The molecular genetics of the telomere biology disorders

ABSTRACT

The importance of telomere function for human health is exemplified by a collection of Mendelian disorders referred to as the telomere biology disorders (TBDs), telomeropathies, or syndromes of telomere shortening. Collectively, the TBDs cover a spectrum of conditions from multisystem disease presenting in infancy to isolated disease presentations in adulthood, most notably idiopathic pulmonary fibrosis. Eleven genes have been found mutated in the TBDs to date, each of which is linked to some aspect of telomere maintenance. This review summarizes the molecular defects that result from mutations in these genes, highlighting recent advances, including the addition of PARN to the TBD gene family and the discovery of heterozygous mutations in RTEL1 as a cause of familial pulmonary fibrosis.

Keywords:

• dyskeratosis congenita
• hTR
• idiopathic pulmonary fibrosis
• PARN
• RTEL1
• telomerase
• telomere
• telomeropathy
• TERC
• TERT

Introduction

The telomere biology disorders (TBDs) constitute a spectrum of clinical conditions that arise from short, dysfunctional telomeres. Telomeres are specialized nucleoprotein structures at chromosome ends, which protect the natural chromosome termini from the DNA damage signaling and repair activities that normally occur at ends created by DNA breaks. In addition, telomeric DNA, which serves as a substrate for the enzyme telomerase, provides a mechanism for the replenishment of terminal chromosomal DNA sequences that are not copied during semiconservative DNA replication as a result of the ‘end replication problem’. In humans, telomeric DNA consists of tandem repeats of duplex TTAGGG sequence, which varies in length considerably among individuals within populations. For example, at birth, the 1^st to 99^th percentile telomere lengths in human lymphocytes range from 8 to 14 kilobases.¹ Telomeric DNA terminates as a 75 to 300 nucleotide long single stranded overhang of the 3′ G-rich strand.^2-4 Although duplex telomeric DNA and the 3′ G-rich overhang are critical for normal telomere function, it is abnormally short duplex telomeric DNA length that is best characterized in the TBDs.⁵

Telomere length decreases 50 to 200 base pairs with each cell division in the absence of a mechanism to counteract the end replication problem.^6,7 Eventually, telomeres reach a critical threshold, which triggers cellular senescence.^7-9 Thus, telomeres function as a clock of replicative lifespan. Telomerase, a specialized reverse transcriptase with catalytic protein and RNA subunits, known as TERT and TR, respectively, counteracts this shortening by synthesising telomeric repeats on to the end of the 3' G-rich strand, using a sequence embedded within TR as the template.¹⁰ Most human somatic cells express insufficient or undetectable levels of telomerase, a result of repression of TERT gene expression.^11-14 Forced expression of TERT rescues the decline in telomere length and circumvents senescence, permitting continuous cell divisions.^15,16 Certain human cell populations such as germ cells, stem cells, and expanding lymphocytes, express telomerase, and, thereby, maintain telomere length and forestall telomere dysfunction-induced cellular senescence.¹⁷ Impaired telomere length maintenance in the germline results in progressively shorter telomere length over generations in murine models and is thought to underlie the phenomenon of disease anticipation that is frequently observed in TBD kinships.^18-20

In addition to sufficient duplex telomere DNA length and 3' overhang end structure, a cadre of proteins that bind directly to duplex telomeric repeats or the 3′ overhang, or indirectly to telomeric DNA via protein-protein interactions, are required for normal telomere maintenance and function. Multiple models systems have been instrumental in the study of these factors (e.g., ciliates, yeast, plants, and mouse), however, in some cases, their function in human cells differs from in these models, including mouse, or has not been adequately studied e.g.,^21,22 Nonetheless, these systems have greatly informed the study of human disease-associated mutations.

Among key factors in mammals are those that constitute the shelterin complex, TRF2, RAP1, TRF1, TIN2, TPP1, and POT1.²³ While these highly abundant proteins assemble into a macromolecular complex and subcomplexes along the telomere,²⁴ the function of individual components varies. TRF2, which binds as a homodimer to double stranded telomeric DNA, inhibits the telomere from inducing ATM signaling and being engaged in canonical nonhomologous end joining resulting in telomere-telomere fusions.^25,26 This is achieved in part by its central role in the formation or maintenance of the telomeric t-loop, a lasso-like structure formed by the invasion of the terminal single stranded 3′ overhang into the proximal duplex telomeric repeats with displacement of the G-rich strand and formation of a D-loop.^27-29 Surprisingly, RAP1, which localizes to telomeres via its interaction with TRF2, has no discernible effect on telomeres in human cells.³⁰ Instead, RAP1 also localizes to a limited set of interstitial sites containing telomeric sequences in both mice and humans, where it regulates gene expression.^31,32

TRF1, which shares a common domain structure with TRF2, homodimerizes and binds duplex telomeric DNA like TRF2.²⁶ Functionally, however, it is distinctly different. Instead of end protection, TRF1 promotes the efficient replication of duplex telomeric DNA. By virtue of the G-richness of the strand replicated by lagging strand synthesis and its propensity to form secondary structures, such as G quartets, the duplex telomeric tract is susceptible to replication fork stalling and collapse.^33,34 Fragile metaphase telomeres, which are multiple telomere fluorescent in situ hybridization (FISH) signals on a single chromatid arm, reflect such replication defects and are prominent in murine cells upon deletion of TRF1.³⁵ TRF1 specifically facilitates replication of the lagging strand and achieves this by recruiting the BLM helicase, which likely dismantles G quartet structures that form on the separated G-rich strand.^35,36 Finally, TRF1 is considered a negative regulator of telomerase.³⁷

A shared and critical function of TRF1 and TRF2 is to recruit the shelterin component TIN2 to telomeres. TIN2 then recruits the TPP1/POT1 heterodimer via interaction with the TPP1 subunit. The TPP1/POT1 heterodimer serves several important roles at the telomere. First, TPP1 is required for the recruitment of telomerase to human telomeres.³⁸ A cluster of glutamate (E) and leucine (L) amino acids on the surface of the OB-fold domain of TPP1, referred to as the TEL patch, is required for this recruitment.^39,40 Residues in the TEN domain of TERT are required for TERT's interaction with TPP1.⁴⁰ Expression of either TEL patch TPP1 or TERT TEN domain mutants in cancer cell lines prevent telomerase recruitment. Combining a TEL patch mutant with a TEN domain mutant rescued the defective TERT recruitment observed with each individual mutant protein, providing strong genetic evidence for their direct interaction. In addition, TPP1s interaction with TERT via the TEL patch promotes the ability of telomerase to catalyze repeat synthesis following the first event, a property referred to repeat addition processivity.^39,40 Finally, TPP1 localizes POT1 to the telomeric end where it functions to inhibit 5′ strand resection, ATR signaling, and sister telomere associations.^25,41

Thus, disruption in any of the above factors has the potential to yield dysfunctional telomeres and human disease. In addition to shelterin constituents, there are a variety of other proteins that function at telomeric and other genomic sites. They, too, have the potential to be pathogenic.

The telomere biology disorders

The clinical manifestations of the TBDs are thought to arise from the loss of critical stem cell populations, particularly in highly proliferative tissues, that is induced by short, dysfunctional telomeres. Several features of dyskeratosis congenita (DC), the first TBD to be described, reflect this. First is the mucocutaneous triad comprised of leukoplakia, a premalignant condition typically involving the oral mucosa, a tissue with high cellular turnover; reticulated skin pigmentation; and nail dystrophy, the latter 2 likely reflecting premature senescence of dermal and ectodermal stem cells, respectively. Bone marrow failure is another prominent feature of DC, with rates varying from 50% by the age of 50 y to 85% by age 30 y depending on the study and likely the criteria used to classify an individual as having DC.^42,43 Depletion of haematopoietic stem cells and bone marrow skeletal stem/progenitor cell defects underlie the bone marrow failure.⁴⁴ Severe and life-threatening gastrointestinal and liver disease, cancer, particularly squamous cell carcinoma of the tongue, and myelodysplastic syndrome,⁴⁵ along with numerous other medical problems, may additionally develop. Lastly, individuals with DC are at risk of pulmonary fibrosis, particularly after haematopoietic stem cell transplantation. Pulmonary fibrosis may also be the sole disease manifestation of a TBD, which presents in adulthood.⁴⁶ Emerging evidence from mouse models indicate that, despite the lung being a tissue of relatively slow cell turnover, telomere-dysfunction induced pulmonary fibrosis is also driven by the failure of stem cells, specifically, type II alveolar epithelial stem cells.^47,48 Lastly, some cases of familial predisposition to adult onset aplastic anemia, myelodysplastic syndrome/acute myeloid leukemia, or cirrhosis, alone or in combination with each other or with pulmonary fibrosis, but in the absence of the mucocutaneous triad, have also been ascribed to an underlying TBD, highlighting the variable presentation of these disorders.^49-52

DKC1, which encodes the nuclear protein dyskerin, was the first gene found mutated in DC.⁵³ The association of dyskerin with active telomerase and the reduced TR and telomerase activity in cells from patients with mutations in DKC1 provided the initial link between DC and defective telomere maintenance.⁵⁴ Subsequently, it was discovered that loss of a single functional allele of TERT or TERC, which encodes human TR (hTR), is sufficient to result in abnormally short telomeres, thereby, revealing the limiting nature of telomerase in normal human cells and telomerase haploinsufficiency as a mechanism underlying DC.^19,55-57

The Hoyeraal-Hreidarsson and Revesz syndromes, which are frequently referred to as severe variants of DC, also present in childhood, typically in infancy.^58,59 Some features, such as intrauterine growth retardation and microcephaly, are apparent at birth. Cerebellar hypoplasia is a defining feature of Hoyeraal-Hreidarsson syndrome and immunodeficiency may be severe and the presenting feature. Revesz syndrome is defined by the presence of bilateral exudative retinopathy or Coats disease. Intracranial calcification is often present. A third entity, referred to as Coats plus, overlaps with Revesz syndrome with the presence of exudate retinopathy or telangiectasias.⁶⁰ In addition, affected individuals have a pattern of asymmetric intracranial calcification, distinct from that observed in Revesz syndrome, with associated leukoencephalopathy and brain cysts; osteopenia with tendency to fracture and poor bone healing; and recurrent gastrointestinal hemorrhage due to vascular ectasias. In each of these conditions, additional features of DC may be present, including the mucocutaneous triad and bone marrow involvement, although the bone marrow involvement is less severe in Coats plus.

To date, 11 genes have been found mutated in patients in the TBDs (Table 1). In addition to the genes that encode TERT, hTR, and dyskerin are the genes that encode the shelterin components TIN2 and TPP1; RTEL1, a DNA helicase with functions at the telomere and at other DNA loci; CTC1 which is a component of the CTC1/STN1/TEN1 complex, which like RTEL1 has functions at telomeric and non-telomeric loci; NOP10 and NHP2, which are involved in telomerase biogenesis; TCAB1, which is required for the recruitment of telomerase to the telomere; and PARN, which has emerging functions in the regulation of telomere-biology related gene transcripts.^{19,53,56,61-68} Autosomal dominant, autosomal recessive, and X-linked recessive inheritance patterns are observed. For certain genes, such as DKC1 and particularly TINF2, which encodes TIN2, de novo mutations are common.^69,70 Notably, disease-associated heterozygous and homozygous or compound heterozygous mutations in TERT, RTEL1 and PARN have been observed, with the biallelic mutations associated with childhood onset and more severe disease.^66,71-74 Dyskeratosis congenita and Hoyeraal Hreidarsson syndrome are associated with mutations in multiple genes, whereas Revesz syndrome and Coats' plus have only been only associated with one gene each, TINF2 and CTC1, respectively. This, intriguingly, may reflect some unique aspect of the function of these factors.

Table 1. DC-associated Genes.

Gene (product)	Inheritance	Disease association	Effect of disease-associated mutations
DKC1 (dyskerin)	XLR, inherited or de novo	DC, HHS	Reduce hTR stability and telomerase activity
TERT (TERT)	AD	IPF, LD, AA, MDS, AL, DC	Reduce telomerase activity, processivity, or recruitment
o	AR	DC, HHS	o
TERC (hTR)	AD	IPF, LD, AA, MDS, AL, DC	Reduce telomerase activity
NHP2 (NHP2)	AR	DC	Reduce hTR stability and telomerase activity
NOP10 (NOP10)	AR	DC	Reduce hTR stability and telomerase activity
ACD (TPP1)	AD	AA	Reduce telomerase recruitment to telomeres
o	AR	DC	o
WRAP53 (TCAB1)	AR	DC	Impair telomerase trafficking through Cajal bodies and recruitment to telomeres
TINF2 (TIN2)	AD, mostly de novoAD, ? inherited	DC, HHS, RS, IPF	Multifactorial? Reduce telomere sister cohesion, reduce telomerase recruitment to telomeres, enhance telomere shortening (telomerase-independent)
RTEL1 (RTEL1)	AR	HHS	Impair t-loop stability resulting in t-loop excision
CTC1 (CTC1)	AR	CP, DC	Impair duplex telomere replication
PARN (PARN)	AR	HHS	Alter telomere biology gene transcripts

AD, autosomal dominant; AR, autosomal recessive; XLR, X-linked recessive.

AA, aplastic anemia; CP, Coats' plus; DC, dyskeratosis congenita; HHS, Hoyeraal Hreidarsson syndrome; AL, acute leukemia; IPF, idiopathic pulmonary fibrosis; LD, liver disease; MDS, myelodysplastic syndrome; RS, Revesz syndrome.

Unifying the impact of mutations in these genes and the pathophysiology of the TBDs is the presence of very short telomeres. Flow cytometry combined with telomere FISH (telomere flow FISH)⁷⁵ has been widely used to obtain average telomere length measurements in leukocyte subsets in patients with DC. Using this method, measured telomere lengths of less than the 1^st percentile in 3 of 4 of the lymphocyte subsets is both highly sensitive and specific for the diagnosis.⁷⁶ As this assay is clinically available, telomere length measurements are now obtained outside a research basis to facilitate the diagnosis of a TBD.

The measurement of telomere length using the telomere flow FISH assay in a large number of individuals within the spectrum of the TBDs in kinships and across studies suggests that the extent of telomere shortening underlies the contrasting infant, childhood, and adult onset presentations e.g..^46,77 This is perhaps best exemplified by the extent of telomere shortening observed in those whose TBD manifests as adult onset idiopathic pulmonary fibrosis (IPF), which is in fact the most common presentation of a TBD.⁵ IPF is a progressive and life-threatening disease of interstitial lung scarring. The estimates vary, but up to 20 percent of IPF probands have a family history of IPF, with their pedigrees displaying a pattern consistent with autosomal dominant inheritance with incomplete penetrance.⁷⁸ Mutations in TERT were the first to be identified in familial and some cases of sporadic IPF, linking a defect in telomere biology to this disease process.⁴⁶ Subsequent studies demonstrated up to 15% of IPF families and 1–3% of sporadic IPF cases to be due to TERT mutations.^79-81 A fewer percentage have mutations in TERC.^46,79 Mutations in DKC1 and TINF2 have also been reported in a few cases and kinships with IPF.^82-84 More recently, mutations in RTEL1 and PARN have been described (these are discussed in more detail below).^85-87 Whereas lymphocyte telomere lengths well below the 1^st percentile for age are present in individuals with childhood onset disease when analyzed by the telomere flow FISH method,⁷⁶ telomere lengths are not as profoundly short and instead most often hover around the 1^st percentile to 10^th percentile for age in those with familial IPF or sporadic IPF with a known telomere associated gene mutation.^46,80 This may account for the variable penetrance and expressivity that typifies the TERT and TERC mutations.

Mechanisms leading to short telomeres in the telomere biology disorders

The 11 genes mutated in the TBDs reflect the complexity of telomere length maintenance. In most cases, the impact of the mutations converges on a failure to maintain telomere length, although, in a few, there is evidence of telomere dysfunction without significant shortening. Here, the effects of the mutations are distributed into different classes, which will serve as a framework to consider potential new genes and avenues for therapy.

Class I: Reduction in telomerase activity, primary and secondary

Multiple mechanisms have been identified that result in the telomere shortening and dysfunction that underlies the TBDs. Most result in a reduction of telomerase activity at chromosome ends. The most common mechanism by which this occurs is through mutations in TERT, TERC, and DKC1. Mutations in TERT and TERC have the potential to impair telomerase catalytic activity. To date, over 75 different TERT mutations have been reported in patients with a TBD.⁸⁸ Most of these are novel. The challenge in interpreting these mutations is that nonsynonymous single nucleotide variants in TERT are not uncommon. TERT consists of 1132 amino acids and over 200 distinct missense variants have been reported in the Exome Aggregation Consortium (ExAC) database, which is comprised of whole exome sequencing data from approximately 60,000 individuals from several disease-specific and population genetic study cohorts. Nine loss-of-function TERT variants (stopgain, frameshift, and splice site) have also been reported. Of the variants identified, over half were found in only one or 2 of the 14,000 – 120,000 alleles sequenced. Thus, excluding the A279T and A1062T polymorphisms, which were present at allele frequencies of 3.6 and 1.3 percent, respectively, the sum total of all the remaining variants suggests that up to 1% of individuals will harbor a rare TERT variant. This underscores the importance of functional analysis of identified TERT variants to determine whether it may be causative of disease.

Several approaches have been used to determine the impact of disease-associated TERT variants on telomerase function, mostly in cell free systems. Some studies have used the telomerase repeat amplification protocol (TRAP) assay, which includes a polymerase chain reaction (PCR) step to amplify telomerase extended oligonucleotides.¹¹ However, because the PCR amplification step includes a primer that can anneal to any of the repetitive telomeric repeats that are added on to the 5′ end of the telomerase extended oligonucleotide, the TRAP assay is only semi-quantitative and does not allow for a measure of repeat addition processivity. An alternative and more informative approach has been the direct primer extension assay, in which telomerase extended telomeric oligonucleotides are analyzed without PCR amplification. This assays allows for a quantitative assessment of enzyme activity as well as repeat addition processivity. Using the direct primer extension, variants affecting activity have been identified as well as others that appear to selectively impact repeat addition processivity e.g.,.^74,89 Strikingly, however, Zaug et al., showed that several putative disease-associated mutations exhibited little discernible defect using this approach. In their carefully controlled, quantitative analysis of 19 mutant telomerases containing disease-associated mutations, they found 10 retained 60% or greater telomerase activity.⁹⁰ Because the vast majority of TERT mutations that have been identified in patients with DC and all that have been identified in patients with isolated adult onset pulmonary fibrosis are heterozygous, this would translate to 80% or greater activity in cells. Importantly, they also found in 3 of 9 cases statistically significant differences in the telomerase activity when it was synthesized in rabbit reticulocyte lysates compared to when it was purified from cells overexpressing transiently transfected TERT and TERC constructs. One mutant exhibited approximately 100% activity in the rabbit reticulocyte lysate but only 60% in HEK lysates. Thus, this analysis revealed yet another potential variable in the assessment of these mutations.

As discussed below, mutations in telomerase may result in a defect that is apparent in cells only, as would be the case, e.g., if it affected the ability of telomerase to be recruited to telomeres. Thus, mutants that appear to track with disease in a kinship yet do not exhibit a defect in vitro should be examined for their impact on telomere elongation or length maintenance in cells. Additionally, it remains to be determined whether minor reductions in telomerase activity may be clinically significant in certain cases. While such reductions may be tolerable under normal conditions, they may be contributory when there are other factors driving telomere shortening (e.g.,, bone marrow stress or smoke inhalation driving proliferation of haematopoietic or lung alveolar stem cells, respectively).

TERC is much less frequently found mutated in the TBDs than TERT. hTR is 454 nucleotides in length and approximately 60 different mutations have been reported.⁸⁸ Some of these are clearly deleterious, deleting large segments of the RNA or targeting essential components, such as the template region. However, others are single nucleotide substitutions that may or may not be of functional significance. Sixty-two unique single nucleotide variants are reported in ExAC. The two most common are still infrequent, present in 0.89 and 0.19 percent of individual sequences. The others are present in 1–10 alleles of the up to 111,000 alleles sequenced, resulting in another 0.13% of individuals carrying a variant containing allele, assuming nearly all of the variants were present in the absence of one or more of the others. Thus, hTR appears to be more conserved in the general population as compared to TERT, although it is important to note that the gene target size is considerably smaller.

As with TERT, functional analysis is helpful to assess the impact of any identified hTR variant on function. A systematic approach to the analysis of disease-associated mutant hTR as described above for TERT has been similarly reported. In this work, Robart and Collins analyzed the impact of 13 hTR mutations on a range of relevant properties such as hTR stability, hTR/TERT interaction, telomerase activity and repeat addition processivity.⁹¹ They, too, found, with a smaller subset of TERT variants, a number with no apparent impact on telomerase activity or repeat addition processivity. This suggests a lower tolerance for hTR variation due to the likely impact of most variants on telomere length maintenance.

In addition to impairing telomerase catalytic activity via mutations in TERT or TERC, decreased levels of telomerase are found in association with mutations in genes that encode proteins required for telomerase biogenesis and hTR stability. These genes include DKC1, as described above, NHP2 and NOP10. hTR is an H/ACA motif-containing RNA, and similar to other H/ACA RNAs, requires dyskerin, NHP2, and NOP10 for its assembly.^54,92 Disease-associated mutations in DKC1, NHP2, and NOP10 have been shown to impact telomerase assembly and stability.^93,94 NAF1, SHQ1, and pontin/reptin are also involved in telomerase biogenesis,^92,95 however, mutations in the genes encoding these factors have not been identified in patients to date.

Class II: Impaired telomerase trafficking and recruitment to telomeres

Once assembled, additional factors are required for telomerase to be recruited to telomeres. One such factor is TCAB1, an essential component of the dynamic subnuclear structures known as Cajal bodies.⁹⁶ TCAB1 binds the 4 nucleotide CAB box present on hTR as well as on small Cajal body-associated RNAs.⁹⁷ Depletion of TCAB1 results in a relocalization of telomerase from Cajal bodies to the nucleolus, a reduction in telomerase recruitment to telomeres, and telomere shortening (Venteicher et al., 2009; Zhong et al., 2011, 2012). The chaperone TCP-1 Ring Complex (TRiC) was recently found to mediate TCAB1 folding and, consistent with this, similarly impacts telomerase recruitment and localization and impairs the addition of telomeric repeats onto chromosome ends.⁹⁸ In contrast to what is observed in the absence of dyskerin, which also binds directly to hTR, these effects are not due to hTR instability.

Compound heterozygous mutations in the gene encoding TCAB1, WRAP53, have been identified in patients with DC, and were the first in this mechanistic class of mutations that impair telomerase recruitment.⁶³ Strikingly, the missense mutations map to the region to which TRiC binds and mediates folding, and result in marked reduction in TCAB1 accumulation.⁹⁸ Cells from these patients, as expected from the TCAB1 depletion studies, retained telomerase activity when measured in cell lysates, but demonstrated reduced telomerase association with telomeres and mislocalization of telomerase to the nucleolus.⁶³

The molecular mechanism underlying TCAB1's essential influence on telomerase recruitment to telomeres remains unknown. Although the recruitment of telomerase to telomeres is also impaired when the formation of Cajal bodies is blocked by depletion of the Cajal body integral component coilin, the recruitment can be rescued by overexpression of TERT and hTR.⁹⁹ Such a rescue is not observed in TCAB1 depleted cells. Together these results indicate that, while TCAB1-mediated trafficking of telomerase to Cajal bodies is important for the recruitment of endogenous telomerase to telomeres, TCAB1 contributes an additional, Cajal body-independent activity. The identification of a TCAB1 separation-of-function mutant that retains the ability to localize telomerase to the Cajal body but can no longer promote telomerase recruitment to the telomere would be a useful tool in uncovering this additional activity.

The association of telomerase with TCAB1 is necessary, but not sufficient for its recruitment to the telomere. As described above, telomerase recruitment also requires the interaction of TERT with the TEL patch on TPP1. Underscoring the importance of this interaction was the identification of 2 unrelated probands with an in-frame deletion of a residue in the TEL patch, K170.⁶⁸ One of the probands had telomere length at the 1^st percentile and he and relatives with the K170del mutation had hematologic diseases and cancer in the TBD spectrum. The second proband and his identical twin, which were reported by another group, had Hoyeraal Hreidarsson syndrome.⁶⁷ They also carried on the other TPP1 allele a missense change, P491T, which may have modified the phenotype to a more severe one compared to that of the other family and their own father and sibling who carried only the TPP1 K170del mutation and had very short telomeres, but minor or no features to date in the TBD spectrum. In support of the possibility that heterozygosity of the K170del allele is insufficient to induce severe telomere shortening and early onset, multisystem disease phenotype, the ExAC database reports 27 unique deleterious (stopgain, frameshift, and splice site) variants in ACD, with each present in 1 to 15 of the on average 113,000 alleles sequenced. While individually these are very low, the frequency of any being present in an individual (0.1%) is greater than predicted for a Hoyeraal Hreidarsson syndrome-inducing variant and individuals with such a severe phenotype would have been unlikely to be in the aggregated cohorts. Identification and characterization of additional patients with ACD/TPP1 mutations will help further clarify this question. Similarly will those with mutations that impact the TERT side of the TPP1-TERT interaction. One such recruitment defective TERT mutant that retains full catalytic activity has been characterized.⁴⁰ The mutation, V114M, maps to the TEN domain and was identified in 2 kinships with IPF and blood dyscrasias.¹⁰⁰

Class III: Impairment in duplex telomere replication and C-strand fill in

Each of the genes described above has a direct connection with telomerase, by either encoding a component of the holoenzyme or a factor required for telomerase biogenesis or recruitment to telomeres. Telomere maintenance also requires fill-in synthesis of the C-rich strand following telomerase-mediated elongation of the G-rich strand and efficient semi-conservative replication of the duplex telomeric tract. Thus, not unexpectedly, mutations have been found in genes that are critical for these processes in patients with TBDs.^65,101-103 The first of these genes is CTC1, which encodes the CTC1 component of CTC1/STN1/TEN1 (CST) complex. CTC1 promotes the re-start of telomere lagging strand synthesis at stalled replication forks and fill-in synthesis of the C-rich strand at the chromosome terminus,^104,105 properties shared by its binding partners STN1 and TEN1.¹⁰⁶ This requires the binding of CTC1 to single-stranded DNA and its association with the replication initiator polα primase complex. While these properties suggest that CTC1 might also have a role in general genomic DNA replication, this is restricted to regions experiencing replication stress, where it stimulates the firing of late or dormant origins.^106,107

Mutations in CTC1 were first identified in patients Coats plus.^65,101 Interestingly, whereas classic DC is a genetically heterogeneous condition, Coats plus is not, with mutations in CTC1 being identified in vast majority of a well-defined Coats plus cohort.⁶⁵ Subsequently, biallelic CTC1 mutations were identified in cohorts with classic DC.^102,103 Notably, a few of these patients lacked retinopathy or brain abnormalities, indicating that these features are not universally found in patients with germline CTC1 mutations.

In contrast to the other gene mutations resulting in a TBD with childhood onset disease, CTC1 mutations do not consistently result in short telomeres.^65,101-103 It is possible that these differences are the result of differences in technique for measuring telomere length. Notably, functional analyses of disease-associated CTC1 mutations revealed that they variably impaired the association of CTC1 with its binding partners STN1, TEN1, and polα-primase, the ability of CTC1 to binding telomeric ss DNA binding in vitro and telomeric chromatin association in vivo, and cellular localization.¹⁰⁸ However, they uniformly lead to the accumulation of internal gaps of single stranded telomeric DNA, indicating a telomere replication defect. While it is predicted, mutations in the genes encoding CTC1's binding partners STN1 and TEN1 have not been reported to date in patients.

Class IV: Telomere instability

Similar to CTC1, RTEL1 (regulator of telomere length 1) is a factor implicated in the TBDs, but acts at both telomeric and nontelomeric sites. RTEL1 is an essential DNA helicase that dismantles D-loop structures, which may form as natural intermediates in recombinational repair.¹⁰⁹ In addition, it disrupts the D-loops found at the base of t-loops, resolving the t-loop structure.¹¹⁰ RTEL1 is recruited to telomeres in late S phase by TRF2, and this is essential to prevent catastrophic telomere loss by SLX1/4 nuclease-dependent excision of double-stranded telomere circles (T-circles).¹¹¹ RTEL1 is also required for the integrity of duplex telomeric DNA replication, through its G-quartet unwinding activity, however, this occurs via a PCNA-dependent, TRF2-independent mechanism.^111,112

RTEL1 mutations have been found in both patients with Hoyeraal Hreidarsson syndrome and familial pulmonary fibrosis. Autosomal recessive transmission of compound heterozygous or homozygous RTEL1 mutations result in Hoyeraal Hreidarsson syndrome and are associated with very short telomeres.^71,113-115 Notably, the R1264H mutation, for which there is a carrier frequency of 1% in the Ashkenazi Jewish population,¹¹⁶ impairs RTEL1 interaction with TRF2 and, therefore, RTEL1 recruitment to telomeres.¹¹¹ Cells from patients with homozygous R1264H mutations demonstrated very short telomeres and elevated T-circles, underscoring the critical nature of this interaction and RTEL1's telomeric role in preventing catastrophic telomere loss.¹¹³ In a mouse model, the R1264H mutation preserved RTEL1's interaction with PCNA and its role in telomere and genome-wide replication.¹¹¹ Thus, this separation-of-function mutant revealed the essential requirement of RTEL1's t-loop unwinding activity in human telomere maintenance.

To date, close to 30 RTEL1 variants have been reported in autosomal recessive disease with pediatric onset or autosomal dominant familial pulmonary fibrosis, and mutations in RTEL1 are projected to be present in up to 5% of IPF families.^85,86,117 Many of the variants are deleterious whereas others are missense changes of uncertain significance. Some map to known domains, such as the helicase, harmonin homology domains, or the C4C4 metal-binding motif to which TRF2 binds, whereas others localize to uncharacterized regions. Whether some of the mutants will only impact RTEL1s role in promoting duplex telomere replication by unwinding G4 quartets will be of interest and inform the relative importance of this function vis a vis its t-loop unwinding activity.

Class V: Multifactorial

In contrast to the mechanisms described above, why mutations in TINF2, which encodes the shelterin component TIN2, result in such severe telomere shortening remains to be fully elucidated. Mutations in TINF2, which are heterozygous and only rarely inherited, result in early-onset disease, with many cases of Hoyeraal Hreidarsson and Revesz syndromes reported.^61,70,118 This contrasts with the observation that heterozygous mutations in telomerase component genes may lead to more severe manifestations of disease and progressively earlier onset disease over multiple generations as well as the adult onset phenotypes observed with heterozygous RTEL1 and PARN mutations. All DC-associated TINF2 mutations reported to date, whether missense or truncating, cluster in a 34-amino acid segment located centrally within both the short (TIN2S_1–354) and long (TIN2L_1–451) TIN2 isoforms. DC-associated mutations in TINF2 have been shown to impair some TIN2 functions, such as binding to heterochromatin protein 1γ (HP1γ), which is crucial for its role in sister telomere cohesion; however, not all mutations impair this interaction.¹¹⁹ An additional study suggested that DC-associated, mutant TIN2 impaired the recruitment of telomerase to the telomere;¹²⁰ however, a mouse model of TIN2-associated DC demonstrated that telomere shortening occurs even in the absence of telomerase and, therefore, is not a result of decreased telomerase activity at telomeres.¹²¹ TIN2 provides the tether for the TPP1/POT1 heterodimer to the telomere, thus, disruption of the TIN2-TPP1 interaction could have markedly deleterious effects. However, in overexpression co-immunoprecipitation studies, a panel of DC-associated TIN2 mutants interacted with TPP1 similarly to wild type.¹²² Thus, the study of the mechanism(s) by which TINF2 mutations impair telomere maintenance is likely to shed new insight into crucial functions of TIN2.

One important consideration is whether the mutations exert a dominant negative effect. Findings in patients with TINF2 mutations and adult onset IPF suggest that this may be the case. Two groups have reported TINF2 mutations in adults with IPF, with and without additional features of DC.^84,123 One described a woman who presented with IPF at the age of 43 y.¹²³ Her leukocyte telomere length was short relative to age-matched controls. Sanger sequencing of a cloned segment of PCR amplified DNA encompassing the DC cluster revealed a deletion of nucleotides 871–874, resulting in a frameshift mutation, TIN2p.R291Ifs11X. Because the patient did not develop severe bone marrow failure during the first decade of life, as typically observed in patients carrying a mutation in TINF2, it was suggested subsequently that perhaps she experienced somatic mosaicism, with reversion of the mutant TINF2 allele in some of the peripheral blood cells, thereby, attenuating the bone marrow phenotype.¹²⁴ The mosaicism might have been missed by sequence analysis of the cloned PCR fragment.

Another report focused on a family with IPF, idiopathic infertility, and cryptogenic liver disease.⁸⁴ The proband was a 49-year-old woman who had IPF but none of the major features of DC. She had lymphocyte telomere lengths at the 1^st percentile for age. Sequence analysis revealed 2 mutations in TINF2, a 15 nucleotide deletion encompassing the exon 6 splice acceptor site, and a T284R missense mutation. Heterogeneity in the presence of one, both, or none of the TINF2 mutations in cloned PCR products amplified from peripheral blood and lung tissue DNA template led the authors to conclude that the deletion mutation was acquired in the haematopoietic cells, whereas the missense mutation was present in the germline. Expression analysis suggested that the 15 nucleotide deletion and T284R double mutant allele was subject to nonsense mediated decay, which was supported by the loss of expression of the TIN2L isoform. Together these results led to the hypothesis that the acquired deletion mutation in the haematopoietic lineage suppressed a dominant negative effect of the missense mutation. However, the T284R mutation had been previously reported in a 66-year-old woman with BMF and minor features of DC who was also reported to have lung disease, raising the likelihood that the T284R mutation is a hypomorphic mutation, quite distinct from the R282H and other mutations reported in children. In summary these studies clearly implicate TINF2 mutations as rare causes of adult onset IPF. It may be possible the effects of the mutations on TIN2 function are less severe than those more commonly found in patients, such as the R282H mutation. A direct test of whether DC-associated TIN2 mutations exert a dominant negative effect is warranted.

Deficiency in PARN

The most recently discovered gene implicated in the TBDs is PARN, which encodes a poly(A)-specific 3' exoribonuclease, comprised of a catalytic nuclease domain, 2 regions with RNA binding properties, known as the R3H and RRM domains, and an unstructured C-terminal tail.¹²⁵ Rare, heterozygous PARN variants were first reported in 6 kindreds with familial pulmonary fibrosis.⁸⁶ Five were stopgain, frameshift, or splice site variants, whereas one resulted in a K to R substitution at residue 421, a conserved residue in the nuclease domain. In most, but not all, cases, the leukocyte telomere length in probands was below the 1^st percentile. In contrast to the presentation with adult onset pulmonary fibrosis with these heterozygous mutations, a subsequent report described 4 cases (2 of which were siblings) with Hoyeraal Hreidarsson syndrome or an overlap of Hoyeraal Hreidarsson/DC syndromes due biallelic variants in PARN.⁶⁶ The siblings were homozygous for an A383V substitution, within the nuclease domain, and had stable expression of the mutant proteins in lymphoblastoid cell lines. The third case was homozygous for a splice donor site variant, which resulted in detectable expression of 2 splice variants, predicted to result in either an in-frame deletion within the nuclease domain or a frameshift mutation leading to loss of the RRM and C-terminal tail. The fourth case was compound heterozygous for an indel, resulting in a frameshift and premature stop codon, and a splice site variant, predicted to result in an in-frame deletion of residues within the R3H domain, although cDNA was not available to confirm stable expression of these transcripts. Functional analysis of transiently expressed PARN with the A383V change resulted in deadenylation defect and in silico analysis suggested structural alteration of the catalytic domain. PARN interacts directly with the m⁷G-cap and poly(A) tail during poly(A) hydrolysis, thus impacts mRNA function.¹²⁵ In addition, PARN is involved in the maturation of certain non-coding RNAs, such as H/ACA box small nucleolar RNAs. Consistent with these functions, cells from the children with the biallelic PARN mutations and PARN-depleted cells exhibited reduced RNA hTR, DKC1, RTEL1, and TRF1 transcripts. Whether these reductions are sufficient to result in the predicted telomeric defects, such as reduced telomerase activity or increased fragile telomeres or T-circles, remains to be determined.

Conclusions

As many as 80% of IPF families and 40% of probands with DC remain genetically uncharacterized.^66,85 It is likely that whole exome sequencing approaches will continue to pare down the percentages of uncharacterized cases and may, excitingly, lead to the discovery of genes not previously implicated in telomere maintenance, as was the case for PARN.⁶⁶ We may also find novel or exceedingly rare variants in known telomere biology genes, in which case, functional analyses will be essential to determining their contribution. As telomere basic science research continues to thrive, discovery of new candidates, such as components of TRiC,^97,126 should regularly prompt re-inspection of existing whole exome sequencing data. Mouse and other models will continue provide valuable insight into the impact of specific genes on telomere biology, however, ever-improving gene editing technologies should now allow evaluation of these variants in relevant primary human cells. Lastly, there is emerging data in support of a new type of TBD that manifests as predisposition to specific tumor types, such as melanoma and glioma, that is due to mutations in POT1 and result in not short, but long, dysfunctional telomeres.^127-129 The study of how these mutations result in telomere dysfunction and cancer predisposition will be important as therapies designed to restore telomere length in the classical TBDs are considered.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.