Moreover, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis demonstrated the top 20 signaling pathways enriched by the mutated gene exons, including extracellular matrix (ECM) receptor interaction, phosphoinositide 3-kinase–protein kinase B (PI3K-Akt) signaling pathway, focal adhesion, pathways in cancer, endocytosis, neuroactive ligand receptor interaction, regulation of actin cytoskeleton, etc. [Supplementary Figure 1B, https://links.lww.com/CM9/C692]. The protein-protein intereaction results showed that the top 10 proteins with the highest degree of interaction were tumor protein p53 (TP53), abelson murine leukemia viral oncogene homolog 1 (ABL1), estrogen receptor 1 (ESR1), E1A binding protein p300 (EP300), Jun proto-oncogene, AP-1 transcription factor subunit (JUN), casitas B-lineage lymphoma (CBL), conserved helix-loop-helix ubiquitous kinase (CHUK), growth factor receptor-bound protein 2 (GRB2), tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein gamma, 14-3-3 gamma (YWHAG), and breast cancer 1 (BRCA1) [Supplementary Figure 1C, https://links.lww.com/CM9/C692]. For variant of uncertain significance (VUS) mutations, the GO analysis revealed that the BPs enriched by the genes containing VUS mutations included serine, cysteine, glutathione, and glycoprotein biosynthesis as well as glycosylation and serine metabolism. The MFs enriched by the VUS-harboring genes included activities of multiple enzymes and the structural components of the extracellular matrix. The CCs consisted of plasma membrane components and those of the extracellular matrix [Supplementary Figure 2A, https://links.lww.com/CM9/C692]. Moreover, KEGG analysis revealed that the top signaling pathways with the highest VUS enrichment included the intestinal immune network for immunoglobulin A (IgA) production, lysine degradation, cell adhesion molecules, and olfactory transduction pathways [Supplementary Figure 2B, https://links.lww.com/CM9/C692]. We validated some of the identified VUS in 24 exons using conventional Sanger sequencing. The results confirmed the presence of eight exonic VUS in four genes, including MUC3A (c.2590_2592delinsGCT, p.S864A; c.3107_3108delinsAC, p.S1036N; c.3978_3979delinsCT, p.(=)), OR4C3 (c.469_470delinsTG, p. N157C; c.850_854delinsATGCT, p.V284_H285delinsML), MADCAM1 (c.742T >C, p.S248P), and POTED (c.28A >G, p.T10A; c.37A >G, p.T13A), in patients with PH-PM [Table 1 and Supplementary Figure 3, https://links.lww.com/CM9/C692]. Table 1 - Validation of the identified variants of unknown significance in patients with pul­monary hypertension-fibrous mediastinitis by Sanger sequencing. dbSNP ID Type Gene Reference Alteration hgvs.c hgvs.p rs386716152 Non-frameshift substitution MUC3A TCC GCT c.2590_2592delinsGCT p. S864A rs796586599 Non-frameshift substitution MUC3A GT AC c.3107_3108delinsAC p. S1036N rs796213668 Non-frameshift substitution MUC3A TC CT c.3978_3979delinsCT p. (=) rs386753295 Non-frameshift substitution OR4C3 AA TG c.469_470delinsTG p. N157C N/A Non-frameshift substitution OR4C3 GTGCA ATGCT c.850_854delinsATGCT p.V284_H285delinsML rs1140822 Non-synonymous SNV MADCAM1 T C c.742T >C p. S248P rs373326718 Non-synonymous SNV POTED A G c.28A >G p. T10A rs200720062 Non-synonymous SNV POTED A G c.37A >G p. T13A Alt: Sample genome allele; dpSNP:Database of single nucleotide polymorphisms; hgvs.c: The mutation results in a change in the nucleotide; hgvs.p: The mutation results in an altered amino acid; hgvs: Human Genome Variation Society; MADCAM1: Mucosal vascular addressin cell adhesion molecule 1; MUC3A: Mucin 3A; N/A: Not available; OR4C3: Olfactory receptor family 4 subfamily C member 3; POTED: POTE ankyrin domain family member D. p.(=): No amino acid change; Ref: reference genome allele. The pathological changes in FM are caused by the proliferation of cellular collagen and fibrous tissue in the mediastinum. We predicted that mucins may be involved in regulating the pathology of fibrosis in FM, as mucins regulate mucus viscoelasticity and play a key role in fibrosis diseases.[2] Indeed, the current study identified SNPs, such as rs386716152, rs796586599, and rs796213668, in the MUC3A (mucin 3A, cell surface associated) gene from PH-FM patients, suggesting the association of these SNPs with the fibrotic and inflammatory processes of this fatal disease. The abnormal expression of mucosal vascular addressin cell adhesion molecule 1 (MAdCAM-1), an adhesion molecule, is associated with the infiltration and activation of various inflammatory cells and the excessive deposition of extracellular matrix during the pathological processes of tissue fibrosis.[3] Therefore, the identification of non-frameshift mutations in the MAdCAM-1-coding gene in PH-FM patients may indicate the involvement of this adhesion molecule in disease progression. Although it is uncertain whether prostate, ovary, testis, and embryo (POTE) proteins are involved in the fibrotic process, this may be an important clue for the further discovery of whether POTE ankyrin domain family member D (POTED) proteins regulate fibrosis by mediating caspase-3, which is a key regulator or inducer of fibrosis formation.[4] Olfactory receptor family 4 subfamily C member 3 (OR4C3) is a member of the olfactory receptor (OR) family and is widely expressed and functions beyond the olfactory organ, such as macrophages,[5] which play a key role in facilitating fibrosis in various tissues. ORs on macrophages are a rich source of unexplored regulators of inflammation. Therefore, it is likely that ORs regulate the process of fibrosis through macrophages. This is another new way for us to study the pathological changes in fibrosing mediastinitis. In our case cohort, the small sample size and the lack of an adequate sample size to validate our results impose significant limitations on this study. However, this further emphasizes the rarity of this disease. In addition, the pathogenesis of this disease remaining still unknown and the lack of suitable animal models for the advanced study of the disease pathogenesis, prevention, and the relevant diagnostic and prognostic markers are important barriers presently faced. These data can be considered exploratory and await replication in independent cohorts. However, our findings suggest that genetic factors may contribute to the development of FM, providing a basis for future studies on the genetic mechanisms of this rare disease. In summary, genomic DNA from the peripheral blood of patients with PH-FM was analyzed by high-throughput WES. High-frequency mutations in four genes, including MUC3A, MAdCAM-1, POTED, and OR4C3, were identified by WES and validated by conventional Sanger sequencing. The findings in this study may shed light on the molecular genetic mechanism underpinning the initiation and progression of PH-FM and provide a clue for developing genetic screening and gene therapy strategies against this rare, but fatal disease. Funding This work was supported by grants from the National Natural Science Foundation of China (Nos. 82070052 and U25A2001), the Tianjin high-level talent selection and training project in the health industry (No. TJSJMYXYC-D2-042) the Natural Science Foundation of Gansu Province of China (No. 24JRRA1049) and the Lanzhou Science and Technology Program of Gansu Province of China (No. LX-62000001-2022-090). Conflicts of interest None.