Monogenic gene variants in lung transplant recipients with usual interstitial pneumonia

Discussion

Our results suggest that almost half of the patients with histologically confirmed UIP receiving lung transplants in Finland have a genetic diagnosis possibly contributing to the development of IPF. 39% of the patients had a molecular diagnosis of telomere disease. A recent report discovered telomere disease-causing variants by using exome sequencing in only 11.8% of lung transplant recipients [13]. The difference probably highlights globally variable indications for lung transplantation, population differences in genetic contribution and differences between the genetic assays to identify difficult-to-detect genetic variants. We only included patients with confirmed histopathological UIP, and it is possible that genetic variants are over-represented in this group compared to populations with other types of IPF. In Finland, 50% of lung transplantations are performed on patients with interstitial lung disease [7], while the number in North America ranges from 22 to 32% [6, 12]. The penetrance of telomere disease-related variants is variable. In this study, we did not measure telomere lengths, which could have provided further evidence that there is a causative relationship between the observed genetic variant and the patients’ phenotypes. It is, however, important to point out that patients can express normal pulmonary phenotypes even though their telomeres are shortened [16, 17]. In addition, patients with IPF often have shortened telomeres even without telomere-related genetic disease [17].

Interestingly, only one of the 11 patients with a molecular diagnosis had a positive family history for IPF, whereas the molecular diagnoses of the remaining 10 patients were novel. Sporadic variants have previously been observed in only one out of 10 IPF patients [16, 17]. It is possible that some familial cases in this study have previously been undiagnosed or that genetic contribution is underestimated or remains undiscovered in general since genetic testing is not utilised routinely in the workup of these patients.

Our study provides statistical evidence that TERT p.(Asp684Gly) and p.(Arg774*) are the most common missense and nonsense variants related to IPF, and that both of the variants are highly enriched in the Finnish population. This may indicate that telomere-related IPF may be more common in Finland than in other countries, but this hypothesis needs further testing by assessing the penetrance of these variants. TERT variants are found in 8–15% of patients with autosomal dominant familial pulmonary fibrosis [16, 18]. Disease caused by mutations in TERT are inherited in both autosomal dominant and recessive manners but is considered dominant with reduced penetrance in relation to pulmonary fibrosis [17]. In the current study, 26% of patients had heterozygous TERT variants. The TERT p.(Asp684Gly) variant has previously been observed in one homozygous and two heterozygote individuals out of 135 patients analysed for suspected telomere-related disease [19].

Genetic variants for PARN are associated with impaired telomere maintenance and are found in ∼5% of cases with familial pulmonary fibrosis [18]. To our knowledge, the observed PARN deletions described here have not been previously published in the relevant medical literature or reported in the disease-related variation databases. There are, however, multiple multi-exon and whole-gene deletions listed in HGMD in patients with telomere shortening-related disease. For PARN c.(1480+1_1481–1)_(1864+1_1865-1)del, in this study there is a partially overlapping de novo deletion (exons 23–24) identified previously in a patient with bone marrow failure and neurological dysfunction [20].

TINF2 encodes a protein that converts uridine to pseudouridine in ribosomal RNA and is part of a complex protecting telomere ends. In familial pulmonary fibrosis, TINF2 variants are observed in 0.6% of cases. In this study, one patient had two variants for TINF2. The missense variant c.840G>C, p.(Lys280Asn) has previously been associated with dyskeratosis congenita [21]. The other c.766C>T, p.(Arg256*) variant was present only in part of the gene reads, suggesting that it was mosaic. Since the DNA sample was extracted from saliva, we could not verify whether the variant was germline mosaic or a somatic mutation. The mosaic variant causes a premature stop codon and loss of normal protein function either through protein truncation or nonsense-mediated mRNA decay from the other allele and might modify the biological mechanism of the other missense variant for this patient as it prevents translation of the other pathogenic missense variant.

HPS1 encodes a transmembrane protein that forms a lysosomal complex together with Hermansky–Pudlak syndrome 4 protein. Variants in HPS1 are associated with autosomal recessive Hermansky–Pudlak syndrome. The syndrome is characterised by albinism, bleeding diathesis, immunodeficiency and pulmonary fibrosis. The variant observed here has previously been found as homozygous in one patient with HPS [22]. In HPS, life-threatening manifestations, such as restrictive lung disease, occur usually starting from the fourth decade of life. There is, however, large variation in disease severity, possibly related to genetic anticipation.

SFTPA2 encodes surfactant protein A2. Surfactant-related genetic variants are thought to increase cellular damage by causing endoplasmic reticulum stress, and genetic variants have been associated with autosomal dominant IPF [23]. The observed variant has not previously been reported, but there was strong association between the mutation and the patient's phenotype. The same patient had a variant for EDN3, encoding Endothelin 3. Variants in this gene have been associated with congenital central hypoventilation syndrome and Waardenburg syndrome. Waardenburg syndrome is characterised by pigmentation abnormalities, hearing loss, musculoskeletal abnormalities and Hirschprung disease [24]. It has not previously been linked to pulmonary fibrosis in the literature but might explain IPF through hypoventilation-related mechanisms. The variant generates a frameshift leading to a premature stop codon in a new reading frame. It is predicted to cause loss of normal protein function either through protein truncation or nonsense-mediated mRNA decay. To our knowledge, the variant has not been published in the relevant medical literature or reported in the disease-related variation databases. Disease caused by EDN3 variants are inherited in both autosomal recessive and dominant manners. The variant was considered likely pathogenic considering the established association between the gene and patient's phenotype, rarity in control populations and frameshift mutation type.

DNAH9 encodes a microtubule-associated motor protein. Variants in this gene are associated with primary ciliary dyskinesia and several congenital defects. It is, however, inherited in an autosomal recessive manner and does not cause disease without another disease-causing variant in the same gene [25]. Variants in SCN4A, encoding α subunits of sodium channels, have been associated with disease that causes myotonia or periodic paralysis, and they are inherited in both autosomal recessive and dominant manners. The phenotype is highly variable, and disease onset usually occurs in infancy or early childhood. Muscle and soft tissue symptoms are the most common manifestations [26]. Inspiratory stridor in early life is the only related respiratory disorder that has been reported previously [27].

The variants for SERPINA1 in this study were heterozygous and are not able to cause disease without a second disease-causing variant in the same gene. The p.(Glu366Lys) variant removes a salt bridge to Lys290 and a hydrogen bond to Thr203, causing misfolding of the protein within the endoplasmic reticulum, which results in a lack of secretion from hepatocytes and a reduction of plasma α-1-antitrypsin levels to 10–15% of normal in homozygotes [28] and to 61–77% in heterozygotes [29, 30]. Heterozygous carrier of p.(Glu366Lys) may have moderately decreased to normal serum α-1 antitrypsin levels and could therefore influence pulmonary phenotype.

In this study, we analysed the incidence of monogenetic disease variants and did not include analysis for common gene variants and genes associated with multigenic susceptibility to IPF. Because of the study design, we did not include survival as an end-point. We investigated possible impact of telomere-related variants on disease progression before transplantation and the incidence of acute rejection and CLAD after transplantation. Previously, these variants have been associated with several complications after transplantation but their impact on CLAD development or survival has been controversial [10, 13, 14]. Based on our results, the risk of developing CLAD and time to CLAD onset was not different between patients with telomere-related and non-telomere-related disease. Patient age at disease onset or transplantation, or the time from diagnosis to transplantation did not differ between the groups. This indicates that disease progression was not related to telomeropathy. Leukopenia or other forms of bone marrow dyscrasia is frequent after lung transplantation [13, 14]. We failed to show any association between bone marrow dysfunction and telomere-related IPF. The overall incidence of post-operative renal dysfunction was more common in this study than in many previous reports, but it was not related to telomere disease. The only malignancies observed in this study were epidermoid skin carcinomas, and they occurred at a similar rate as previously reported [13, 14], with no significant over-representation in telomeropathy patients.

This study has several limitations. It was a retrospective study with small sample size. Patient selection bias is possible because we only included patients that were eligible for lung transplantation and were alive several years after transplantation. These results are not necessarily generalisable to all IPF patients. All patients in this study were included based on the same inclusion criteria, and the two study groups are thereby comparable.